&EPA
EPA/635/R-03/012
www.epa.gov/iris
TOXICOLOGICAL REVIEW
OF
n-HEXANE
(CAS No. 110-54-3)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
November 2005
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document has been reviewed in accordance with United States Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
11
-------
CONTENTS—TOXICOLOGICAL REVIEW OF n-HEXANE
(CAS No. 110-54-3)
LIST OF TABLES AND FIGURES v
FOREWORD viii
AUTHORS, CONTRIBUTORS, AND REVIEWERS ix
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS 3
3. TOXICOKINETICS RELEVANT TO ASSESSMENTS 4
3.1. ABSORPTION 4
3.2. DISTRIBUTION 5
3.3. METABOLISM 8
3.4. ELIMINATION 17
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS 21
4. HAZARD IDENTIFICATION 25
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY AND CASE REPORTS 25
4.1.1. Oral Exposure 25
4.1.2. Inhalation Exposure 25
4.1.2.1. Subchronic Exposure 25
4.1.2.2. Chronic Exposure 25
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION 47
4.2.1. Oral Exposure 47
4.2.1.1. Subchronic Studies 47
4.2.1.2. Chronic Studies 49
4.2.2. Inhalation Exposure 49
4.2.2.1. Subchronic Studies 49
4.2.2.2. Chronic Studies 58
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND
INHALATION 58
4.3.1. Oral Studies 58
4.3.2. Inhalation Studies 59
4.4. OTHER STUDIES 63
4.4.1. Acute Toxicity Data 63
4.4.2. Studies with Mixtures Containing n-Hexane 67
4.4.2.1. Oral Exposure 68
4.4.2.2. Inhalation Exposure 68
4.4.2.2.1. Subchronic studies 68
in
-------
4.4.2.2.2. Chronic exposure 69
4.4.2.2.3. Reproduction/developmental studies 70
4.4.3. Potentiation and Antagonism Studies 73
4.4.4. Mode of Action Studies 82
4.4.5. Genotoxicity Studies 90
4.5. SYNTHESIS AND EVALUATION OF MAJORNONCANCER EFFECTS AND
MODE OF ACTION—ORAL AND INHALATION 93
4.5.1. Oral Exposure 93
4.5.2. Inhalation 94
4.5.3. Mode of Action Information 104
4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER
CHARACTERIZATION 105
4.6.1. Summary of Overall Weight-of-Evidence 105
4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence 106
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 108
4.7.1. Possible Childhood Susceptibility 108
4.7.2. Possible Gender Differences 108
5. DOSE RESPONSE ASSESSMENT 110
5.1. ORAL REFERENCE DOSE 110
5.2. INHALATION REFERENCE CONCENTRATION Ill
5.2.1. Choice of Principal Study and Critical Effect — with Rationale and
Justification Ill
5.2.2. Methods of Analysis 116
5.2.2.1. Adjustment to a Human Equivalent Exposure Concentration .119
5.2.3. RfC Derivation—Including Application of Uncertainty Factors (UFs) . 120
5.2.4 Previous RfC 122
5.3. CANCER ASSESSMENT 123
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE 125
6.1. HUMAN HAZARD POTENTIAL 125
6.2. DOSE RESPONSE 126
6.2.1. Noncancer 126
6.2.2. Cancer 127
7. REFERENCES 128
APPENDIX A: SUMMARY OF EXTERNAL PANEL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION A-l
APPENDIX B: BENCHMARK DOSE (BMD) ANALYSIS B-l
IV
-------
LIST OF TABLES AND FIGURES
Table 3-1. Concentration of n-hexane in blood and tissues of pregnant F344 rats immediately
after a 6-hour exposure to 1000 ppm n-hexane 6
Table 3-2. Tissue distribution of radioactivity in male F344 rats 72 hours after a 6-hour
inhalation exposure to various concentrations of [1,2-14C 6
Table 3-3. Apparent steady state concentration of n-hexane concentrations in male F344 rats
after 6 hours inhalation exposure to [l,2-14C]-n-hexane 7
Table 3-4. Metabolism of n-hexane following a 6-hour exposure of pregnant F344 rats on
gestation day 20 11
Table 3-5. Apparent kinetic parameters for n-hexane hydroxylation in rat liver and lung
microsomes 13
Table 3-6. Metabolites excreted in urine during a 72-hour period following inhalation exposure
to n-hexane in male F344 rats 20
Table 3-7. n-Hexane metabolite levels in urine of Wistar rats coexposed to n-hexane and
toluene 21
Table 4-1. Persistent and transient neurological symptoms following occupational exposure to
n-hexane in a tungsten carbide alloy factory 27
Table 4-2. Results of neurological tests in control subjects and those occupationally exposed to
n-hexane in a tungsten carbide alloy factory 28
Table 4-3. Nerve stimulation in control subjects and those occupationally exposed to n-hexane
in a tungsten carbide alloy factory 29
Table 4-4. Motor neurographic findings in patients with n-hexane polyneuropathy 37
Table 4-5. Nerve conduction study findings in printers with n-hexane-induced
polyneuropathy 38
Table 4-6. FM-100 hue test error scores of n-hexane-exposed and nonexposed groups 40
Table 4-7. Nerve conduction changes in male Wistar rats exposed 12 hours/day for 24 weeks to
200 or 500 ppm n-hexane 50
Table 4-8. Experimental protocol for Phases I and II of a 6-month inhalation study of n-hexane
and a mixture containing hydrocarbon isomers plus n-hexane in male Sprague-Dawley
rats 53
-------
Table 4-9. Relative organ weights of male Sprague-Dawley rats exposed to n-hexane 22
hours/day, 7 days/week for 6 months 54
Table 4-10. Incidence of nasal turbinate and neuropathological lesions in B6C3F1 mice exposed
to n-hexane for 13 weeks 56
Table 4-11. Skeletal variations in live fetuses of pregnant Sprague-Dawley rats exposed to n-
hexane via inhalation 60
Table 4-12. Total red blood cells and nucleated cells in bronchial lavage from n-hexane-
challenged New Zealand white rabbits 63
Table 4-13. Enzyme activities in lung homogenates of rabbits exposed to n-hexane 64
Table 4-14. Concentration of biochemicals and enzyme activities in bronchial lavage fluid from
male Sprague-Dawley rats exposed to n-hexane 64
Table 4-15. Changes in sciatic and sural nerve action potentials induced by n-hexane and 2,5-
hexanedione 67
Table 4-16. Incidence of liver and pituitary tumors in male and female B6C3F1 mice exposed to
commercial hexane for 2 years 70
Table 4-17. Time-to-onset for the appearance of axonal swelling in explanted cultures of fetal
mouse spinal cord incubated with mixtures of n-hexane and methyl ethyl ketone 76
Table 4-18. Effect of 2,5-hexanedione, acetone, ethanol, and mixtures of 2,5-hexanedione with
acetone or ethanol in drinking water on average MCV 79
Table 4-19. Pyrrole adduct formation in proteins from y-diketone-treated rats 87
Table 4-20. Summary of in vitro assays on the mutagenicity/genotoxicity of n-hexane 90
Table 4-21. Summary of in vivo assays on the mutagenicity/genotoxicity of n-hexane 91
Table 4-22. Summary of in vivo and in vitro assays on the mutagenicity/ genotoxicity of
commercial hexane mixtures 92
Table 4-23. Toxicity findings in inhalation studies for n-hexane 101
Table 5-1. BMD results of n-hexane inhalation toxicity studies for selection of the principal
study 118
Table B-l. Mean nerve conduction velocity in rats exposed to n-hexane B-3
VI
-------
Table B-2. Parameters and modeling results for the n-hexane exposure MCV data B-4
Table B-3. Mean total mixed nerve conduction velocity in rats exposed to n-hexane B-5
Table B-4. Mean total distal nerve conduction velocity in rats exposed to n-hexane B-5
Table B-5. Estimates of BMCL and goodness of fit statistics for total mixed nerve conduction
velocity B-7
Table B-6. Estimates of power model BMCL estimates and goodness of fit statistics for distal
mixed nerve conduction velocity B-8
Table B-7. BMCL estimates for power and quadratic models with various parameters B-9
Table B-8. Mean of litter mean pup body weight in rats exposed to n-hexane B-10
Table B-9. Results from fitting BMDS models to mean fetal weight data B-l 1
Table B-10. Mean reduced MCV in rats exposed to n-hexane for 8, 12, and 16 weeks .... B-12
Table B-l 1. Results of fitting the Hill function dose-response model to rat MCV data for 8, 12,
and 16-week exposures to n-hexane B-13
Table B-12. Benchmark dose modeling results of n-hexane inhalation toxicity studies for
selection of the principal study B-13
Figure 3-1. Biotransformation of n-hexane 9
Figure 3-2. Physiologically based toxicokinetic model of the distribution of n-hexane in the
body and the urinary excretion of 2,5-hexanedione 22
vn
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to n-hexane.
It is not intended to be a comprehensive treatise on the chemical or toxicological nature of
n-hexane.
In Section 6, Major Conclusions in the Characterization of Hazard and Dose Response,
EPA has characterized its overall confidence in the quantitative and qualitative aspects of hazard
and dose response by addressing knowledge gaps, uncertainties, quality of data, and scientific
controversies. The discussion is intended to convey the limitations of the assessment and to aid
and guide the risk assessor in the ensuing steps of the risk assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at 202-566-1676.
Vlll
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER AND AUTHORS
Jamie Benedict, Ph.D. (Chemical Manager)
National Center for Environmental Assessment
Office of Research and Development
George Holdsworth, Ph.D.
Oak Ridge Institute for Science and Education
Oak Ridge Associated Universities
Elizabeth Dupree Ellis, Ph.D.
Oak Ridge Institute for Science and Education
Oak Ridge Associated Universities
CONTRIBUTING AUTHOR
Allan Marcus
National Center for Environmental Assessment
Office of Research and Development
REVIEWERS
This document and the accompanying IRIS Summary have been peer reviewed by EPA
scientists and independent scientists external to EPA. Comments from all peer reviewers were
evaluated carefully and considered by the Agency during the fmalization of this assessment.
During the fmalization process, the IRIS Program Director achieved common understanding of
the assessment among the Office of Research and Development; Office of Air and Radiation;
Office of Prevention, Pesticides, and Toxic Substances; Office of Solid Waste and Emergency
Response; Office of Water; Office of Policy, Economics, and Innovation; Office of Children's
Health Protection; Office of Environmental Information, and EPA's regional offices.
INTERNAL EPA REVIEWERS
Philip Bushnell, Ph.D.
Neurotoxicology Division
National Health and Environmental Effects Research Laboratory
Office of Research and Development
Joyce M. Donohue, Ph.D.
Health and Ecological Criteria Division
Office of Science and Technology
Office of Water
Reeder Sams II, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
ix
-------
EXTERNAL PEER REVIEWERS
James J. Chen, Ph.D.
National Center for Toxicological Research
Food and Drug Administration
Lucio G. Costa, Ph.D.
Toxicology Program
Department of Environmental and Occupational Health Sciences
School of Public Health and Community Medicine
University of Washington
Hugh L. Evans, Ph.D.
Nelson Institute of Environmental Medicine
School of Medicine
New York University
Doyle G. Graham, M.D., Ph.D.
Independent consultant
Bernard Weiss, Ph.D.
Department of Environmental Medicine
School of Medicine and Dentistry
University of Rochester
Summaries of the external peer reviewers' comments and public comments and the
disposition of their recommendations are provided in Appendix A.
-------
LIST OF ABBREVIATIONS AND ACRONYMS
1SD
ACGIH
ADJ
AIC
API
ATSDR
BAEP
BEI
BMC
BMCL
BMD
BMDL
BMDS
BMR
BRRC
CA
CASRN
CHL
CHO
CI
CUT
CMAP
CNS
CYP450
DCV
DL
DOF
EMG
ENM
EP
EPA
EPL
ERG
EROD
FM
FOB
GD
HEC
HSDB
IRDC
IRIS
One standard deviation
American Conference of Governmental Industrial Hygienists
Adjusted for continuous exposure
Akaike's Information Criterion
American Petroleum Institute
Agency for Toxic Substance and Disease Registry
Brainstem auditory evoked potentials
Biological exposure index
Benchmark concentration
95% lower bound on the benchmark concentration
Benchmark dose
95% lower bound on the benchmark dose
Benchmark dose software
Benchmark response
Bushy Run Research Center
Chromosomal aberrations
Chemical Abstracts Service registry number
Chinese hamster lung
Chinese hamster ovary
Confidence interval
Chemical Industry Institute of Toxicology
Compound muscle action potential
Central nervous system
Cytochrome P450
Distribution of conduction velocity
Distal latency
Degree of freedom
Electromyography
Electroneuromyography
Evoked potentials
U.S. Environmental Protection Agency
Experimental Pathology Laboratories
Electroretinogram
Ethoxyresorufm O-deethylase
Farnsworth-Munsell
Functional observation battery
Gestation day
Human equivalent concentration
Hazardous Substances Data Bank
International Research and Development Corporation
Integrated Risk Information System
XI
-------
LDH
LOAEL
MAP
MCV
MDL
MN
NADPH
NCE
NSC
NCTR
NF
NL
NOAEL
NTP
OR
PBTK
PCE
PND
PNS
ppm
PROD
RfC
RfD
RvR
SCE
SCV
SD
SDS-PAGE
SEM
SEP
SNAP
TLV
TWA
UF
UPDRS
VEP
WBC
Median lethal concentration
Median lethal dose
Lactate dehydrogenase
Lowest-observed-adverse-effect level
Motor nerve action potential
Motor nerve conduction velocity
Minimum detection limit
Micronucleus
Reduced nicotinamide adenine dinucleotide phosphate
Nonchromatic erythrocytes
National Safety Council
National Center for Toxicological Research
Neurofilament
Nested logistic
No-observed-adverse-effect level
National Toxicology Program
Odds ratio
Physiologically based toxicokinetic
Polychromatic erythrocyte
Postnatal day
Peripheral nervous system
Parts per million
Pentoxyresorufin O-depentylase
Reference concentration
Reference dose
Rai and van Ryzin
Sister chromatid exchange
Sensory nerve conduction velocity
Standard deviation
Sodium dodecyl sulfate polyacrylamide gel electrophoresis
Standard error of the mean
Somatosensory evoked potential
Sensory nerve action potential
Threshold limit value
Time weighted average
Uncertainty factor
Unified Parkinson Disease Rating Scale
Visual evoked potential
White blood cell
xn
-------
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of n-hexane.
IRIS Summaries may include an oral reference dose (RfD), inhalation reference concentration
(RfC), and a carcinogenicity assessment.
The RfD and RfC provide quantitative information for noncancer dose-response
assessments. The toxicity values are based on the assumption that thresholds exist for certain
toxic effects such as cellular necrosis but may not exist for other toxic effects such as some
carcinogenic responses. In general, the RfD is an estimate (with uncertainty spanning perhaps an
order of magnitude) of a daily exposure to the human population (including sensitive subgroups)
that is likely to be without an appreciable risk of deleterious noncancer effects during a lifetime.
It is expressed in units of mg/kg-day. The inhalation RfC is analogous to the oral RfD but
provides a continuous inhalation exposure estimate. The inhalation RfC considers toxic effects
for both the respiratory system (portal-of-entry) and for effects peripheral to the respiratory
system (extrarespiratory or systemic effects). It is generally expressed in units of mg/m3.
The carcinogenicity assessment provides information on the carcinogenic hazard potential
of the substance in question and quantitative estimates of risk from oral and inhalation exposure.
The information includes a weight-of-evidence judgment of the likelihood that the agent is a
human carcinogen and the conditions under which the carcinogenic effects may be expressed.
Quantitative risk estimates are presented in three ways to better facilitate their use: (1) generally,
the slope factor is the result of application of a low-dose extrapolation procedure and is presented
as the risk per mg/kg-day of oral exposure; (2) the unit risk is the quantitative estimate in terms of
either risk per g/L drinking water or risk per g/m3 continuous airborne exposure; and (3) the 95%
lower bound and central estimate on the estimated concentration of the chemical substance in
drinking water or air presents cancer risks of 1 in 10,000, 1 in 100,000, or 1 in 1,000,000.
Development of these hazard identification and dose-response assessments for n-hexane
has followed the general guidelines for risk assessment as set forth by the National Research
Council (1983). The United States Environmental Protection Agency (EPA) guidelines that were
used in the development of this assessment includes the following: Guidelines for
Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Guidelines for Reproductive
Toxicity Risk Assessment (^JS. EPA, 1996), Guidelines for Neurotoxicity Risk Assessment (U.S.
EPA, 1998a), Guidelines for carcinogen risk assessment (U.S. EPA, 2005a), Supplementary
-------
Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens (U.S. EPA.,
2005b), Recommendations for and Documentation of Biological Values for Use in Risk
Assessment (U.S. EPA, 1988), Interim Policy for Particle Size and Limit Concentration Issues
in Inhalation Toxicity (U.S. EPA, 1994a), Methods for Derivation of Inhalation Reference
Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b), Use of the
Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 1995), Science Policy
Council Handbook: Peer Review (U.S. EPA, 2000a), Science Policy Council Handbook: Risk
Characterization (U.S. EPA, 2000b), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000c), and A Review of the Reference Dose and Reference Concentration Processes (U.S.
EPA, 2002).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service registry number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document.
-------
2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS
n-Hexane (CASRN 110-54-3) is a straight-chain, fully saturated hydrocarbon with six
carbon atoms also referred to as hexane and hexyl hydride. The chemical is also referred to by
trade names such as Skellysolve B and NCI-C60571. Some physical and chemical properties of
n-hexane are shown below (HSDB, 2005; ATSDR, 1999).
Chemical formula C6H14
Molecular weight 86.18
Melting point -95 °C
Boiling point 69 °C
Density 0.66 g/mL (at 20 °C)
Water solubility 9.5 mg/L (at 25 °C)
LogKow 3.29
LogKoc 2.9
Vapor pressure 150 mm Hg (at 25 °C)
Henry' s Law constant 1.69 atm-m3/mol
Conversion factor 1 ppm = 3.5 mg/m3; 1 mg/m3 = 0.28 ppm
(at 25 °C, 760 mm Hg)
n-Hexane is a solvent that has many uses in the chemical and food industries, either in
pure form or as a component of the mixture commercial hexane. Highly purified n-hexane is
primarily used as a reagent for chemical or chromatographic separations. Commercial hexane is
a mixture that contains approximately 52% n-hexane; the balance is made up of varying amounts
of structural isomers and related chemicals, such as methylpentane and methylcyclopentane.
Mixtures containing n-hexane are also used in processes for the extraction of edible fats
and oils in the food industry, as cleaning agents in textile and furniture manufacturing, and in the
printing industry. n-Hexane is the solvent base for many commercial products, such as glues,
cements, paint thinners, and degreasers (NSC, 2003; ATSDR, 1999).
The chemical is a minor constituent of crude oil and natural gas and, therefore, represents
a variable proportion of different petroleum distillates. For example, n-hexane comprises about
11.6% of unleaded gasoline and about 2% of JP-4 aviation fuel (ATSDR, 1993a, b).
-------
3. TOXICOKINETICS RELEVANT TO ASSESSMENTS
3.1. ABSORPTION
No oral exposure studies evaluating absorption of n-hexane in humans or laboratory
animals are available. However, absorption following oral exposure has been suggested by the
identification of n-hexane and its metabolites in expired air, serum, and urine (Baelum et al.,
1998; Ono et al., 1981; Krasavage et al., 1980). For example, increased levels of n-hexane in
exhaled air and a major metabolite of n-hexane (2,5-hexanedione) in urine were observed
following exposure of human volunteers to n-hexane (0.3 and 1.0 mg/minute for 60 minutes) by
a gastric feeding tube (Baelum et al., 1998). Krasavage et al. (1980) showed increased levels of
2,5-hexanedione in serum of rats exposed to n-hexane by gavage. In addition, neurotoxicity
observed following oral exposure of rats to n-hexane also suggests oral absorption of the
chemical (Ono et al., 1981; Krasavage et al., 1980).
There is also limited evidence in humans inferring absorption following inhalation
exposure to n-hexane. Mutti et al. (1984) measured n-hexane in the inhaled and expired air of 10
workers who routinely breathed in solvent vapors during their shift at a shoe factory. n-Hexane
concentrations in breathing zone air (8-hour time weighted average [TWA] median concentration
of 69 parts per million [ppm] or 243 mg/m3) were monitored with personal monitors. The
authors noted that other structural isomers and related chemicals of n-hexane (2-methylpentane,
3-methylpentane, cyclohexane, and n-heptane) were also present in the breathing air. Samples
(inhaled and exhaled air) were collected simultaneously for 5 minutes; the last 100 mL of the
tidal volume represented alveolar air. Alveolar uptake was determined by the following
equation:
D = kQ(TWA)VaRt
where D = alveolar uptake or dose (mg), k = factor converting ppm to mg/L (3.5 x 10"3), Ci(TWA)
= time weighted average of breathing zone levels of n-hexane (ppm), Va = alveolar ventilation, R
= pulmonary retention coefficient, and t = time (minutes). Alveolar retention (difference
between inhaled and alveolar concentrations of n-hexane) was approximately 25%. The authors
stated that the absorption rate, taking into account both retention and alveolar ventilation, was
approximately 17%. Further evidence for absorption following inhalation exposure of n-hexane
was suggested by the presence of metabolites of n-hexane in the urine, which was monitored at
the beginning and end of each shift, and 15 hours after exposure.
-------
Veulemans et al. (1982) studied the respiratory uptake and elimination of n-hexane in six
healthy male volunteers. Subjects were exposed at rest to 360 or 720 mg/m3 (102 or 204 ppm) of
n-hexane for 4 hours and to 360 mg/m3 (102 ppm) under various levels of exercise. A 2-week
rest period was allowed between experiments. Inhaled and exhaled air were monitored (for up to
4 hours after exposure), and blood samples were collected. The authors reported an approximate
22% and 24% retention of n-hexane from inhaled air (360 and 720 mg/m3, respectively) at rest.
The corresponding absorption rates were calculated as 0.84 and 1.59 mg/minute, respectively.
Absorption rates for n-hexane (360 mg/m3) also increased with increasing physical activity.
3.2. DISTRIBUTION
Following inhalation exposure, n-hexane is absorbed into the circulation and transported
to the liver, the major site of metabolism. In the liver, n-hexane is metabolized to various
metabolites that are then distributed in the blood to various organs and tissues, including the
liver, kidney, and brain. Several inhalation studies in humans and animals demonstrate the
distribution of n-hexane.
Perbellini et al. (1985) reported partition coefficient values (olive oil: air, human blood:
air, and human tissues: air) for n-hexane. The following values were presented: olive oil: air,
146; blood: air, 0.80; and 1.0, 2.8, 5.0, 5.0, 3.0, 5.2, and 104 for lung, heart, muscle, brain,
kidney, liver, and fat: air, respectively. These values are similar to those reported in some tissues
from F344 rats (2.9 for muscle, 5.2 for liver, and 159 for fat [Gargas et al., 1989]). In addition,
in vitro blood: air (2.13) and breast milk:air (4.66) partition coefficients for humans (eight
volunteers) were used to calculate a milk: blood partition coefficient of 2.10 (Fisher et al., 1997).
No studies investigating the distribution of n-hexane following oral exposure in humans
or laboratory animals are available. However, one human study evaluating the distribution of n-
hexane following inhalation exposure is available. Veulemans et al. (1982) measured blood
levels of n-hexane at various intervals for 4 hours after exposure (resting and during physical
activity) to n-hexane (102 and 204 ppm for 4 hours). Blood levels of n-hexane rapidly decreased
to approximately half the steady state exposure values within the first 10 minutes and reached
steady state levels by 100 minutes. The average half life for n-hexane in blood was 1.5-2 hours.
Studies in laboratory animals indicate that n-hexane is distributed to a variety of tissues
following inhalation exposure. Bus et al. (1979) monitored n-hexane in the blood of pregnant
F344 rats that were exposed via inhalation to 1000 ppm (3520 mg/m3) n-hexane (purity not
stated) for 6 hours on gestation day (GD) 20. n-Hexane was measured in maternal blood, liver,
-------
kidney, brain, and fetuses using gas chromatography-mass spectrometry at 0, 1, 2, 4, and 8 hours
after exposure. Maximum tissue concentrations of n-hexane were observed immediately after
cessation of exposure. Average concentrations of n-hexane in blood and tissues are presented in
Table 3-1.
Table 3-1. Concentration of n-hexane in blood and tissues of pregnant F344
rats immediately after a 6-hour exposure to 1000 ppm n-hexane
Tissue
Blood
Liver
Kidney
Brain
Fetus
Concentration (jig/g wet weight
or |j,g/mL)
0.45 ±0.11
0.85 ±0.13
6.33 ±0.75
0.04 ± 0.00
0.61 ±0.14
Source: Busetal., 1979.
Bus et al. (1981) administered a single exposure or five daily 6-hour exposures of 1000
ppm n-hexane to male F344 rats (three/group) and observed the appearance of n-hexane in
blood, liver, kidney, brain, and sciatic nerve. In the single dose experiment, concentrations of
n-hexane in blood, liver, kidney, brain, and sciatic nerve were 0.50 ±0.1 |ig/mL, and 1.23 ±
0.14, 5.8 ± 0.3, 3.0 ±0.11, and 46 ± 10 |ig/g, respectively. However, levels of n-hexane were
undetectable in blood after 1 hour and in liver, kidney, and brain after 4 hours.
Bus et al. (1982) also monitored the distribution of radioactivity in a range of tissues for
up to 72 hours after exposing male F344 rats to a single 6-hour exposure to either 0, 500, 1000,
3000, or 10,000 ppm n-hexane (95.5% pure) containing 11.8-54.9 |iCi/mmol [l,2-14C]-n-hexane
(Table 3-2).
Table 3-2. Tissue distribution of radioactivity in male F344 rats 72 hours after a 6-
hour inhalation exposure to various concentrations of [l,2-14C]-n-hexane
Tissue
Liver
Lung
Exposure (ppm)
500
1000
3000
10,000
Tissue concentration (nmol-equivalents/g wet weight or mL)
63.5 ±0.9
58.4 ±0.7
90.9 ±3.0
85.3 ±13.1
313.8 ±34.6
176.9 ±9.5
189.0 ±10.1
118.1±2.8
-------
Tissue
Kidney
Testes
Brain
Sciatic nerve
Blood
Exposure (ppm)
500
1000
3000
10,000
Tissue concentration (nmol-equivalents/g wet weight or mL)
69.2 ±3.4
33.5 ±0.9
24.6 ±0.7
53.0 ±1.9
27.8 ±1.3
88. 1± 6.6
48.6 ±1.4
33.1 ±2.2
84.8 ± 27.7
23.5 ±2.2
242.2 ±6.6
ND
ND
ND
ND
135.4 ±5.5
67.8 ±5.4
57.5 ±3.3
ND
79.0 ±7.3
ND = Not determined.
Source: Busetal., 1982.
The authors stated that since the n-hexane was labeled at the C-l and C-2 positions, it
was likely that the radioactivity found in the various tissues was due to incorporation by
metabolism of 2-hexanone, but it was possible that residual tissue radioactivity was due to
binding of reactive metabolites to macromolecules, which is supported by in vitro studies
showing 2,5-hexanedione forming Schiff bases and/or pyrrole derivatives with lysine amino
groups (Sanz et al., 1995; DeCaprio et al., 1988 , 1982; Boekelheide, 1987; Lapadula et al.,
1986; Anthony et al., 1983a, b; Graham et al., 1982b).
Similarly, after a single 6-hour exposure of n-hexane to male F344 rats (0, 500, 1000,
3000, or 10,000 ppm), n-hexane was detected in blood, liver, kidney, sciatic nerve, testis, brain,
and lung (Baker and Rickert, 1981) (Table 3-3).
Table 3-3. Apparent steady state concentration of n-hexane concentrations in male
F344 rats after 6 hours inhalation exposure to [l,2-14C]-n-hexane
Tissue
Liver
Lung
Kidney
Testes
Brain
Sciatic nerve
Blood
Exposure (ppm)
500
1000
3000
10,000
Tissue concentration (nmol-equivalents/g wet weight or mL)
2.6 ±0.9
3.1 ±0.2
7.0 ± 1.0
3.5 ±0.8
1.8 ±0.0
12.0 ±1.0
1.3 ±0.2
6.7 ±0.3
8.8 ±0.7
22.0 ±2.0
20.0 ±2.0
19.1 ±2.3
48.0 ±5.0
2.2 ±0.2
22.8 ±2.3
27.7 ±2.9
41.4 ±4.2
27.3 ±2.4
36.1 ±2.8
130.3 ±17.4
8.4 ±0.8
72.4 ±4.8
89.3 ±11.6
54.4 ±1.8
53.9 ±6.7
54.2 ±1.6
430.5 ±59.4
20.9 ±0.3
-------
Data are means ± standard error of the mean (SEM) (n = 3).
Source: Baker and Rickert, 1981.
3.3. METABOLISM
n-Hexane is principally metabolized in the liver. As shown in Figure 3-1, n-hexane is
initially hydroxylated by the action of mixed function oxidases to form either 1- or 3-hexanol in
a detoxification pathway or 2-hexanol in a bioactivation pathway. Through the bioactivation
pathway, 2-hexanol is converted to 2-hexanone and 2,5-hexanediol. Both of these metabolites
are then further metabolized to 5-hydroxy-2-hexanone, 2,5-hexanedione, and 4,5-dihydroxy-
2-hexanone. 2,5-Hexanedione is believed to be the major toxic metabolite produced in humans
following acid hydrolysis of urine samples prior to analysis by gas chromatography (Perbellini et
al., 1981).
Evidence that the liver is the primary location for the initial hydroxylation step for
bioactivation of n-hexane comes from the measurement of hydroxylating activity in isolated
microsomes from liver, lung, brain, and the extensor digitorum longus and soleus skeletal
muscles (Crosbie et al., 1994). Microsomes were incubated with n-hexane in the presence of
reduced nicotinamide adenine dinucleotide phosphate (NADPH) and the rates of production of
the primary alcohols, 1-, 2-, and 3-hexanol, were compared. Liver microsomes produced
significantly more 2-hexanol (1104 ± 205 pmol/minute-mg protein) than microsomes from lung,
brain, and soleus and extensor digitorum longus skeletal muscles (132 ± 25, 3 ± 2, 4 ± 1, 28 ± 5
pmol/minute-mg protein, respectively). Similarly, in liver microsomes, 2-hexanol production
occurred at a much faster rate than that of 1- or 3-hexanol. In contrast, the production of
1-hexanol appeared to occur at a much faster rate in microsomal preparations from the lung.
Metabolism of n-hexane in humans primarily forms 2,5-hexanedione. Perbellini et al.
(1981) identified this metabolite along with 2,5-dimethylfuran, y-valerolactone, and 2-hexanol in
the urine of 41 shoe workers exposed to 11-250 mg/m3 commercial hexane (a mixture
containing n-hexane). The mean concentrations, determined following acid extraction of the
urine, were 5.4 ± 4.9, 3.7 ± 4.1, 3.3 ± 2.7, and 0.19 ± 0.26 mg/L, respectively. n-Hexane
exposure correlated well both with total metabolites (r = 0.7858) and with 2-hexanol (r = 0.6851)
and 2,5-hexanedione (r = 0.6725) individually.
-------
0
Bioactivation
Pathway (predominates in liver)
(CYP450 dependent)
OH
2-Hexanol
)H
2,5-Hexanediol
OH
OH
5-Hydroxy-2-hexanone
Detoxification
Pathway (predominates in lung)
(CYP450 dependent)
beta-oxidation of fatty acids
r
2,5-Hexanedione
4,5-Dihydroxy-2-hexanone
D
2,5-Dimethylfuran
Figure 3-1. Biotransformation of n-hexane.
Source: Adapted from Soriano et al., 1996; Couri and Milks, 1982.
-------
The time course of n-hexane metabolism in humans was determined by van Engelen et al.
(1997). Volunteers were exposed to approximately 60 ppm n-hexane for 15.5 minute intervals in
the morning and afternoon. 2,5-hexanedione blood levels peaked between 16.2 and 19.8 minutes
after the start of exposure (no difference was found between morning and afternoon exposures).
Other studies indicate that 2,5-hexanedione levels identified in the urine of humans may
be an artifact of the extraction method used (dos Santos et al., 2002; Fedtke and Bolt, 1986).
Specifically, studies indicate that hydrolysis of urinary conjugates with acid may lead to the
conversion of 4,5-dihydroxy-2-hexanone to 2,5-hexanedione. It is also possible that acid
hydrolysis could result in the release of 2,5-hexanedione that is bound to protein amino groups
by a reversible imino bond. For example, dos Santos et al. (2002) observed increased levels
(approximately 10 times higher) of urinary 2,5-hexanedione in acidified urine samples from 52
Brazilian shoe workers exposed to n-hexane compared to samples from exposed workers that
were not subjected to acid hydrolysis (0.94 mg/L after acid hydrolysis versus 0.09 mg/L without
hydrolysis).
Fedtke and Bolt (1986) used the acidification pretreatment procedure to detect a small
amount of 2,5-hexanedione in the urine of 12 subjects who had not been exposed to n-hexane.
The range of values was 0.12-0.78 mg/L with an arithmetic mean of 0.45 ± 0.20 mg/L. The
authors speculated that the low levels of 2,5-hexanedione detected in the urine might be
explained by the metabolism of endogenously produced n-hexane rather than environmental
exposure. Studies in humans and laboratory animals have shown that rodents exhale alkanes as
metabolic products of lipid peroxidation of liver phospholipid fatty acids (Vaz and Coon, 1987;
Gelmont et al., 1981; Kivits et al., 1981). Fedtke and Bolt (1986) also investigated (in a single
urine sample) the effect of acid hydrolysis on formation of 2,5-hexanedione. A pH range of 3-7
had no effect on the amount of 2,5-hexanedione liberated, but a pH below 3 (down to 0.1)
increased the amount of 2,5-hexanedione released.
A study in rats suggested that 2-hexanol may be the major metabolite of n-hexane
following inhalation exposure. Fedtke and Bolt (1987) exposed three male Wistar rats/group to
mean concentrations of 0, 50 ± 3, 102 ± 6, 248 ± 6, 504 ± 20, 1003 ± 74, or 3074 ± 96 ppm
n-hexane for 8 hours. The authors collected urine samples during and after exposure (8, 16, 24,
32, 40, and 48 hours postexposure). The formation of n-hexane metabolites was dependent on
exposure concentrations up to approximately 300 ppm. Formation of 1-, 2-, and 3-hexanol and
2-hexanone was evident during exposure, but had ceased by 8 hours postexposure. Levels of
2,5-hexanedione and 4,5-dihydroxy-2-hexanone were initially low and the metabolism of
n-hexane to these metabolites had ceased by 16 and 40 hours postexposure, respectively. The
10
-------
primary metabolite formed in rats following inhalation exposure was 2-hexanol (approximately
twice all other metabolites observed), followed by 4,5-dihydroxy-2-hexanone. These
metabolites together accounted for about 90% of the total metabolites formed. The level of 4,5-
dihydroxy-2-hexanone was approximately 10 times higher than 2,5-hexanedione (calculated by
subtraction of the concentration of free 2,5-hexanedione measured without acid hydrolysis from
the concentration of 2,5-hexanedione measured following complete acid hydrolysis).
Krasavage et al. (1980) exposed male COBS, CD(SD) BR rats to single gavage doses of
0, 6.6, 13.2, and 46.2 mmol/kg n-hexane and evaluated peak 2,5-hexanedione levels in serum.
The peak serum concentrations of 2,5-hexanedione measured at each dose were 24, 44, and 53
|ig/mL, respectively.
Bus et al. (1979) observed the metabolism of n-hexane in the fetus and maternal tissues
of pregnant F344 rats that had been administered a single 6-hour exposure of 0 or 1000 ppm
n-hexane on GD 12 or 20. n-Hexane and its metabolites, 2,5-hexanedione and 2-hexanone, were
detected in the liver, kidney, brain, blood, and the developing fetus at time points up to 18 hours
after exposure. The metabolism of n-hexane to 2,5-hexanedione and 2-hexanone was rapid
(Table 3-4). 2,5-hexanedione was the major metabolite observed in all maternal tissues
evaluated and in the fetus. Levels reached peak concentrations in these tissues at 4 hours
postexposure.
Table 3-4. Metabolism of n-hexane following a 6-hour exposure of pregnant F344
rats on gestation day 20
Tissues
Time after exposure (hours)a
0
1
2
4
8
12
18
Maternal
Liver
n-Hexane
2-Hexanone
2,5-HD
Kidney
n-Hexane
2-Hexanone
2,5-HD
Brain
n-Hexane
0.85 ±0.13
0.16 ±0.03
0.56 ±0.03
6.33 ±0.75
1.04 ±0.31
1.29 ±0.10
0.04 ± 0.00
0.33 ±0.06
0.06 ± 0.00
0.75 ±0.04
3. 16 ±0.97
0.32 ±0.01
0.73 ±0.04
ND
0.15 ±0.04
0.03 ± 0.00
0.74 ± 0.04
1.15 ±0.15
0.14 ±0.02
1.22 ±0.24
ND
ND
0.03 ±0.01
1.30 ±0.30
0.67 ± 0.27
0.08 ±0.03
1.37 ±0.07
ND
ND
0.01 ±0.00
0.85 ± 0.08
ND
0.03 ±0.00
0.80 ±0.03
ND
ND
ND
0.36 ±0.06
ND
ND
0.24 ± 0.06
ND
ND
ND
0.11 ±0.02
ND
ND
0.07 ±0.03
ND
11
-------
Tissues
2-Hexanone
2,5-HD
Blood
n-Hexane
2-Hexanone
2,5-HD
Time after exposure (hours)a
0
0.69 ±0.13
2.41 ±0.30
0.45 ±0.11
0.70 ±0.10
1.06 ±0.27
1
0.29 ±0.02
1.79 ±0.10
0.30 ±0.05
0.30 ±0.05
0.93 ±0.12
2
0.11 ±0.01
3. 10 ±0.34
0.13 ±0.02
0.10 ±0.01
1.51 ±0.21
4
0.03 ±0.01
3.61 ±0.40
0.04 ±0.01
0.04 ±0.01
1.73 ±0.30
8
0.01 ±0.00
2.07 ±0.09
ND
0.01 ±0.00
0.74 ± 0.09
12
ND
0.29 ±0.01
ND
ND
0.33 ±0.03
18
ND
0.16 ±0.01
ND
ND
0.14 ±0.04
Fetus
n-Hexane
2-Hexanone
2,5-HD
0.61 ±0.14
0.51 ±0.08
1.17±0.15
0.31 ±0.12
0.18 ±0.00
0.97 ±0.16
ND
0.10 ±0.01
1.24 ±0.09
ND
0.03 ±0.00
1.67 ±0.16
ND
0.01 ±0.00
0.80 ±0.05
ND
ND
0.29 ±0.07
ND
ND
0.07 ±0.01
a Values are ug/mL or ug/g wet weight ± SEM.
ND = Not detected.
Source: Busetal., 1979.
The kinetics of the metabolism of n-hexane has also been investigated in vitro using
microsomal preparations from the liver and lung of male Sprague-Dawley rats (Toftgard et al.,
1986). The concentrations of the metabolic products formed and the reaction velocities were
determined. The kinetic data were plotted using an Eadie-Scatchard transformation. An Eadie-
Scatchard transformation is a plot of velocity/substrate concentration on the y-axis against
velocity on the x-axis. It is used to estimate the Km and Vmax for an enzyme. The estimated
parameters for n-hexane hydroxylation in the liver and lung are presented in Table 3-5.
The values for 1- and 2-hexanol suggested that a two-enzyme system is responsible for
the metabolism of n-hexane to these metabolites in liver tissue. The lower a Km value, the higher
the affinity of an enzyme for a substrate. The data indicate that one of the metabolic enzymes
has a high affinity for n-hexane as a substrate, while the other has a lower affinity. The
metabolite of greatest interest in the liver is 2-hexanol because of its conversion to
2,5-hexanedione, a lexicologically active metabolite. The enzyme represented by Kml in Table
3-5 with a Km of 6.0 jiM is primarily responsible for the production of 2-hexanol. The second
enzyme system (Km2) involved in the production of 2-hexanol has a Km of 1,100 jiM and thus a
far lower affinity for n-hexane than the first system. This suggests that the first system is likely
to play the major role in the production of 2-hexanol in the liver. The production of 1-hexanol in
the liver also appears to involve two enzymes with considerably different affinities for the
substrate. The enzyme represented by Kml with a Km of 0.4 jiM has a greater affinity for hexane
12
-------
than the enzyme represented by Km2 with a Km of 300 jiM.
Table 3-5. Apparent kinetic parameters for n-hexane hydroxylation in rat liver and
lung microsomes
Tissue parameter
Liver
K™, (UM)
Vmaxl (nmoles/mg-min)
K^ (uM)
Vmax2 (nmoles/mg-min)
Lung
K^uM)
Vmax(nmoles/mg-min)
Product formed
1-hexanol
0.4
0.09
300
1.2
9
2.2
2-hexanol
6
1
1100
4.6
50
1.3
3-hexanol
ND
ND
290
0.5
65
0.2
ND = No data.
Source: Toftgardetal., 1986.
The liver data for the production of 3-hexanol suggest that there is only one enzyme
involved in the metabolism of n-hexane to this product. The affinity of this enzyme for n-hexane
is similar to the low affinity enzyme system responsible for the production of 1-hexanol. The
authors concluded that there were at least four enzymes involved in the metabolism of n-hexane
to 1-, 2-, and 3-hexanol, in the liver but could not identify these enzymes from the kinetic data.
The Km and Vmax values indicate that 1- and 2-hexanol are the favored hydroxylation products in
the liver. The reaction requirement for NADPH suggests that these enzymes may be cytochrome
P450 (CYP450) isozymes.
The Eadie-Scatchard plots for lung microsomes suggest that a single enzyme is
responsible for the hydroxylation of n-hexane to 1-, 2-, and 3-hexanol in this tissue. The kinetic
parameters for each of the lung metabolites are presented in Table 3-5. Based on the low Km,
and accompanying Vmax , 1-hexanol is the favored product in the lungs. The enzymes
responsible for the formation of 2-hexanol and 3-hexanol have similar affinities for n-hexane.
CYP450 enzymes catalyze the initial steps (either detoxification or bioactivation)
involving hydroxylation in the metabolism of n-hexane. Specifically, the enzymes responsible
for the metabolism of n-hexane have been investigated in vivo. Nakajima et al. (1991)
characterized the CYP450 enzymes that are induced following exposure to n-hexane in male
13
-------
Wistar rats. The authors used phenobarbital, n-hexane, 2-hexanone, and 2,5-hexanedione to
induce different CYP450s to which they also raised monoclonal antibodies. The enzyme
activities of the CYP450 isozymes 2E1, 2C2/6, 1A1/2, and 2B1/2 were measured indirectly by
benzene aromatic hydroxylase activity, toluene side chain oxidation, ethoxyresorufm
O-deethylase (EROD) activity, and pentoxyresorufin O-depentylase (PROD) activities,
respectively. There was increased activity of benzene aromatic hydroxylase in liver microsomes
from n-hexane-treated rats, indicating the induction of CYP2E1. Conversely, there was no
increase in PROD or EROD activities in microsomal preparations from n-hexane-treated rats
compared to control preparations, indicating that n-hexane did not specifically induce CYP2A1/2
or CYP2B1/2. 2,5-hexanedione induced CYP2E1 and, to some extent, CYP2B1/2, suggesting
that more than one CYP450 species may be involved in the overall conversion of n-hexane to its
metabolic products. n-Hexane and 2-hexanone increased CYP2E1 to a similar extent when
measured in an immunoinhibition assay of toluene side-chain oxidation. In addition, 2-hexanone
induced CYP2B1/2 to a lesser extent than phenobarbital treatment.
Iba et al. (2000) demonstrated in in vivo studies that CYP2E1 may be involved in the
metabolism of n-hexane to 2,5-hexanedione. CYP2E1 knockout mice and control mice were
administered daily intraperitoneal injections of n-hexane (200 mg/kg) for up to 21 days.
CYP2E1 knockout and control mice had similar urinary levels of 2,5-hexanedione on day 10 of
administration (6.1 and 4.3 |ig/mL in the CYP2E1 knockout and control mice, respectively).
Levels of 2,5-hexanedione continued to increase in control mice on days 14 and 21 (22.9 and
16.1 |ig/mL) but not in CYP2E1 knockout mice. These data indicate that CYP2E1 may be
involved in the metabolism of n-hexane to 2,5-hexanedione following prolonged daily
exposures.
In vitro studies also indicate the involvement of CYP450 enzymes in n-hexane
metabolism (specifically the CYP2B1 enzyme), primarily leading to the formation of 2- and
3-hexanol (bioactivation pathway). Toftgard et al. (1986) evaluated the role of the
phenobarbital-inducible CYP450 isozymes (CYP2B1 and 2B2) and p-naphthoflavone-inducible
CYP450 isozyme (CYP1A1) in n-hexane hydroxylation. Specific isozyme preparations were
isolated from rat livers after exposure to the appropriate inducer and the isozymes involved in
the production of hexanols were identified. Production of 1-, 2-, and 3-hexanol was measured
relative to time and expressed as nmol metabolite/minute-nmol enzyme protein (turnover
number). The enriched CYP2B1 preparation produced 2-, 3-, and 1-hexanol with a turnover
ratio of approximately 30:10:1. The CYP2B2 turnover ratio was similar to CYP2B1 (20:8:1 for
2-, 3-, and, 1- hexanol, respectively). Turnover numbers were highest for CYP2B1 and lowest
14
-------
for CYP2A1. CYP2B1/2 primarily produced 2-hexanol, whereas CYP2A1 primarily produced
3-hexanol. All three isozymes had low turnover numbers for production of 1-hexanol.
Although Toftgard et al. (1986) did not unequivocally demonstrate the identities and
numbers of the CYP450 isozymes involved in the hydroxylation of n-hexane in the liver, the
isozyme turnover data in combination with the kinetic data presented above support the
hypothesis that 2-hexanol is the primary n-hexane metabolite in rat liver. In addition, Toftgard
et al. (1986) used antibodies to inhibit these isozymes to evaluate the role of each isozyme in
n-hexane metabolism. Anti-CYP2Bl inhibited the formation of 2- and 3-hexanol but not
1-hexanol. Anti-CYPlAl had little antagonistic effect on the formation of any of the metabolic
products.
Crosbie et al. (1997) used metyrapone, a specific inhibitor of CYP2B1/2 to monitor the
appearance of 1-, 2-, and 3-hexanol in liver and lung microsomes from male Wistar rats that
were incubated with n-hexane. The inhibitor did not affect the amounts of 1-hexanol produced
by liver and lung microsomes, but 2-hexanol levels were reduced by 33% in liver microsomes
and by 74% in lung microsomes. 3-hexanol levels were reduced by 31% in liver microsomes
(not statistically significant compared to preparations without inhibitor) and by 92% in lung
microsomes. This near complete abolition of the 3-hexanol-producing activity of n-hexane-
incubated lung microsomes suggests that CYP2B1 is important for n-hexane hydroxylation and
detoxification in this organ. The partial reductions of 2-hexanol production in metyrapone-
treated liver and lung microsomes suggests the partial involvement of CYP2B1 in hydroxylation
for bioactivation.
Several studies have demonstrated that the presence of other chemicals may affect the
metabolism of n-hexane. Van Engelen et al. (1997) examined the effects of coexposure to
methyl ethyl ketone on the toxicokinetics of n-hexane in human volunteers. Subjects (four to
five persons) were exposed to 60 ppm n-hexane for 15.5 minutes with or without coexposure to
200 or 300 ppm methyl ethyl ketone. Each subject served as their own control by being exposed
on the same day to n-hexane or methyl ethyl ketone alone and to the mixture of both solvent
vapors. Methyl ethyl ketone had no effect on the concentration-time course for exhaled
n-hexane, but the concentration-time course for 2,5-hexanedione appearance in serum was
threefold lower after coexposure to methyl ethyl ketone. The authors suggested that one of the
intermediate steps in the conversion of n-hexane to 2,5-hexanedione may have been inhibited by
coexposure with methyl ethyl ketone.
Studies in laboratory animals also demonstrate the effect of coexposure to other solvents
on n-hexane metabolism. Robertson et al. (1989) demonstrated that 1.87 mL/kg methyl ethyl
15
-------
ketone given by gavage to male F344 rats 4 days prior to a single 6-hour inhalation exposure to
1000 ppm n-hexane increased the concentration of 2,5-hexanedione in blood, sciatic nerve, and
testis up to 10-fold. 2,5-Dimethylfuran was also detected in increased quantities as a result of
coexposure to methyl ethyl ketone.
Shibata et al. (1990a) monitored the appearance of 2-hexanol, 2,5-hexanedione, and
2,5-dimethylfuran in the urine for up to 48 hours after the start of a single 8-hour exposure of six
male Wistar rats/group to either 2000 ppm n-hexane alone or 2000 ppm n-hexane containing
either 200, 630, or 2000 ppm methyl ethyl ketone. Both total (free and conjugated) and free
n-hexane metabolite levels were decreased in urine of rats exposed to a mixture of n-hexane and
methyl ethyl ketone (significant in the 2000 ppm n-hexane plus 2000 ppm methyl ethyl ketone
exposure group).
Shibata et al. (1990b) also demonstrated lower concentrations of 2,5-hexanedione and
2-hexanone in the serum of rats coexposed to 2000 ppm n-hexane and 2000 ppm methyl ethyl
ketone compared to rats receiving n-hexane alone. The area under the serum concentration
curve for 2,5-hexanedione was 109.35 |ig/mL/24 hours in rats exposed to n-hexane alone
compared to 23.7 |ig/mL/24 hours in rats coexposed to n-hexane and methyl ethyl ketone.
Iwata et al. (1983) treated five male Wistar rats/group with a single 8-hour inhalation
exposure of either 1000 ppm n-hexane, 1000 ppm n-hexane plus 1000 ppm toluene, 1000 ppm
n-hexane plus 1000 ppm methyl ethyl ketone, or fresh air. The authors evaluated the levels of
2,5-dimethylfuran, 2-hexanone, 2-hexanol, 2,5-hexanedione, and y-valerolactone in urine
(samples were acid-hydrolyzed) following exposure to either n-hexane alone or to solvent
mixtures. The total concentrations of metabolites decreased by approximately one-sixth
following coexposure to n-hexane and toluene and one-fourth following coexposure to n-hexane
and methyl ethyl ketone.
In a more recent study, Cardona et al. (1996) reported the effects of acetone on n-hexane
metabolism and elimination. These authors analyzed the relationship between exposure to these
solvents and the concentrations of free and total (samples were acid-hydrolyzed)
2,5-hexanedione in the urine. Environmental monitoring and urinary samples were obtained
from 87 workers in the shoe industry in Spain. Environmental concentrations of n-hexane,
toluene, and acetone were monitored at each subject's workplace for the final 2-4 hours
(averaging 200 minutes) of a work shift. Urine and alveolar (exhaled) air samples were collected
15 minutes after the end of the work shift. The median concentrations of n-hexane, toluene, and
acetone in the workplace were 47 mg/m3 (range of 4-652 mg/m3), 57 mg/m3 (range of 12-683
mg/m3), and 109 mg/m3 (70 cases evaluated; range of 1-1826 mg/m3), respectively. The level of
16
-------
free 2,5-hexanedione in the urine of exposed workers was about 12% of total urinary
2,5-hexanedione. Total urinary 2,5-hexanedione concentration was significantly correlated
(p<0.001) with environmental n-hexane exposure (r = 0.936) and n-hexane in exhaled air (r =
0.7435). Acetone was statistically significantly correlated (p<0.001) with the ratios of total or
free 2,5-hexanedione and atmospheric n-hexane concentrations (r = 0.6459 and 0.6965,
respectively). In addition, there was significant correlation (p<0.001; r = 0.96626 and 0.94217,
respectively) between total and free 2,5-hexanedione concentrations and n-hexane and acetone
environmental exposures, cutaneous absorption (glove use), interaction of n-hexane and acetone,
and day of the week.
3.4. ELIMINATION
A single study in humans suggests elimination following oral exposure to n-hexane.
Specifically, Baelum et al. (1998) collected urine from human volunteers immediately following
exposure to n-hexane via a gastric feeding tube and at 1, 2, 3, and 4.5 hours postexposure. These
samples contained 2,5-hexanedione at a mean concentration of 0.22 ± 0.10 |_imols/L. No oral
exposure studies in laboratory animals are available indicating elimination of n-hexane.
Several human inhalation studies have provided evidence for the elimination of n-hexane
and metabolites following occupational and voluntary exposures to n-hexane. Imbriani et al.
(1984) measured the amount of parent n-hexane in the urine of 30 shoe workers who were
exposed to n-hexane in the workplace. The employees wore personal samplers that provided
estimates of the amount of n-hexane in the air that ranged from 13 to 197 mg/m3 (3.7-56 ppm).
The median concentration of n-hexane in the urine was 4.8 |ig/L, with an overall correlation
coefficient of 0.84 for the 30 subjects.
Mutti et al. (1984) monitored 10 workers exposed to n-hexane in a shoe factory (8 hour
TWA of 243 ppm n-hexane). Alveolar excretion of n-hexane accounted for 10% of the total
uptake. Among the metabolites eliminated in urine (samples were collected pre-shift, end of
shift, and the next morning and metabolites were measured following acid hydrolysis) were
2,5-hexanedione, 2,5-dimethylfuran, 2-hexanol, and y-valerolactone. The authors indicated that
end of shift 2,5-hexanedione levels were the best estimate of n-hexane exposure. Urinary
2,5-hexanedione excretion of approximately 3 mg/g creatinine was considered to indicate
exposure to 50 ppm n-hexane.
In a follow-up study, Mutti et al. (1993) observed a weaker correlation between n-hexane
exposure in the workplace and the amount of 2,5-hexanedione in the urine of exposed
individuals. The study authors indicated that 2,5-hexanedione levels may build up during the
17
-------
course of a workweek; therefore, urinary levels may not consistently reflect the ambient
n-hexane exposure concentration.
Several other human studies support the assertion that 2,5-hexanedione levels in urine are
the best estimate of n-hexane exposure in the workplace. For example, Ahonen and Schimberg
(1988) documented 2,5-hexanedione excretion in the urine of four healthy female shoe workers
who were exposed to varying amounts of n-hexane, along with acetone, toluene, and other
solvents. n-Hexane and other organic solvent concentrations were measured in the breathing
zone of the workers. Urine samples were collected from each worker during the experiment and
the following weekend. Regression data between the 8-hour TWA concentration of n-hexane in
the air and urinary 2,5-hexanedione for the three most heavily exposed workers at the different
sampling times indicated that a 180 mg/m3 (50 ppm) 8-hour TWA concentration of n-hexane in
air would result in a urinary 2,5-hexanedione concentration of 10 ± 3 |imol/L.
Saito et al. (1991) correlated the amount of 2,5-hexanedione in urine with exposure of 50
individuals to n-hexane at various concentrations. Saito et al. (1991) performed acid hydrolysis
to ensure that any conjugated urinary metabolites of 2,5-hexanedione, such as 4,5-dihydroxy-
2-hexanone, were converted to 2,5-hexanedione prior to assay, giving a total value for the
subject metabolite. Urinary concentrations of 2,5-hexanedione measured in this manner showed
a good correlation with exposure to n-hexane (r = 0.973).
Cardona et al. (1993) analyzed working conditions and environmental exposure to
solvents in 27 shoe factories in Italy and Spain and measured end-of-shift total 2,5-hexanedione
concentrations in urine. They reported that urinary concentrations of 2,5-hexanedione tended to
increase during the workweek, although a significant linear correlation was obtained between
mean environmental concentrations of n-hexane and urinary concentrations of the metabolite.
Concentrations of 2,5-hexanedione in end-of-shift urine ranged from 0.2 to 24.2 mg/L, with an
arithmetic mean of 6.3 ± 4.9 mg/L. Variability in the correlation was thought to have been due
to differing practices among the subjects in the use of protective clothing and rubber gloves.
Percutaneous absorption of n-hexane was thought to have occurred in some cases.
Mayan et al. (2001) analyzed urine samples from 45 Portuguese shoe workers for total
2,5-hexanedione and correlated these values with measured amounts of n-hexane in workplace
air. The urine samples, which were collected 1 hour before the end of the shift, had a geometric
mean 2,5-hexanedione concentration of 2.68 mg/g creatinine. The individual values ranged from
0.6 to 8.5 mg/g creatinine and correlated (r = 0.85) with personal air sample n-hexane
concentrations ranging from 6 to 70 ppm.
A similar study by the same research group in 111 shoe workers showed a positive
18
-------
correlation between workplace n-hexane concentrations ranging from 5-70 ppm and total
2,5-hexanedione concentrations in urine of 0.12-14.25 mg/g creatinine (Mayan et al., 2002).
Dos Santos et al. (2002) evaluated the amounts of free and total 2,5-hexanedione in the
urine of 52 Brazilian shoe workers and categorized the subjects according to the mean
concentrations of n-hexane to which they were exposed during the course of their work. No
numerical exposure data for n-hexane were provided in the report, but subjects (14) who applied
glue with a paintbrush had higher concentrations of 2,5-hexanedione in their urine than
individuals who used a glue handgun (total 2,5-hexanedione 1.5 versus 0.7 mg/L). Eleven
subjects who worked under a fume hood had the lowest urinary 2,5-hexanedione concentrations
(0.08 mg/L).
Prieto et al. (2003) monitored free and total 2,5-hexanedione and 4,5-dihydroxy-
2-hexanone in the urine of 132 Spanish shoe workers who were exposed to n-hexane ranging
from 4-709 mg/m3 (1-200 ppm). Most subjects were exposed to other solvents during the
course of their work, such as toluene, methyl ethyl ketone, other hexane isomers, heptane,
acetone, and ethyl acetate. The amounts of total urinary 2,5-hexanedione gave the best
correlation with levels of exposure to n-hexane (r = 0.91). Concentrations of the metabolite
ranged from 0.3-32.46 mg/L.
Studies in animals are also available, suggesting the elimination of n-hexane metabolites
in urine following exposure to n-hexane via inhalation. Bus et al. (1982) exposed F344 rats
(three/group) to a single 6-hour exposure of either 500, 1000, 3000, or 10,000 ppm n-hexane
(95.5% pure) containing 11.8-54.9 |iCi/mmol [l,2-14C]-n-hexane. More than 50% of the
recovered radioactivity was expired as 14CO2 or excreted in the urine. Similarly, Baker and
Rickert (1981) administered a single 6-hour inhalation exposure of 500, 1000, 3000, or 10,000
ppm n-hexane to male F344 rats. Urinary elimination of metabolites during the 72-hour period
following exposure included 2-hexanone, 2,5-hexanedione, 5-hydroxy-2-hexanone, 2-hexanol,
and dimethylfuran. The total amounts of dimethylfuran and 2,5-hexanedione were higher in
urine samples that were acid-hydrolyzed compared with untreated urine (Table 3-6). This
change was probably caused by the dehydration or hydrolysis of conjugated metabolites to
dimethylfuran or 2,5-hexanedione under the acidic conditions (Fedtke and Bolt, 1986).
19
-------
Table 3-6. Metabolites excreted
inhalation exposure to n-hexane
in urine during a 72-hour period following
in male F344 rats
Exposure
concentration
(ppm)
500
1000
3000
Metabolite
2-Hexanone
Dimethylfuran
2,5-Hexanedione
5-Hydroxy-2-hexanone
2-Hexanol
2-Hexanone
Dimethylfuran
2,5-Hexanedione
5-Hydroxy-2-hexanone
2-Hexanol
2-Hexanone
Dimethylfuran
2,5-Hexanedione
5-Hydroxy-2-hexanone
2-Hexanol
Total metabolites formed (ug)a
No treatment
0.4 ±0.1
7.0 ±4.0
4.0 ±1.0
3.3 ±0.7
ND
1.3 ±0.2
1.2 ±0.2
3. 3 ±0.2
2.9 ±0.01
ND
8.6 ±0.3
17.0 ±2.0
44.4 ±0.5
41.5 ±0.8
1.3 ±0.2
Hydrolysis with
P-glucuronidase
0.8 ±0.0
14.3 ±0.5
3.8 ±0.2
0.5 ±0.1
1.5 ±0.3
1.3±0.1
86.2 ±5.1
4.8 ±0.6
33.0 ±2.2
0.6 ±0.1
15.1 ±1.0
357.4 ±49.2
50.3 ±4.3
38.1 ±2.4
9.0 ±3.1
Hydrolysis
with 3N HC1
1.5 ±0.2
162.0 ±2.0
9.9 ±0.2
2.3 ±0.0
ND
4.0 ±0.1
194.3 ±60.2
72.4 ±6.1
5.8 ±0.9
ND
10.1 ±2.4
879.3 ±231.1
222.4 ±21.0
45.4 ±2.3
3.5 ±0.2
a Values are means for three animals ± SEM.
ND = Not detected.
Source: Baker and Rickert, 1981.
Frontal! et al. (1981) exposed Sprague-Dawley rats (six to nine/group) to 500, 1000,
2500, or 5000 ppm n-hexane, 9-10 hours/day, 5 days/week for up to 30 weeks. Some animals
were transferred to metabolic cages after exposure to permit the collection of overnight urine
samples. Metabolites of n-hexane identified in P-glucuronidase and acid hydrolyzed urine
samples included 2,5-dimethylfuran, y-valerolactone, 3-hexanol, 2-hexanol, and
2,5-hexanedione.
Perbellini et al. (1982) studied the metabolic interaction between n-hexane and toluene in
vivo in six male Wistar rats/group following intraperitoneal administration of either 200 mg/kg
20
-------
n-hexane, 200 mg/kg n-hexane plus 200 mg/kg toluene, or 200 mg/kg toluene alone. Amounts
of n-hexane metabolites obtained in 24-hour urine samples were lower in animals receiving
n-hexane mixed with toluene (Table 3-7).
Table 3-7. n-Hexane metabolite levels in urine of Wistar rats coexposed to
n-hexane and toluene
Metabolite
2-Hexanol
2,5-Hexanedione
2,5-Dimethylfuran
y-Valerolactone
Treatment
n-hexane (200 mg/kg)a
230.42 ± 145.83
138.97 ±58.91
91.57 ±35. 83
47.67 ±20.71
n-hexane (200 mg/kg) and toluene
(200 mg/kg)a
88.74 ± 54. 16b
72.93 ±41.06b
32.50 ±11.94C
16.70 ±11.56C
aData are means ± SD.
Significantly different from n-hexane: bp<0.005, cp<0.0l, as calculated by the authors.
Source: Perbellini et al., 1982.
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
A series of reports by Perbellini and coworkers described the construction of a
physiologically based toxicokinetic (PBTK) model for the distribution of n-hexane and its
metabolites among eight functionally described compartments of the body (Perebellini et al.,
1990, 1986, 1985). The compartments represent (1) the site of intake (lungs), (2) highly
vascularized tissues, (3) a muscle group, (4) a fat group, (5) a metabolizing center, and (6) three
other compartments important in the kinetics of metabolism (biotransformation, water, and
urinary compartments). The scheme, displayed in Figure 3-2, shows the catabolism of n-hexane,
with the production of 2,5-hexanedione and its subsequent transfer to the water and urinary
compartments. It is assumed that the chemical instantly establishes a balance between alveolar
air and venous blood and that the chemical is in equilibrium with each tissue compartment.
However, in the model the liver is theoretically considered to be the only site at which
metabolites of n-hexane are formed, and all rate constants are assumed to be first order
(Perbellini et al., 1986).
21
-------
Q(dv)
C(inh)\
I
r~ •
r
(
Lung Compartment
Vessel Rich Group
Muscle Group
Fat Group
Liver Compartments
KlandK2
,K3
Water Compartment
VK4
Urine Compartment
Q(alv)
1
C(exh)
Qptver)
Figure 3-2. Physiologically based toxicokinetic model of the distribution of
n-hexane in the body and the urinary excretion of 2,5-hexanedione.
Source: Perbellini et al., 1986.
22
-------
Perbellini and coworkers obtained much of their data on tissue volumes and blood flows
from the scientific literature (Perbellini et al., 1985; Mapleson, 1973). This information was
combined with experimental data on the partition of n-hexane between air and various media and
extracts of human tissue obtained at autopsy (Perbellini et al., 1985). At the core of the PBPK
model was a series of differential equations that described: (1) the concentration of n-hexane
reaching the lungs in venous blood; (2) the rate of change of n-hexane in the pulmonary
compartment and the rate of change of the arterial blood n-hexane concentration; (3) the rate of
change of n-hexane in the liver and the resulting amount of hepatic 2,5-hexanedione; (4) the
amount of 2,5-hexanedione in the water compartment; and (5) the rate of change of
2,5-hexanedione in the urine. Solving the set of differential equations simultaneously for various
hypothetical exposure scenarios allowed the model to be tested against experimental data on
n-hexane and 2,5-hexanedione concentrations that had been reported for human volunteers by
Veulemans et al. (1982). Key findings from this comparison were that when modeling
hypothetical human exposures (approximately 100 ppm for 4 hours) the resulting n-hexane
concentration in the venous blood (176 |-ig/L) was similar to the measured concentration in
human volunteers (207 ± 32 |-ig/L) (Veulemans et al., 1982). The concentration of n-hexane
declined rapidly at cessation of exposure, with the model showing close agreement to the charted
experimental data (Perbellini et al., 1986). The latter was best described by the regression
equation:
Cv(mg/L) = Ae-bt
with values for the constants of 0.114 (mg/L) for A, and 0.0074 (min)"1 for b. The half-life for
n-hexane in the blood was 94 minutes (Veulemans et al., 1982).
The key utility of the model was to simulate occupational conditions that had been found
in factories where n-hexane-containing products are used and where exposure is constant. When
the hypothetical exposure duration was extended to 8 hours, the concentration of n-hexane in the
fat compartment was shown to follow an upward trend, though with broad fluctuations,
representing the interval between shifts. Similarly, the concentration of 2,5-hexanedione in urine
displayed peaks and troughs without dropping to negligible levels until a time point after the last
exposure representing the latter part of the weekend. Perbellini et al. (1990) drew attention to
the persistence of n-hexane in the fat compartment and reported a half life of 64 hours in this
tissue group. This half-life suggests that accumulated n-hexane in fat could not be completely
excreted by the start of the following workweek, and that near to complete excretion of n-hexane
23
-------
in fat would require more than 10 days of no further exposure.
Perbellini et al. (1990) used their approach to evaluate the likely impact on the biological
exposure index (BEI) of the 50 ppm threshold limit value (TLV) proposed for n-hexane for
1988-1989 by the American Conference of Governmental Industrial Hygienists (ACGIH, 2003).
As simulated by the model, urinary concentrations of 2,5-hexanedione ranged from 2.4 to
2.9 mg/L before the start of the first shift of the work week and from 3.3 to 4.3 mg/L on the
morning of the following work days. Fisher et al. (1997) used a generic human lactation PBTK
model that was developed using published human and animal PBTK model parameters to
simulate the transfer of 19 volatile chemicals, including n-hexane, from a nursing mother to her
infant during breast feeding. The model was used to estimate the amount of chemical that would
be transferred during a given nursing schedule, assuming resumed occupational exposure after
childbirth and maternity leave. Specifically, the five-compartment model of Ramsey and
Andersen (1984) was adapted by the incorporation of a milk compartment that changed in
volume in response to a nursing infant. For n-hexane, rodent tissue solubility and allometrically
scaled metabolic rate constants available in the published literature were used to estimate human
tissue metabolic parameters for the model. Blood:air and milk:air partition coefficients were
determined by running the model for a simulated maternal exposure at the TLV of 50 ppm
n-hexane. This simulation predicted the amount of chemical that would be ingested by an infant
over a 24-hour period. The amount of n-hexane ingested by an infant was 0.052 mg (Fisher et
al., 1997). The Fisher et al. (1997) model does not specifically address target tissues or
extrapolate between species or routes and has not been validated. n-Hexane levels in breast milk
have not been quantified for measured exposures to n-hexane. The authors suggested that the
absence of exposure and toxicokinetic data on lactational transfer of chemicals such as n-hexane
to nursing infants is a disadvantage of this model.
24
-------
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY AND CASE REPORTS
4.1.1. Oral Exposure
No studies were identified that address the toxic effects of n-hexane in humans via the
oral route.
4.1.2. Inhalation Exposure
4.1.2.1. Subchronic Exposure
No subchronic exposure studies were identified that addressed the toxic effects of
n-hexane in humans via the inhalation route.
4.1.2.2. Chronic Exposure
Beall et al. (2001) conducted a nested case control study evaluating the relationship
between the occurrence of intracranial tumors among employees at a petrochemical plant and
exposure to agents, including ionizing radiation, methylene chloride, acrylonitrile, vinyl
chloride, formaldehyde, n-hexane, and various other chlorinated, halogenated, volatile, and
aromatic hydrocarbons and nitroso compounds. The workers were also exposed to
organometallic and elemental metallic catalysts. The study authors selected subjects from
approximately 2595 plant workers. The workers were mailed questionnaires that evaluated work
history in the plant, and a total of 12 cases of intracranial tumors was identified that had been
diagnosed among respondents after they had been hired at the plant. All cases were confirmed
by review of medical records and pathology specimens by four neuropathologists. Six of these
cases, all of which were men, had primary brain cancers or gliomas (two astrocytomas, two
oligodendrogliomas, and two glioblastomas ). Six cases had benign intracranial tumors, of
which two were diagnosed as vestibular schwannomas (observed in one man and one woman),
two as meningiomas (both in men), and two pituitary adenomas (observed in one man and one
woman). Ten healthy controls were matched to each case by gender, birth year (± 2 years), race,
and a start date for work in the building complex that preceded the tumor diagnosis date for the
matched case. The median length of employment at the facility was 16.8 years for cases and
10.9 years for controls.
25
-------
Work histories were obtained from company records or interviews, the latter providing
information about complete work history, exposures encountered, extent of hands-on work at
each job, and incidence of certain other nonoccupational factors that may be related to risk of
occurrence of brain cancers and intracranial tumors (exposure to diagnostic irradiation, use of
anticonvulsant and ototoxic drugs, history of head trauma, seizures, meningitis, use of cellular
phones and radiation badges, amateur radio operation, pesticide application, furniture
refmishing, and history of hearing loss). Exposure information was obtained from company
accounting records that detailed hours worked on projects during each year of employment and
self-reported workplace exposure to chemicals of interest. The authors compared cases and
controls with respect to self-reported exposure to chemicals of interest, project-based work
histories indicating the potential use of chemicals of interest, and self-reported exposure to any
of the other nonoccupational factors that may be related to the risk of brain cancers. Conditional
regression was used and maximum likelihood estimates of odds ratios (ORs) with a 95%
confidence interval (CI) were reported.
The authors showed that the OR for self-reported exposure to n-hexane was statistically
significantly elevated (OR, infinity), with a CI of 1.4 to infinity (6 cases and 26 controls
evaluated) for gliomas. The OR for potential exposure to n-hexane based on job-related
exposure estimates was 2.3 (CI, 0.4-13.7; four cases and 26 controls evaluated) for gliomas.
Analyses by duration indicated a statistically significantly elevated OR of 16.2 (CI, 1.1-227.6;
two cases and two controls evaluated) for potential long-term exposure to n-hexane (>48
months) for gliomas. No relationship was found between exposure to n-hexane and the
occurrence of intracranial tumors.
Sanagi et al. (1980) compared peripheral nervous system (PNS) function in n-hexane-
exposed and unexposed workers. Fourteen employees working in the mixing and drying jobs
(during the study period) at a factory producing tungsten carbide alloy for 1-12 years (average of
6.2 years) served as the n-hexane exposed group. In addition, a group of five workers who had
been mixers in the past (exposure for 1-16 years with an average of 5.2 years), but who were not
engaged in these jobs at the time of the study, were classified as exposed in the past. Fourteen
workers from the same factory who were not exposed to any solvents served as controls in this
study. All subjects were males under 50 years of age and free of metabolic diseases and lifestyle
factors that may affect the PNS. Twenty-two breathing zone monitoring samples taken twice a
year over a 2-year period indicated an 8-hour TWA of 58 ppm for n-hexane and 39 ppm for
acetone. No other solvent vapors were detected. Medical examinations consisted of interviews,
questionnaires, clinical neurological examinations, and neurophysiological testing. The
26
-------
questionnaire was comprised of 23 questions concerning neurological symptoms. Three
questions regarding hearing deficit, vision disturbance, and writing deficit were answered with
either present or absent. The remaining 20 questions were answered with always, sometimes, or
absent. The neurological exams and neurophysical tests were conducted by the study authors
without knowledge of participants' study group status.
Overall, no individual worker had obvious signs of PNS damage. However, compared to
unexposed workers, exposed workers as a group reported a statistically significant increased
incidence of headache, hearing deficit, dysesthesia in limbs, and muscle weakness (specific
muscles not indicated) as reported on a questionnaire (Table 4-1). The authors classified these
symptoms as either persistent or persistent and transient.
Sanagi et al. (1980) also carried out a number of neurological tests to identify exposure-
related neurological signs of n-hexane toxicity (Table 4-2). No objective neurological symptoms
related to muscle strength by manual testing, muscle wasting, or muscle tone were reported.
Statistically significant exposure-related deficits in muscle strength (as determined by jump test
on one foot) and reduced vibration sensation of the radial processes (determined by the tuning
fork test for vibration sensation) in the exposed group (average group values) were observed
compared to controls.
Table 4-1. Persistent and transient neurological symptoms following occupational
exposure to n-hexane in a tungsten carbide alloy factory
Symptom
Headache
Heaviness in head
Vertigo/dizziness
Anosmia/dysnosia
Vision disturbance15
Double vision
Tinnitus
Hearing deficitb
Dysphagia
Dysarthria
Incidence of symptoms (%)
Exposed
86a
71
50
46
57
36
36
71a
29
14
Exposed in the past
60
40
60
20
40
20
40
20
0
40
Controls
43
43
38
14
43
21
29
14
29
21
27
-------
Pain in neck/arm
Lumbago
Arthrodynia
Muscle pain
Sensitivity to cold
Limb dysesthesia
Limb numbness
Stiff shoulders
Tired arms
Tired legs
Muscle weakness
Writing impairment15
Unsteady gait
71
54
36
50
8
29a
21
64
57
79
29a
14
21
80
60
20
20
20
40
40
100
40
80
40
40
20
43
50
29
14
14
0
0
64
36
46
0
36
7
a Statistically significant from controls (£><0.05).
b Symptoms classified as persistent, no data are presented as persistent and transient.
Source: Sanagietal., 1980.
Table 4-2. Results of neurological tests in control subjects and those occupationally
exposed to n-hexane in a tungsten carbide alloy factory
Test (units)
Exposed group3
Controls3
Muscle strength
Grip power (kg)
Jumping on one foot (cm)
45.3 ±2.9
21.3±3.6b
44.9 ±5.2
26.0 ±6.2
Vibration sensation
Radial processes (s/16s)
Medial malleoli (s/16s)
13.8±2.4b
12.2 ±2.1
15.4 ± 1.6
13.4 ±2.0
Position sense
Barrany's test (cm)
Mann's test (%)
0.8 ±0.4
21
0.7 ±0.5
0
Coordination skills
Knee slapping (times/15s)
24.8 ±4.8
24.5 ±2.8
28
-------
Floor tapping (times/15s)
39.9 ±7.7
42.6 ±6.0
a Values are means ± SD.
b Statistically significant versus controls (p<0.05).
Source: Sanagietal., 1980.
Neurophysiological findings indicated a slowing of MCV in the posterior tibial nerve,
with delayed recovery in exposed groups compared to controls (Table 4-3). These findings are
consistent with the neurological signs and subject-reported symptoms generally associated with
n-hexane exposure.
Table 4-3. Nerve stimulation in control subjects and those occupationally exposed
to n-hexane in a tungsten carbide alloy factory
Test (units)
Motor nerve conduction velocity (m/s)
Muscle action potential ratio (%)
Residual latency (ms)
Conduction velocity of slow fibers (m/s)
Distal sensory conduction velocity (m/s)
Mixed nerve conduction velocity (m/s)
Right median and ulnar nerves3
Exposed group
57.3 ±3.4
97.2 ±5.2
2.26 ± 0.46
48.5 ±4.5
66.4 ±6.9
72.5 ±3.4
Controls
57.5 ±3.2
100.3 ±5.0
2.19 ±0.32
49.9 ±4.4
65.2 ±5.9
71.3 ±3.8
Right posterior tibial nervea
Exposed group
46.6±2.3b
90.1 ±7.4
2.55 ± 0.48b
38.6 ±2.2
42.6 ±5.0
59.1 ±3.4
Controls
48.3 ±2.1
88.9 ±11.8
2.21 ±0.34
39.1 ±1.5
41.7 ±3.9
60.2 ±3. 3
a Values are means ± SD.
b Significantly different from controls (/><0.05).
Source: Sanagietal., 1980.
Mutti et al. (1982a) monitored MCV in a group of 95 shoe factory workers exposed to a
hydrocarbon mixture containing n-hexane, cyclohexane, methyl ethyl ketone, and ethyl acetate,
and compared them to 52 unexposed workers from the same factory. Gender, age, and
employment time were similar in the exposed and referent groups. Both groups were free of
individuals suffering from diseases affecting the PNS. Neurological symptoms occurred more
frequently among the exposed than the unexposed workers. These included statistically
significant increases in the frequency of self-reported sleepiness, dizziness, weakness in the
limbs, paresthesia (burning or tingling sensation in limbs), and hypoesthesia (partial loss of
sensation and/or diminished sensibility). Electroneurographic measurements identified a
statistically significant increased motor nerve action potential (MAP) duration and decreased
29
-------
MCV in the median and ulnar nerves in exposed workers compared with unexposed workers.
Exposed workers were divided into two groups (mild- and high-exposure) based on hydrocarbon
exposure in their main jobs. The two groups were determined by exposure score, taking into
account time in job and hygienic effect (defined as the ratio between the measured concentration
of n-hexane and the proposed 1979 ACGIH TLV values for n-hexane). The TWA for n-hexane
of the 108 breathing zone samples taken was 243 mg/m3 (69 ppm) in the mildly exposed group
and 474 mg/m3 (134 ppm) in the highly exposed group. The MCV of the median nerve and the
MAP duration of the ulnar nerve were related to hydrocarbon exposure. The median for the
hygienic effect was 0.81 and 1.91 for the mild and highly exposed groups, respectively. The
authors stated that estimates of past exposure concentrations were most likely underestimated
and hygienic effects were higher due to industrial improvements implemented prior to the study.
Mutti et al. (1982b) also compared 15 women in a shoe factory and 15 healthy, age-
matched female workers at four other shoe factories with no known exposure to neurotoxic
substances or metabolic diseases affecting the PNS. Exposure times ranged from 2 to 8 years.
The breathing zone TWA of organic solvents was measured 36 times at each workplace over 3
years. Over 50% of the samples exceeded the ACGIH TLV for technical grade hexane (a
mixture containing n-hexane, cyclohexane, methyl ethyl ketone, and ethyl acetate) and was
occasionally as high as 5000-7000 mg/m3 (1422-1990 ppm). The median value for n-hexane
was 448 mg/m3 (127 ppm), and the median hygienic effect was 1.24. The study began 3 months
after industrial hygiene improvements had been made to the factory (solvent levels decreased to
trace amounts) and continued for 6 more months. The authors stated that the subjects were
considered as past-exposed and therefore any detectable toxicity could be considered as a
chronic effect. The results of neurophysiological examinations of the peroneal, ulnar, and
median nerves in exposed subjects showed a significant reduction in the maximal MCV and the
distal sensory nerve conduction velocity (SCV) compared with controls. There was also
increased latency in the somatosensory evoked potential (SEP) of exposed workers compared
with the unexposed workers. The distal SCV and the latency in the proximal segment of the
tibial nerve were negatively correlated, reflecting a peripheral neuropathy in which increased
signal latency and reduced conduction velocities were apparent in the same nerve. In addition,
the SEP was flatter in the exposed group than in the referent group, indicating a block in central
conduction and suggesting the possibility of n-hexane effects on the central nervous system
(CNS). Indeed, studies in rats (Schaumburg and Spencer, 1976) showed that neurofilament
(NF)-filled axonal swellings developed in the subterminal regions of the longest axons in the
PNS of rats exposed to n-hexane for 45 days.
30
-------
Governa et al. (1987) investigated the correlation between electrographic changes
indicative of polyneuropathy and urinary excretion of metabolites indicative of exposure to
n-hexane. Forty workers were randomly chosen from four small shoe factories. All workers
handled a type of glue or solvent that contained over 50% n-hexane without protective
equipment for about 7 hours/day. All subjects exhibited no more than mild or nonspecific
symptoms of polyneuropathy and were free of other known risk factors for nervous system
impairment. A urine sample was collected at the end of a shift, and then a neurophysiological
examination (MCV, SCV, and associated distal latencies [DL]) was carried out the following
day. Reference values were obtained from 41 unexposed individuals. A semi quantitative rating
scale of the electroneuromyography (ENM) responses was used as a cumulative index score of
the findings (Allen et al., 1975). The scale ranged from 0, where no ENM abnormalities were
observed, to 10, where decreases in conduction velocities and increases in DL were indicative of
impaired electrophysiological performance (Governa et al., 1987).
The urinary concentrations of cyclohexane, trichloroethylene, and n-hexane metabolites
were measured in 40 workers, but only those for two of the five n-hexane metabolites were
above minimum detection limits (MDLs): 2,5-hexanedione (mean, 6.80 mg/L) and y-
valerolactone (mean, 3.31 mg/L). Trichloroethylene was the only other chemical for which a
significant portion of the workers had metabolite levels above the MDL. However, the
trichloroethylene levels were much lower than those associated with the recommended
occupational exposure limit. Urinary concentrations of the n-hexane metabolite, 2-hexanol, were
below 0.1 mg/L in 29 of 40 workers. The results of the neurological exam found 26 workers
with ENM results within normal limits (ENM score <4), 3 workers with increased DL with or
without a decreased SCV or MCV (ENM score of 4 or 5), and 11 workers with these changes in
DL and SCV and MCV values plus changes in MAPs in at least one muscle (ENM score >8).
While length of exposure to n-hexane was unrelated to ENM scores, a statistically significant
dose-response relationship for these scores was found for 2,5-hexanedione and y-valerolactone.
Looking at the utility of urinary 2,5-hexanedione concentrations as a screening device to detect
significant ENM abnormalities (defined as ENM score >3), Governa et al. (1987) identified a
threshold value of 7.5 mg/L as being closely related to the incidence of abnormalities. However,
some variation from this relationship was apparent, because Governa et al. (1987) identified
three workers with 2,5-hexanedione urinary concentrations of 3.0, 3.3, and 4.5 mg/L, all of
whom displayed ENM changes.
Pastore et al. (1994) measured urinary 2,5-hexanedione in 20 asymptomatic workers with
prolonged exposure to solvents containing n-hexane. These workers were free from known
31
-------
diseases affecting the nervous system or from risk factors for alterations in nervous system
function. Urine samples, taken at the end of the shift, were all in excess of the recommended
ACGffl BEI of 5 mg/L 2,5-hexanedione, with a mean of 11.02 ± 4.5 (range 5.3 to 24.2) mg/L.
The neurological findings in these workers were compared with those obtained during the
previous 8 years in healthy adults of a similar age who were not occupationally exposed to any
toxic substance. No significant anomalies were identified in neurological examinations or
worker responses to questionnaires about neurophysiological problems. However, the results of
electrographic evaluations showed significant decreases in the amplitude of sensory nerve action
potential (SNAP) for the median, sural, and ulnar nerves. These results were unrelated to urinary
2,5-hexanedione levels. However, the SNAP amplitude for the sural and median nerves was
significantly related to the number of years exposed to n-hexane. Adjusting for age did not alter
these results. No differences were found in values of the SCV, MCV, compound muscle action
potential, and F wave latency (a more precise indication of small variations in conduction) for
the nerves evaluated.
Murata et al. (1994) studied the effects of solvent exposure on the autonomic nervous
system and cerebellar function in shoe and leather workers exposed to n-hexane, xylene, and
toluene. 2,5-hexanedione, hippuric acid, and methylhippuric acid concentrations in urine
samples (taken the morning prior to electrophysiological examination) were determined.
Urinary concentrations of 2,5-hexanedione were 0-3.18 (mean 1.39) mg/L; concentrations of
hippuric acid were 0.05-2.53 (mean 0.41) g/g creatinine; and concentrations of methylhippuric
acid were 0.10-0.43 (mean 0.19) g/g creatinine for occupationally exposed workers. In
unexposed workers, the urinary concentration of 2,5-hexanedione was 0.1-0.8 g/g creatinine and
hippuric acid was < 1.5 g/g creatinine; methylhippuric acid was not found. Exposure
concentrations for n-hexane, xylene, or toluene were not reported by the study authors. The
study subjects were free of known confounding factors related to nervous system function and
were similar in their reported use of alcohol and tobacco. Exposed workers had worked in
household factories for a period of 18-42 years (31 ± 6 years). Murata et al. (1994) measured
the distribution of MCVs and SCVs of the median nerve and the variation in the
electrocardiographic duration of the ventricular cardiac cycle (R-R interval) in 30 workers and in
25 healthy controls unexposed to solvents. The SCV and MCV of the median nerve were
significantly slowed in exposed workers compared with unexposed. Variations in the R-R
interval and the respiratory sinus arrhythmia component of the R-R interval also were
significantly lower in the exposed group. The SCV in the forearm was significantly correlated to
the variation in the Mayer sign wave arrhythmia component of the R-R interval. Duration of
32
-------
exposure, concentration of urinary metabolites for solvent exposure, age, or alcohol consumption
were not significantly related to any of these electrophysiological results. While the results
imply that both the PNS and the autonomic nervous system were affected by solvent exposure,
failure to identify a dose-response relationship and the mixture of solvents to which the workers
were exposed led to equivocal results for n-hexane. Because the urine was collected more than
12 hours after exposure, the concentrations of metabolites in urine may have been an
underestimation of the actual solvent exposure.
In a study of the same workers as those monitored by Murata et al. (1994), Yokoyama et
al. (1997) evaluated 29 subjects and 22 healthy unexposed controls for postural sway frequency
in order to assess subclinical cerebellar dysfunction. Subjects were male workers in shoe,
sandal, and leather factories who routinely were exposed to n-hexane, xylene, and toluene during
the course of their work. Postural balance was measured quantitatively using a strain-gauge-type
force platform on which subjects were asked to stand for 60 seconds with their eyes open and
then for 60 seconds with their eyes closed. Lengths of displacement of the body's center of
pressure in the mediolateral and anteroposterior directions were used as indicators of the extent
of postural sway in each direction. Mean urinary concentration of 2,5-hexanedione was 1.20
mg/L (range 0.41 to 3.06 mg/L), and the estimated mean level in workplace air was 40 ppm
(range 13 to 100 ppm) n-hexane. The measurements of postural balance, specifically
spinocerebellar afferent type of sway, showed a significant positive association with
2,5-hexanedione concentration in urine. The authors indicated that xylene could possibly inhibit
the effects of n-hexane exposure on sway. Specifically, there was an inverse correlation between
urinary methylhippuric acid from xylene exposure and vestibulocerebellar type of sway.
Passero et al. (1983) screened 654 workers in 44 shoe factories and 86 home shops
during a period from 1973 to 1981. Evaluation by clinical and electrodiagnostic examination
identified 184 workers with some degree of neurological abnormality. Of these 184 subjects, 9
had other neurological disorders (the authors reported the most common of which was
radiculopathy due to intervertebral disc disease), 77 displayed minimal changes and were
considered normal following repeated examination by the study authors, and 98 manifested overt
polyneuropathy. The majority of the workplace solvent samples collected contained commercial
hexane. The commercial hexane was determined to contain greater than 60% of total mass as
hydrocarbons such as pentane, 2-methyl-pentane, 3-methyl-pentane, n-hexane, heptane,
cyclopentane, cyclohexane, and methyl-cyclopentane. In 7 of 12 samples taken from workplaces
of individuals with the most severe polyneuropathy, over 99% of the total solvent was composed
of these hydrocarbons. No relationship was found between length of exposure and severity of
33
-------
disease. In the cases of polyneuropathy, the neurological pattern showed an insidious onset of
loss of distal motor and sensory function with marked reflex loss. General symptoms, such as
nausea or vomiting, epigastric pain, and insomnia, preceded or accompanied the neuropathy.
Clinical symptoms were weakness, paresthesia (burning or tingling sensation in limbs), and
cramp-like pain with related motor impairment, hypoesthesia (partial loss of sensation and/or
diminished sensibility), changes in tendon reflexes, and muscle trophism and tone. These
symptoms were usually confined to distal portions of the limbs and occurred with varying
degrees of intensity depending on the extent of exposure. All 98 polyneuropathy cases exhibited
abnormal MAPs, regardless of severity. The occurrence of fibrillations, positive waves,
fasciculations, and slowing of MCV increased with disease severity. Several of the most
affected cases exhibited CNS involvement with alterations in electroencephalogram or spasticity
in the lower limbs and increased deep tendon reflexes. The clinical course of these 98 cases was
followed for up to 8 years. Except for the most severe cases, patients improved slowly when
removed from the affected environment. However, deterioration continued for some even after
exposure ceased.
A group of 122 cases of polyneuropathy among workers in 72 shoe factories was
evaluated for severity of neurological impairment in relation to duration of exposure to a mixture
of solvents (Abbritti et al., 1976; Cianchetti et al., 1976). Every worker with polyneuropathy
was questioned about work experience; type of chemical material used on the job; specific job
function performed at onset of disease, and in the years previous to onset of disease; symptoms;
and the order of appearance and evolution of symptoms. All patients were given an
electromyographic examination and were determined to be free of lifestyle or medical conditions
such as diabetes and alcoholism that would cause neurological impairment. None of the subjects
had a history of exposure to other chemicals that might cause neuropathy, such as lead, arsenic,
carbon disulfide, or drugs such as sulfonamides. The workers were divided into three groups
based on severity in the reduction of the MCV of the peroneal nerve. Group I had a maximum
MCV of less than 35 m/s; Group II had an MCV of 35-44 m/s; and Group III had a maximum
MCV of 45 m/s or greater. No quantitative air measurements were taken, but samples of five
glues and two cleaners from five factories in which 20 cases worked were analyzed for several
solvents of interest. Six of these samples contained at least 40% n-hexane, with other solvents,
such as pentane, 2-methylpentane, 3-methylpentane, toluene, and cyclohexane, usually present.
No correlation was identified between severity of neuropathy and length of employment in the
factory.
Sobue et al. (1978) identified 93 cases of polyneuropathy among 1662 shoe workers
34
-------
screened. These workers were divided into three groups according to the presence of (1) sensory
symptoms only (53 subjects), (2) sensorimotor symptoms (32 subjects), or (3) sensorimotor
symptoms with amyotrophy (eight subjects). Follow-up of a subgroup of workers continued for
up to 18 months. All 93 cases were engaged in operations that used pastes consisting of at least
70% n-hexane. Air concentrations of n-hexane in the workrooms ranged from 500 to 2500 ppm.
The degree of neurological disorder was related to hygienic conditions in the workplace. A
further reduction in MCV was noted among some of the workers with the most severe
polyneuropathy, even after removal from the solvent exposure. In addition to signs of sensory
disturbance that are typical of peripheral neuropathy, such as numbness and hypoactive reflexes,
Sobue et al. (1978) found evidence of cranial nerve impairment in some cases. These were
indicated by blurred vision (in 13 subjects), constriction of the visual field (seven subjects), and
numbness over the face (five subjects). In an earlier report on the same group of patients,
Yamamura (1969) tabulated the results of semi quantitative laboratory analyses of biochemical
parameters in blood and urine, most of which were unremarkable except for apparently
depressed levels of serum cholinesterase activities, elevated levels of serum lactate
dehydrogenase (LDH) activity, and positive urobilinogen.
Paulson and Waylonis (1976) also found reduced serum cholinesterase levels among
eight printing room workers who had polyneuropathy and were exposed to n-hexane. Air levels
of n-hexane taken over a 2-month period were found to be as high as 4060 mg/m3 (1152 ppm) in
the plant.
Wang et al. (1986) evaluated a group of 59 press proofing workers from 16 factories who
were employed for at least 2 months. All but four of these workers had regular contact with
solvents in the process of cleaning the rollers. Two exposure measures using personal air
samplers were taken in 14 of the 16 factories. Samples of the bulk cleaning solvents were found
to contain n-hexane at concentrations ranging from 10-65%. Referent neurological data were
collected from 150 healthy individuals (50 persons from three age groups, 10-35, 36-50, and
51-80 years, sex not stated). MCVs among workers exposed to n-hexane were consistently
lower than among controls. The results of the neurological examination identified 15 workers
with polyneuropathy and two asymptomatic workers with abnormal MCVs. All but one of these
workers were employed in factories that used solvents with n-hexane concentrations in excess of
50%. While no association was found with length of employment, statistically significant
associations existed between frequency of polyneuropathy and abnormal MCV and n-hexane
concentration in the cleaning solvents and between the frequency of polyneuropathy and
n-hexane air concentrations. Among the workers with polyneuropathy, a high percentage
35
-------
worked in factories with n-hexane air concentrations greater than 100 ppm. However, a
significant reduction in the MCV was found among workers exposed to air concentrations less
than 25 ppm, a result that the authors considered to be related to the prolonged exposure due to
overtime work.
The 15 cases of polyneuropathy from the study by Wang et al. (1986) were included in a
group of 28 color printers with polyneuropathy studied by Chang and Yip (1987). This study of
electromyography (EMG) changes also included five subclinical cases, 45 workers with no
apparent symptoms, and 72 normal subjects who served as the control group. Among the
clinical and subclinical cases, a significant decrease in MCVs and in amplitude of MAPs and
SNAPs and a significant prolongation of latencies were seen compared with controls. Among
the exposed workers with no apparent symptoms, MCVs were slower, motor DLs were
prolonged, and SNAP amplitudes were attenuated compared with controls. The percentage
difference in these electromyographic changes from the control values increased with increasing
severity of symptoms.
Chang (1990) followed 11 of the 28 polyneuropathy cases for 4 years. The authors
observed the patients monthly for the first 2 years, bimonthly for the third year, and once every 3
months for the final fourth year. All 11 cases had moderate to severe polyneuropathy. There
was some worsening of motor function and electrographic findings in nine of the cases even after
exposure to n-hexane ceased. Delayed worsening of sensory function was not observed.
Sensory disturbances usually disappeared within 4 months. All patients, including the most
severely affected, who was a quadraplegic, regained full motor nerve capacity within 1-4 years.
Tightness in the legs, which appeared early in the course of recovery for six of the more severe
cases, was replaced by muscle cramps, which persisted up until the last clinical visit 4 years after
the onset of neuropathy. Two of the six also had hyperreflexia and residual muscle atrophy in
the lower extremities, and one had only residual atrophy. The inability of two of the subjects to
perceive colors correctly (dyschromatopsia) persisted until the end of the study. These patients
also had macular retinopathy.
In addition, Chang (1991) documented the electrophysiological performance of the 11
cases from their initial diagnosis to complete recovery of motor nerve capacity. Compound
muscle action potentials (CMAPs), DLs, and nerve conduction velocities were measured in
motor and sensory nerves, and pattern visual evoked potentials (VEPs) were assessed in relation
to those of a group of unexposed control individuals. Full recovery was associated with a return
to normal values for patterned VEPs, CMAPs, and DLs of both motor and sensory nerves.
However, the nerve conduction velocity of motor nerves remained significantly lower than
36
-------
normal values even when the patients had apparently made a complete recovery (Table 4-4).
Table 4-4. Motor neurographic findings in patients with n-hexane polyneuropathy
Median nerve
CMAP amplitude (mV)
DL(ms)
MCV (m/s)
Ulnar nerve
CMAP amplitude (mV)
DL(ms)
MCV (m/s)
Peroneal nerve
CMAP amplitude (mV)
DL(ms)
MCV (m/s)
Tibial nerve
CMAP amplitude (mV)
DL(ms)
MCV (m/s)
Motor neurographic findings (mean ± SD)
Initial study (n = 11)
4.1±2.1
6.1 ±0.7
42.9 ±7.1
4.2 ± 1.7
5.0 ±0.8
40.8 ±4.6
1.8 ±1.0
8.2±1.1
31.9±3.3
4.0 ±2.4
7.6 ±1.3
34.5 ±4.3
Final study (n = 11)
10.0 ±2.1
3.9 ±0.4
57.3±2.5a
8.5 ±1.1
3.1 ±0.2
52.8±4.1b
5.6 ±2.5
4.8 ±0.9
46.1±4.9C
10.7 ±5.7
4.5 ±0.2
46.3±4.8b
Controls (n = 72)
8.9 ±3.0
3.7 ±0.4
61.9 ±4.6
7.8 ±2.2
3.0 ±0.4
55.4 ±4.0
5.9 ±1.9
4.7 ±1.0
53.1 ±4.4
10.2 ±3.7
4.5 ±0.8
49.6 ±3. 9
a (^<0.005);b (^<0.05); andc (/X0.0005): Significantly different from controls, as calculated by
the authors.
Source: Chang, 1991.
Chang et al. (1992) further described the effects of n-hexane exposure on the
neurological function of 56 offset machine workers in a printing factory. These workers, who
were free of other known risk factors related to neurological function, such as alcoholism and/or
diabetes mellitus, were the primary users of a cleaning solution containing 14-20% n-hexane.
Subjects typically worked 12 hours/day, 6 days/week, and used a solvent-soaked cloth or sponge
to manually clean the printer roller blanket surface 2 to 3 times each hour. They did not wear
respirators, but all except four workers wore latex gloves when performing this cleaning
operation. While other solvents were present in these cleaning solutions (toluene), n-hexane was
predominant and the only one with TWA air concentrations above the TLV (50 ppm). The
37
-------
workers were also exposed to lead (0.6-8.2 |ig/g) and mercury (0.05-0.95|ig/g) in the printing
inks. The mean TWAs for n-hexane in air samples taken at the plant was 63 ppm (range 30-110
ppm) for general air and 132 ppm (range 80-210 ppm) for breathing zone air of the offset
printers. The findings from the neurological examination of these workers were compared with
the neurological findings of 20 age- and gender-matched unexposed controls. Among the 56
printers, 10 were asymptomatic and showed no clinical findings (classified as healthy workers);
26 were asymptomatic, but had subclinical nerve conduction deficits (classified as subclinical
workers); and 20 were symptomatic for peripheral neuropathy by clinical and electrophysical
findings (classified as symptomatic workers).
No relationship was observed between length of employment and the development of
neuropathy. Workers that were symptomatic reported experiencing symptoms in the upper and
lower extremities such as numbness (8/20 and 14/20, respectively), paraesthesia (5/20 and 13/20,
respectively), pain (2/20 and 9/20, respectively), and weakness (7/20 and 15/20, respectively) of
the feet and distal portion of the legs. Subjects in the symptomatic group had more severe
symptoms in the lower than the upper limbs, and sensory disturbances usually appeared before
motor disturbances. Four workers who did not use gloves when cleaning the roller blanket
showed symptoms in the upper limbs first. No autonomic neuropathy was reported by any
worker. Clinical symptoms were not reported for healthy workers or for those who were
asymptomatic, but classified as having subclinical neuropathy. The authors also measured
electrophysical deficits in each group of workers (Table 4-5). In healthy workers, there was a
decreased amplitude of the median nerve SNAP compared with controls. Subclinical workers
had more significant decrements in the SNAP amplitude and MCV. There was also a mild
reduction in mean MAP amplitudes and prolongation of mean DL. The symptomatic workers
were reported to have a marked reduction in mean SNAP, MAP, and MCV and prolonged DL.
Table 4-5. Nerve conduction study findings in printers with n-hexane-induced
polyneuropathy
Type of nerve
conduction study
Control
(n = 20)a
Healthy worker
(n = 10)a
Subclinical worker
(n = 26)a
Symptomatic
worker (n = 20)a
Amplitude of SNAP (jiV)
Median
Ulnar
Sural
37(11)
15(4)
24 (10)
27 (6)b
14(3)
22(6)
24 (8)b
12(5)
18 (7)b
15 (5)b
7(4)b
11 (8)b
38
-------
Amplitude of MAP (mV)
Median
Ulnar
Posterior tibial
Common peroneal
7(2)
5.7(2.1)
6.6 (2)
4.4 (1.5)
8(3)
6.5(1.9)
6.7 (2.3)
4.3 (1.4)
6.7 (2.4)
4 (2.2)b
53 (2.5)
3.6 (1.4)
4.6 (2.2)b
3.6(1.5)b
2.9 (1.7)b
1.8(1.4)b
DLofSNAP(ms)
Median
Ulnar
Sural
2.3 (0.3)
2.1 (0.3)
3.3 (0.3)
2.3 (0.1)
2.1(0.1)
3.1 (0.2)
2.6 (0.3)b
2.3 (0.3)
3.3 (0.3)
2.9 (0.3)b
2.7 (0.4)b
3.7 (0.6)b
DLofMAP(ms)
Median
Ulnar
Posterior tibial
Common peroneal
2.9 (0.4)
2.2 (0.3)
4.1(0.6)
3.9 (0.5)
3.0 (0.2)
2.3 (0.3)
3.9 (0.6)
3.5 (0.4)
3.6 (0.5)b
2.6 (0.5)b
4.4 (0.6)
4.2 (0.7)
4.3 (1.2)b
3.0 (0.7)b
5.6 (1.2)b
5.4 (1.2)b
MCV (m/s)
Median
Ulnar
Posterior tibial
Common peroneal
59 (5.9)
61 (5.8)
50 (6.4)
51 (4.5)
57(5)
59(6)
46 (3.4)
46 (3.8)
55 (6.7)b
55 (7.8)b
45 (4.7)b
45(5.1)b
46 (6.5)b
48 (7.5)b
38 (6.5)b
37(7.1)b
a Values presented as means with SDs in parenthesis.
b Significantly different from controls (p<0.05), as calculated by the authors.
Source: Chang et al., 1992.
Several studies (Issever et al., 2002; Seppalainen et al., 1979; Raitta et al., 1978) have
investigated vision changes in relation to n-hexane exposure. Seppalainen et al. (1979)
compared the VEP and electroretinograms (ERGs) of 15 workers to those of 10 healthy subjects
with no occupational exposure to solvents or other neurotoxic chemicals. The highest recorded
n-hexane levels in the two factories where the workers were exposed ranged from 2000 to 3250
ppm. In both factories, exposure was to technical grade hexane, which contains other aliphatic
hydrocarbons with no known neurotoxic effects. Maculopathy, color discrimination deficits,
flatter VEPs, and diminished peak-to-peak amplitudes of the ERGs were more common among
39
-------
cases than controls.
An earlier study by the same researchers described visual defects in this same group of
15 workers, 12 of whom displayed impaired color vision (Raitta et al., 1978). The Farnsworth-
Munsell (FM)-IOO hue test showed 12 of the subjects to have impaired color vision, one of
which was probably due to a congenital abnormality. The other cases of color vision impairment
were acquired, mostly in the blue-yellow axis. In 11/15 subjects there was evidence of
associated maculopathy (damage of vessels in eye that leak fluid into the center of the retina
causing loss of central vision), in most cases characterized by pigment dispersion.
Issever et al. (2002) compared the color vision of 26 workers with diagnosed
polyneuropathy resulting from n-hexane exposure to that of 50 nonexposed healthy controls.
The workers had been exposed to n-hexane in the leather industry (exposure concentration not
stated), and all complained of asthenia of upper and lower limbs, paraesthesia in hands and arms,
and difficulty walking. EMG results indicated myelinic and axonal lesions of distal nerves.
None had been screened for color vision during employment. All study subjects were free of any
visual disorder or inherited color vision impairment. However, in the FM-100 hue test, there
was significant impairment in color vision in n-hexane-exposed subjects (Table 4-6). Color
vision defects did not show specificity for the blue-yellow or red-green axes but appeared to be
distributed across the entire range of color vision deficits.
Table 4-6. FM-100 hue test error scores of n-hexane-exposed and nonexposed
groups
Eye
Right
Left
Chromatic focus
Blue-green
Red-green
Total
Blue-yellow
Red-green
Total
Exposed group3
104.3 ±37.3b
64.1±38.4b
168.3 ±70.5b
96.7±55.8b
81.9±51.8b
181.5±103.0b
Control group3
22.2 ±13.5
13. 8 ±8.3
36.0 ±19.8
21.6±11.0
14.0 ±9.4
35.6 ±18.2
a Values are presented as mean ± SD.
b Significantly different from controls (/?<0.001), as calculated by the authors.
Source: Issever et al., 2002.
40
-------
Huang and Chu (1989) used evoked potentials (EPs) to examine the extent to which
exposure to n-hexane in the workplace might bring about subclinical effects on the central
nervous system. The study involved five workers who had developed peripheral neuropathy as a
result of being exposed to n-hexane in a press proofing factory. The subjects worked without
protection in a poorly ventilated room and were exposed to three types of solvent mixtures, one
of which contained 65% n-hexane. Two 1-hour air samples collected from the work room
showed ambient air concentrations of 55 ppm for n-hexane and 9.65 ppm for benzene. Motor
nerve conduction data, brainstem auditory evoked potential (BAEP), VEP, and somatosensory
evoked potential (SEP) were obtained in exposed subjects and compared with those of
unexposed subjects of comparable age. Four of the five exposed subjects displayed typical signs
of neurological impairment, such as muscle cramps, weakness, and distal numbness. In contrast
to the fifth (unaffected) subject, these four individuals had slept in the workroom between shifts.
As tabulated by the authors, the four most severely affected subjects also displayed such
electrophysiological deficits as lower conduction velocities and amplitudes in the median, ulnar,
peroneal, and tibial nerves. Statistically significant changes also were seen in SEPs and BAEPs
when the five exposed subjects were compared as a group with controls. The SEPs and BAEPs
of the fifth (less severely affected) subject were higher than those of control subjects, suggesting
that evoked potentials may be discriminating parameters for revealing subclinical neuropathies.
Scelsi et al. (1980) observed neuropathological symptoms and morphological changes in
a small number of subjects (three women) occupationally exposed to an adhesive agent
containing approximately 80% n-hexane for 2 months to 3 years. These workers experienced
bodily discomfort and increasing weakness in the lower limbs. Biopsies of the sural nerves
revealed axonal swelling and irregular and swollen myelin sheaths. In general, there appeared to
have been a dissolution of neurotubules and an increase in the number of microfilaments.
Specifically, histopathology and electron microscopy showed polymorphous changes in both
myelin sheaths and axons of large diameter fibers of the sural nerve. In addition, there were
irregular and swollen myelin sheaths and segmental swelling of axons with dissolution of
neurotubules and increase of NFs. Polymorphous inclusion bodies were also identified in
Schwann cells. The authors also observed atrophy and focal degenerative myopathic changes
with lymphocytic infiltrates and phagocytosis in the soleus muscle. In one of the subjects, these
changes in the muscle appeared to be associated with muscular denervation, lymphocytic
infiltrations, and phagocytosis.
A follow-up study on the same subjects showed differing severities in the
neuropathological responses (Scelsi et al., 1981). For example, one of the subjects suffered so
41
-------
much from muscular wasting that she was unable to walk. However, neuropathy in the others
was less severe, although symptoms were sufficient for a diagnosis of motor polyneuropathy.
Electrophysiological measurements showed motor and sensory conduction velocities to be
reduced. Light and electron microscopy of the sural nerve showed the presence of large
diameter, irregularly shaped myelinated fibers, degenerated myelin, vacuoles, and abnormal
organelles in two of the subjects. The third subject had scattered large fibers, with thin myelin
sheaths and enlarged axons filled with packed NFs.
Yokoyama et al. (1990) examined the relationship between impaired nerve conduction
velocities and morphological changes seen at biopsy in three workers exposed to n-hexane in a
plant manufacturing parts for jet engines. A single air sample taken in the middle of the
workroom showed the n-hexane concentration to be 195 ppm. All three workers developed
progressive muscular weakness, with numbness and tingling sensations in the lower extremities.
Their status progressed to the extent of work disability with difficulty walking or holding tools.
The study authors measured the distribution of conduction velocities (DCVs) of sensory fibers in
sural nerve at 1-2 months, 4-9 months, and 11, 23, and 36 months after exposure was ended.
Values were obtained in two patients that fluctuated below the lower limit of normal as
determined in 11 male subjects who had no history of exposure to chemicals. A sural nerve
biopsy was obtained from one of the patients 10 weeks after cessation of exposure and showed
degeneration of myelinated nerve fibers and paranodal swelling.
Electrophysiological deficits with subsequent recovery upon removal of subjects from
n-hexane exposure were also demonstrated in four workers who had been exposed to n-hexane
during the course of their work at a ball-manufacturing plant in Taiwan (Huang et al., 1989).
The main source of exposure was solvent evaporation from a cement coating and nylon fiber
winding facility. The n-hexane concentration in the poorly ventilated room averaged 108.9 ppm,
the vapors resulting from a bulk solvent that contained 14.1% n-hexane by weight. The four
workers displayed overt symptoms of peripheral neuropathy and showed electrophysiological
deficits in motor conduction of the median, ulnar, peroneal, and tibial nerves. After cessation of
exposure, there was an initial worsening in muscle strength, sensory deficit, and nerve
conduction that lasted for 2-5 months, with recovery occurring over the course of the following
year.
Huang et al. (1991) carried out a follow-up study of subjects potentially exposed to
n-hexane at the factory described above. Forty-four workers were interviewed and classified
according to their potential exposure to solvents. Five individuals were assigned to Group I
because they had been involved in cement coating and the winding process; eight were assigned
42
-------
to Group II because of their involvement in a gas injection process; and the remaining 31 were
placed in Group III, reflective of a lower potential risk of exposure. All subjects were
administered a questionnaire detailing personal history and describing subjective symptoms.
The subjects also were given a physical and neurological examination. Electrophysiological
determinations of nerve conduction and latencies, and electromyographic measurements were
compared with those of 26 age- and sex-matched controls.
Personal air monitoring samples gave values of 109 ppm n-hexane in the cement coating
section (two samples), 86.4 ppm in the nylon fiber winding section (three samples), and 75 ppm
in the gas injection section (one sample). Neuropathy was evident in all subjects in Group I.
Four of these individuals had severe symptoms while the other subject's symptoms were
considered to be moderate. Two of eight individuals in Group II, but none in Group III,
displayed mild polyneuropathy. In the EMG studies, fibrillations, positive sharp waves, and
increasing polyphasic waves were found in all five Group I cases, whereas only increased
polyphasic waves were noted in the eight workers in Group II. In motor nerve conduction
studies, decreased nerve conduction velocity, prolonged distal latency, and reduced evoked nerve
potential were observed in all four nerves studied. Workers in Group I had more severe
neurological symptoms, consistent with exposure to higher n-hexane concentrations in the
cement coating and nylon fiber winding sections. The factory owner installed a new ventilation
system and enclosed portions of the organic solvent operations. n-Hexane concentrations in air
fell to 12.9 ppm in cement coating, 14.7 ppm in nylon fiber winding, and 1.0 ppm in gas
injection. There were no new cases of polyneuropathy in a 2-year follow-up at the factory, and
all seven originally affected subjects recovered.
Valentino (1996) described an occupational exposure study in which 27 female and 63
male workers in the shoe industry were assessed for polyneuropathy in relation to n-hexane
exposure. All of the subjects, who had been free of exposure for at least a year, were divided
into groups according to whether they had been exposed to n-hexane either during the last 10
years or more than 10 years ago. In general, more recently exposed subjects had a higher
incidence of neuropathological symptoms. In electrophysiological measurements, Valentino
(1996) showed recovery of the motor component, while recovery of the sensory component was
incomplete even 10 years after n-hexane exposure.
Bachman et al. (1993) discussed the cases of 63 workers in a metal can factory. Subjects
received a neurological examination, including measurements of motor function, proprioception,
sensitivity to vibrations, and vibriotactile thresholds. The workers were placed in low and high
exposure groups on the basis of their duration of exposure at the subject location at which, over a
43
-------
3-year period, 17 personal sampling measurements had been taken. These indicated an n-hexane
concentration range of 181-2436 mg/m3 (50-690 ppm). However, the subjects in the study
showed few neurotoxic effects of n-hexane, with no overt cases of clinical neuropathy.
Smith and Albers (1997) presented a case report of an individual subacutely exposed to
n-hexane vapor by sniffing glue. The authors noted that the 25-year-old male subject had a
history of recreational drug use and was exposed to several other non-hexane constituents from
sniffing glue. The 25-year-old male showed progressive weakness, pain, diminished sensations
in the feet, and numbness in the hands. An electrophysiological examination of the subject
revealed reduced SNAPs and SC Vs, increased DLs, and an apparent block of conduction in the
median and ulnar nerves (forearm), peroneal nerve (knee), and tibial nerve (leg). A biopsy of the
sural nerve showed the classic pathological response to n-hexane exposure, including swollen
axons, reduced numbers of myelinated fibers, and the appearance of abundant NFs. An initial
increase in severity of the symptoms on cessation of exposure caused the subject to be confined
to a wheelchair. However, a subsequent, slow improvement resulted in his strength returning to
normal after a year. Reflexes were reestablished and changes to electrophysiological parameters
paralleled the clinical recovery.
In a published abstract, Pezzoli et al. (1989) described a female leather goods worker
who had been exposed to n-hexane for many years prior to developing symptoms of
Parkinsonism. Symptoms of axonal neuropathy were evident in the patient, as would be typical
in subjects chronically exposed to n-hexane. In a follow-up of the same patient, Pezzoli et al.
(1996) reported that the subject's disease progressed even after she withdrew from the work
environment. Pathological examination and immunohistochemical analysis of the brain showed
severe and widespread dopaminergic neuronal loss, severe gliosis in the substantia nigra, and
near total loss of immunostaining of tyrosine hydroxylase in the striatum. The authors stated that
the patient's disease resembled a rigid akinetic form of parkinsonism with levodopa-induced
dystonias. However, pathological examination of the brain revealed similarities to human 1-
methyl-4-phenyl-l,2,4,6-tetrahydropyridine-induced parkinsonism. In general, interpreting these
results as indicative of the etiological importance of n-hexane exposure to the onset of
Parkinsonism is complicated by the fact that some symptoms are typical of the condition while
others differ. The extent to which the solvent may be causative also is difficult to gauge.
The same researchers presented a case report of a 53-year-old man who had worked since the
age of 15 in a leather-processing factory (Pezzoli et al., 1995). The glue used at his workstation
consisted of 50% n-hexane, 17% 3-methylpentane, 15% 2-methylpentane, 13%
methylcyclopentane, and 4% cyclohexane. A 24-hour urine sample from the patient contained
44
-------
0.79 mg/L 2,5-hexanedione. Neurological examination of the patient showed some signs of
parkinsonism (tremor, bradykinesia, and rigidity). Other symptoms and findings, when taken
together, contributed to a weighted Unified Parkinson Disease Rating Scale (UPDRS) motor
score of 13. UPDRS is a three part evaluation of various physical and mental characteristics
(motor skills, mental status, behavior, mood, and activities of daily living) on a 0 to 5 scale. The
cumulative score gives an index of severity of Parkinson's disease (on a scale of 0 to 199, with
199 being the most severe cases of Parkinson's disease). Motor and sensory nerve conduction
studies showed essentially normal results, except that the medial and lateral plantar nerve
sensory action potentials were bilaterally absent. Following diagnosis, the subject was removed
from n-hexane exposure scenarios at work. The level of urinary 2,5-hexanedione declined to
0.25 mg/L, a value thought to be indicative of no occupational exposure. However, signs of
parkinsonism worsened, and the patient achieved a weighted UPDRS motor score of 23, 30
months after the onset of the disease.
Vanacore et al. (2000) evaluated the possible association between chronic exposure to
n-hexane and parkinsonism on the basis of the clinical and occupational history of a 55-year-old
patient who had worked for 17 years in an environment where mixtures of aliphatic
hydrocarbons (53% n-hexane) were used. Results of neurophysiological, neuroradiological, and
neuropsychological tests suggested that n-hexane may affect the CNS. For example, magnetic
resonance imaging showed a marked cortical cerebral atrophy, and the neuropsychological
assessment revealed impaired visuomotor response, as well as loss of short- and long-term
memory. The patient's apparent parkinsonism did not change much over the next 5 years.
In a further study to evaluate the effect of environmental influences on the pathogenesis of
parkinsonism, Canesi et al. (2003) measured the total urinary concentration of 2,5-hexanedione
and 2,5-dimethylpyrrole adducts in the urine of 108 patients with parkinsonism compared with
108 unaffected subjects. The 2,5-dimethylpyrrole adduct production is a result of the interaction
of 2,5-hexanedione with NF protein lysine residues. Urinary excretion of both 2,5-hexanedione
and 2,5-dimethylpyrrole were significantly reduced in patients with parkinsonism compared with
controls, although there was a gradual decline in urinary metabolite levels with age in both
Parkinsonism patients and controls. The data did not indicate whether the lower urinary levels
were possibly the result of reduced conversion of n-hexane to its metabolites or of an increase in
their further catabolism.
Karakaya et al. (1996) investigated the effects of n-hexane on the immune system.
Immunological parameters in 35 workers exposed to n-hexane were compared with 23 age-
matched controls with no history of n-hexane exposure. Exposure to n-hexane was measured by
45
-------
TWA concentration in the breathing zone air and urinary 2,5-hexanedione (following acid
hydrolysis) at the end of each shift. The mean TWA for n-hexane in air was 123 ppm (range
23-215 ppm), and the mean level of 2,5-hexanedione in urine was significantly higher in
exposed workers (2.39 |ig/g creatinine) compared with unexposed workers (0.41 |ig/g
creatinine). Whole blood samples were analyzed for immunoglobulins IgG, IgM, and IgA, and
white blood cell (WBC) types. Immunoglobulin concentrations in exposed workers were
significantly lower than in unexposed workers. A significant inverse correlation also was found
between the immunoglobulin levels and the 2,5-hexanedione concentrations in the exposed
group. No differences were detected in WBC counts between exposed and unexposed subjects.
Yucesoy et al. (1999) examined the effects of occupational coexposure to n-hexane,
toluene, and methyl ethyl ketone on natural killer cell activity and some immunoregulatory
cytokine levels in shoe workers. Twenty three male shoe workers were studied in comparison
with 18 unexposed controls. Levels of solvents in breathing zone air were 58.07 ± 28.09 ppm
(range 4.3-300 ppm) for n-hexane, 26.62 ± 10.27 ppm (range 5.37-116.2 ppm) for toluene, and
11.39 ± 4.86 ppm (range 2.43-47) for methyl ethyl ketone. Urinary levels of 2,5-hexanedione
and hippuric acid were measured, and natural killer cell cytotoxic activity in peripheral
lymphocytes using human erythroleukemic cells (K562) as targets was monitored. Serum levels
of interleukin-2 and y-interferon also were monitored. The authors suggested that the levels of
2,5-hexanedione in the urine of exposed subjects (3.22 ± 0.44 mg/g creatinine versus 0.98 ± 0.22
mg/g creatinine in controls) indicated that the workers had been exposed to n-hexane. Natural
killer cytotoxic activity and serum levels of interleukin-2 and y-interferon did not differ between
exposed and control groups.
To protect against the onset of subclinical and clinical neuropathological symptoms of
n-hexane exposure, ACGIH proposes a BEI of 0.4 mg/L as an acceptable concentration of
2,5-hexanedione in urine (non-acid hydrolyzed samples) and a BEI of 5 mg/g creatinine (acid
hydrolyzed samples) at the end of shift on the last day of a workweek (ACGIH, 2003; ACGIH,
2001). The ACGIH TLV-TWA of 50 ppm is based on a review of the peer-reviewed literature
indicating that commercial hexane mixtures contain approximately 50-70% n-hexane. Further,
the skin notation was assigned because studies indicate that dermal exposure to n-hexane leads to
peripheral neuropathy in humans (ACGIH, 2001).
46
-------
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION
4.2.1. Oral Exposure
4.2.1.1. Subchronic Studies
A few studies have addressed the toxicity of n-hexane when administered via the oral
route to experimental animals. Krasavage et al. (1980) compared the neurotoxicity of
2-hexanone, 2-hexanol, 2,5-hexanedione, 2,5-hexanediol, 5-hydroxy-2-hexanone, n-hexane, and
practical grade hexane. Groups of five male COBS CD(SD)BR rats/group received equimolar
doses of 6.6 mmol/kg of the chemicals by gavage, 5 days/week, for 90 days. The mg/kg
equivalents of the equimolar daily doses were 660 mg/kg 2-hexanone, 780 mg/kg
2,5-hexanediol, 755 mg/kg 2,5-hexanedione, 765 mg/kg 5-hydroxy-2-hexanone, 675 mg/kg
2-hexanol, 570 mg/kg n-hexane, and 4000 mg/kg practical grade hexane. After a month of
treatment, additional groups of five rats were administered n-hexane at 13.2 and 46.2 mmol/kg
(1140 and 3980 mg/kg, respectively). The period of treatment and observation was extended to
120 days for those animals receiving 46.2 mmol/kg n-hexane to ensure that an overt
neuropathological endpoint was detected in rats exposed to the chemical. The onset of
neuropathy was assessed by the initial appearance of hind-limb paralysis, at which point the
animal was sacrificed and examined histopathologically.
Two rats in the 13.2 mmol/kg exposure group, one rat in the 46.2 mmol/kg n-hexane
exposure group, and three rats in the practical grade hexane group died due to chemical
pneumonitis following intubation and were not included in histological analyses. The authors
observed the clinical manifestation of hind-limb paralysis from exposure to the following
chemicals (in order of decreasing potency) 2,5-hexanedione (5/5), 5-hydroxy-2-hexanone (5/5),
2,5-hexanediol (5/5), 2-hexanone (5/5), 2-hexanol (4/5), and high dose n-hexane (3/5). Practical
grade hexane and the lower doses of n-hexane did not produce hind-limb paralysis during the 90-
day testing period. The relative potency of the test chemicals was compared with 2-hexanone.
Specifically, the authors calculated a neurotoxic index (based on the ratio of number of days
until hind-limb paralysis developed in 2-hexanone-treated animals to number of days until hind-
limb paralysis developed following exposure to the test chemical). 2,5-Hexanedione, 5-hydroxy-
2-hexanone, and 2,5-hexanediol had higher neurotoxic indexes than n-hexane and practical grade
hexane. Neurotoxic index correlated with peak serum concentrations of 2,5-hexanedione
produced and the area under the serum concentration-time curve for 2,5-hexanedione.
47
-------
Decreased body weight gain and food consumption was correlated with the neurotoxic index of
each chemical (statistical significance not reported by study authors). All chemicals except 570
and 1140 mg/kg n-hexane produced giant axonal swelling, adaxonal myelin infolding, and
paranodal myelin retraction (incidences not reported by the study authors). The authors noted
that there was an obvious gradation of histopathological response in the frequency of giant axons
between chemicals with the highest neurotoxic index and those with lower neurotoxic indices.
Krasavage et al. (1980) also evaluated testicular tissue by histopathology. The authors
observed various stages of atrophy of the testicular germinal epithelium following the
administration of 2,5-hexanedione, 5-hydroxy-2-hexanone, 2,5-hexanediol, 2-hexanone,
2-hexanol, and the high dose of n-hexane (incidence and severity of effect not reported by the
study authors). The correlation of serum concentration of 2,5-hexanedione with the onset of
neuropathological symptoms suggested that a threshold serum concentration of 50 |ig/mL of
2,5-hexanedione was necessary for the induction of the neuropathological effects. Only those
animals receiving n-hexane at the highest dose (46.2 mmol/kg, equivalent to 3980 mg/kg-day)
built up a serum concentration of 50 |ig/mL 2,5-hexanedione and developed the hind-limb
paralysis characteristic of hydrocarbon-induced neuropathy.
The effects of n-hexane on peripheral nerve transmission were observed by Ono et al.
(1981). Male Wistar rats (5-7/group) were administered n-hexane by gavage in olive oil daily
for 8 weeks. The exposure regimen consisted of administration of 0.4 mL solvent in 0.6 mL
olive oil for the first 4 weeks, 0.6 mL solvent in 0.4 mL olive oil for a subsequent 2 weeks, and
1.2 mL solvent in 0.8 mL olive oil for the final 2 weeks, while a control group received olive oil
alone. Body weight was measured every 2 weeks during the experimental period, resulting in
dose calculations of 811 mg/kg-day (after 2 weeks), 759 mg/kg-day (2-4 weeks), 1047 mg/kg-
day (4-6 weeks), and 2022 mg/kg-day (6-8 weeks). Peripheral nerve activity was measured by
administering a differential pulse to electrodes inserted at different points along the tail of
unanesthetized animals. Transmission of electrical charge was then detected at other points
along the tail. The group mean MCV was measured at the start of the experiment and every 2
weeks until termination.
There was no change among the groups in the rates of body weight gain throughout the
experiment. MCV in groups receiving n-hexane (after 4, 6, and 8 weeks exposure) was reduced
by approximately 5-10% compared with controls. These changes achieved statistical
significance at the 4- and 8-week time point (statistical test not stated). Distal latencies
decreased as the rats grew, but there were no statistically significant differences between
n-hexane-exposed and control animals. However, there were statistically significant reductions
48
-------
in the proximal (approximately 6-8%) and distal (approximately 5-8%) mixed MCVs of animals
receiving n-hexane compared with controls after 4, 6, and 8 weeks, respectively. In general,
such changes were less severe or absent in animals exposed to isomers of n-hexane or
structurally related chemicals (2- and 3-methylpentane, methylcyclopentane). In the absence of
histological studies, the significance of minor MCV changes after exposure to 2-and 3-
methylpentane and methylcyclopentane cannot be fully evaluated. It has been suggested that in
order for an alkane to be neurotoxic it must be metabolized to a gamma diketone.
4.2.1.2. Chronic Studies
No chronic exposure studies were identified that involved the administration of n-hexane
to experimental animals via the oral route.
4.2.2. Inhalation Exposure
4.2.2.1. Subchronic Studies
Ono et al. (1982) examined the subchronic effects of n-hexane on MCV and DL in eight
male Wistar rats/group exposed 12 hours/day for 24 weeks to 200 and 500 ppm n-hexane. Using
a modification to a previously described experimental protocol for assessing nerve conduction
(MCV and mixed nerve conduction velocity) and performance (Ono et al., 1981), the authors
measured MCV, mixed nerve conduction velocity, and DL before the start of the experiment and
then prior to sacrifice (i.e., 0, 4, 8, 12, 16, 20, and 24 weeks exposure). The authors used the
following equation to fit the data to a time-series curve of the seven observations in each rat:
Y = A - C x e"Bt (where Y = data for a particular endpoint over the seven observation periods; A,
B, C = constant; t = weeks of exposure; and e = base of natural logarithm). The statistical
differences in the average of constant A (estimated as a limiting value of the curve) were
evaluated by Student's t-test. One animal from each group was examined histopathologically in
an attempt to link any functional deficits to morphological changes that may have taken place
over the duration of the experiment. The authors stated that they did not observe any definite
clinical signs of neuropathy in any of the exposed groups. Peripheral nerve function was
impaired by exposure to 500 ppm n-hexane and impaired to a lesser extent in rats exposed to 200
ppm n-hexane. Specifically, MCV and distal and total (distal plus proximal) mixed MCV were
statistically significantly decreased at >200ppm (Table 4-7). DL and proximal mixed MCV
were statistically significantly decreased at the low dose but not at the high dose. Degeneration
of the myelinated axons was evident in the peripheral nerves of all exposed groups.
49
-------
Table 4-7. Nerve conduction changes in male Wistar rats exposed 12 hours/day for
24 weeks to 200 or 500 ppm n-hexane
Dose
group
Control
200 ppm
500 ppm
MCV(m/s)
49.6 ±7.6
42.9±3.5a
36.2±l.lb
DL (ms)
1.81 ±0.12
1.91 ±0.10
2.13±0.20b
Total mixed nerve
conduction velocity
(m/s)
59.3 ±4.1
53.3±3.0b
50.9 ± 1.3b
Distal mixed nerve
conduction velocity
(m/s)
52.3 ±3.3
47.6±1.7b
45.3 ± 1.6b
Proximal mixed
nerve conduction
velocity (m/s)
68.6 ±12.5
58.8 ±9.0
5.0±3.2a
Curves: Y = A - C * e"Bt; Y, data for a particular endpoint over the seven observation periods; A, B, C,
constant; t, weeks of exposure; and e, base of natural logarithm.
*p< 0.05
V
-------
and weanlings within 2 weeks of the beginning of exposure. However, subsequent treatments
affected this parameter more in young adults than in weanlings. Older rats exhibited earlier and
more severe signs of hind-limb paralysis compared with younger animals. Conversely, there was
little difference in neuropathological responses between rats in the different age categories,
including tail nerve conduction time and BAEP.
Cavender et al. (1984a, b) exposed F344 rats (15/sex/group) to n-hexane at 0, 3000,
6500, and 10,000 ppm for 6 hours/day, 5 days/week for 13 weeks. There were no n-hexane-
related clinical signs of toxicity, effects on food consumption, ophthalmological findings, or
changes in neurological function. However, there was a lowering of the urinary pH in high-dose
males. There were increased organ/body weight ratios for liver, kidney, and testis in high-dose
males and kidney in mid-dose males. Histopathological examination of the tibial nerves
revealed paranodal axonal swelling in mid- and high-dose males (1/5 and 4/5, respectively).
Male Wistar rats (eight/group) were exposed to 0, 500, 1200, or 3000 ppm n-hexane
(>99% pure) for 12 hours/day, 7 days/week for 16 weeks (Huang et al., 1989). The authors
measured MCV in the tail nerve along with body weight before exposure and after 4, 8, 12, and
16 weeks of exposure to n-hexane. One animal from each group was sacrificed after 16 weeks of
exposure for histopathological evaluation of the nerve fibers in the tail. In addition, Huang et al.
(1989) measured the levels of neuron-specific enolase and P-S100. These nerve-specific
proteins are part of a family of calcium-binding proteins that are involved in processes such as
cell to cell communication, cell growth, intracellular signal transduction, and development and
maintenance of the central nervous system. Some members of the SI 00 protein family are
released into the extracellular space (depending on concentration of protein) by an unknown
mechanism and modulate cell proliferation, act as chemoattractants for leukocytes, stimulate
neuronal survival and/or differentiation of astrocyte proliferation, increase apoptosis of neurons,
and regulate macrophage activation. Data indicates that SI00 proteins may be an extracellular
biomarker for natural aging or damage to the CNS or PNS (i.e., dementia associated with
Alzheimer's and Parkinson's diseases) (Donate, 2001; Donate, 1999; Fano et al., 1995). Dose-
dependent, statistically significant reduction in body weight gain was observed in the mid-dose
(at 12 weeks) and high-dose (at 8 weeks) rats (Huang et al., 1989). Additionally, there were
some neurological deficits in mid- and high-dose rats, including a reduction in grip strength and
a comparative slowness of motion from week 12 of exposure. However, no hind-limb paralysis
was observed by the time of termination of the experiment. Rats exposed to the mid and high
doses of n-hexane showed a reduction in MCV. This reduction was statistically significant
during weeks 8-16 of the exposure period compared with controls. Increased incidence of
51
-------
paranodal swelling, along with some evidence of demyelination and remyelination was present
in the peripheral nerves at both mid and high doses. However, these histopathologic findings
were more severe in the high-dose group. Among the biochemical changes were dose-dependent
reductions in nervous system-specific proteins, particularly the P-S100 protein in tail nerve
fibers, which was significantly reduced by approximately 75% at all dose levels. The
neurophysiologic deficits and histopathologic effects that were evident in mid- and high-dose
rats suggested a NOAEL of 500 ppm.
Huang et al. (1992) exposed male Wistar rats (seven/group) to 0 or 2000 ppm (99% pure)
n-hexane, for 12 hours/day, 6 days/week for 24 weeks. Body weight gain decreased at the fourth
week of exposure. Effects of treatment included an overall reduction (statistically significant) in
MCV after 8 weeks that gradually decreased thereafter. The authors also observed an increase in
DL after 12 weeks that was even greater after 16 weeks. There was a reduction in the activity or
amount of neuron-specific enolase (y-enolase), creatine kinase-B, and the P-S100 protein with
neurophysiologic deficits that were most evident in the distal segment of the sciatic nerve (64,
71, and 76%, respectively). Levels of these nerve-specific markers were positively correlated
with MCV and negatively correlated with DL in the distal sciatic nerve.
The American Petroleum Institute (API) sponsored a number of toxicological studies of
n-hexane in experimental animals, including a 26-week inhalation toxicity study in Sprague-
Dawley rats (Biodynamics, 1978). This study, originally submitted to the EPA under the Toxic
Substances Control Act, featured a complex protocol in which 12 rats/sex/group were exposed to
0, 5, 25, or 125 ppm n-hexane, 6 hours/day, 5 days/week (mean concentrations of 6, 26, and 129
ppm) for up to 34 weeks or for 21 hours/day, 7 days/week (mean concentrations of 5, 27, and
126 ppm) for up to 34 weeks. Neuropathological examinations were carried out on a subset of
each group after 8, 18, 26, 31, and 34 weeks. Hematological and clinical chemistry parameters
were evaluated after rats had been exposed for 13 and 26 weeks. Body weights were monitored
weekly through week 12, then bimonthly until the end of the study.
The authors noted a number of fluctuations in clinical chemistry, including higher fasting
glucose levels in male rats exposed to 5 ppm and 125 ppm n-hexane at 26 weeks and lower
blood urea nitrogen in female rats exposed to 125 ppm n-hexane for the same duration. There
were also fluctuations in hematological parameters, including reductions in hemoglobin
concentration and hematocrit in females exposed at all n-hexane concentrations and durations at
the 13-week interim evaluation. However, these changes showed little relationship to dose,
remained within normal limits, and were not apparent in blood samples taken after 26 weeks.
52
-------
Accordingly, the study authors considered the changes not to be related to treatment. An
addendum to the report concluded that no animal in the study displayed signs of nervous system
degeneration characteristic of n-hexane exposure.
The International Research and Development Corporation (IRDC), sponsored by Phillips
Petroleum Co., continuously exposed male Sprague-Dawley rats to n-hexane and a C6-isomer
mixture consisting of n-hexane, methylcyclopentane, 3-methylpentane, and 2-methylpentane as
major components for 22 hours/day, 7 days/week for 6 months (IRDC, 1992a, b). This study
was conducted in two phases; the exposure groups are shown in Tables 4-8 and 4-10.
Table 4-8. Experimental protocol for Phases I and II of a 6-month inhalation
study of n-hexane and a mixture containing hydrocarbon isomers plus
n-hexane in male Sprague-Dawley rats
Group
I
II
III
IV
V
VI
VII
VIII
IX
X
Treatment
Controls
125 ppm n-Hexane
125 n-Hexane + 125 ppm
125 n-Hexane + 375 ppm
C6 isomersa
C6 isomers
125 ppm n-Hexane + 1375 ppm C6 isomers
500 ppm n-Hexane
Controls
500 ppm C6 isomers
500 ppm n-Hexane + 500
ppm C6 isomers
500 ppm n-Hexane
aC6 isomers were a mixture of n-hexane-depleted C6 hydrocarbons containing methylcyclopentane, 3-
methylpentane, and 2-methylpentane as major components
Source: IRDC, 1992a, b.
In both phases of this study, animals were examined daily for signs of clinical toxicity,
and body weights were monitored weekly. Two controls and four rats from Group VI (see Table
4-8) were taken from their exposure groups every month for the first 5 months. These animals,
plus four from all groups after 6 months of exposure, were examined histopathologically for
changes to the cervical spinal cord. All surviving animals (10/group) were necropsied at term,
and the weights of their major organs were recorded. Excised pieces of tissue from a variety of
organs and tissues were fixed for histopathological examination, including all abnormal masses,
adrenal gland, abdominal aorta, bone marrow, brain, Zymbal's gland, esophagus, epididymis,
53
-------
eye and optic nerve, tongue, Harderian gland, neuroganglia, liver, kidney, lung, lymph nodes,
mammary gland, pancreas, parathyroid, pituitary, prostate, salivary gland, skeletal muscle, skin,
nasal turbinates, gonads, lacrimal gland, heart, thymus, thyroid, peripheral nerve, small intestine,
large intestine, spinal cord, spleen, seminal vesicle, stomach, and urinary bladder.
Rats exposed to 500 ppm n-hexane (Group VI) showed an abnormal gait (10/34). These
symptoms were evident in a single rat after 14 weeks but increased in incidence and severity
throughout the remainder of the experiment. However, the true incidence of this response may
be underestimated because of subjects withdrawn from the study before symptoms could appear.
Some changes in absolute and relative organ weights were considered to be responses to
treatment, including statistically significant increases in the liverbody weight and kidney:body
weight ratios (Table 4-9). Signs of liver necrosis, marked by raised discolored areas on the
organ
surface and, in some cases, necrotic foci when examined histopathologically, were evident in
3/10 rats exposed to 125 ppm n-hexane (Group II) and 2/10 rats exposed to 500 ppm n-hexane
(Group VI). Degenerative and regenerative changes in the kidney were observed in 4/10 rats
exposed to 500 ppm n-hexane (Group VI). The authors described these kidney and liver lesions
as trace to mild severity. Reevaluation of the renal pathology of the kidneys showed a solvent-
induced exacerbation of the hyaline droplet nephropathy that is characteristic for male rats of this
and other strains as reported by Experimental Pathology Laboratories (EPL, 1992).
Table 4-9. Relative organ weights of male Sprague-Dawley rats exposed to n-hexane
22 hours/day, 7 days/week for 6 months
Organ
Spleen
Liver
Kidney
Adrenal
Testis
Heart
Group"
I
0.18 ±0.029
2.88 ±0.26
0.56 ± 0.047
1.01 ±0.23
0.75 ±0.11
0.29 ±0.041
II (125 ppm)
0.15±0.02b
3.01 ±0.27
0.55 ±0.059
0.91 ±0.10
0.67 ±0.08
0.27 ±0.45
VI (500 ppm)
0.21 ±0.12
3.31±0.11C
0.71±0.17C
1.22 ±0.21
0.62 ±0.10
0.27 ±0.021
a Values are percent of body weight ± SD.
b Statistically significant compared with controls (p<0.05).
0 Statistically significant compared with controls (/?<0.01).
54
-------
Source: IRDC, 1992a, b.
Exposure to n-hexane at 500 ppm (Group VI) was associated with axonal degeneration,
myelin vacuolation, and skeletal muscular atrophy. Axonal degeneration occurred in the sciatic
and tibial nerves (0/10, 0/10, and 7/10) and in the spinal cord thoracic, lumbar, and sacral nerves
0/10, 0/10, and 8/10), respectively. Skeletal muscle atrophy was observed in the high-dose
animals (9/10). No other treatment groups exhibited neuropathologic or myopathic changes.
In phase II of the study (Table 4-8), rats (10/group) were exposed to either filtered air
(Group VII), 500 ppm C6 isomers alone (Group VIII), 500 ppm n-hexane plus 500 ppm C6
isomers (Group IX), or 500 ppm n-hexane alone (Group X). An abnormal gait developed in rats
exposed to 500 ppm n-hexane (Group X) at week 16 with increasing incidence and severity over
time. Atrophy of the sciatic and anterior tibial nerves (14/16) and skeletal muscle atrophy (3/10)
were observed in rats exposed to 500 ppm n-hexane. No histopathologic evidence of neuropathy
was seen in controls. The authors noted a slightly increased incidence and severity of chronic
renal nephritis in both controls and n-hexane-exposed rats (6/11 and 10/10, respectively), and a
significant increase in relative kidney weights (0.532 ± 0.053% in controls and 0.844 ± 0.098%
in the n-hexane-exposed group).
The National Toxicology Program (NTP) sponsored a 13-week inhalation study of
n-hexane in B6C3F1 mice, the findings of which have been reported in the peer-reviewed
scientific literature (NTP, 1991; Dunnick et al., 1989). Groups of 10 mice/sex/group were
exposed to 0, 500, 1000, 4000, or 10,000 ppm n-hexane (>99% pure), 6 hours/day, 5 days/week
for 13 weeks, while another group of 10 mice was exposed to n-hexane at 1000 ppm for 22
hours/day, 5 days/week for 13 weeks. Separate groups of eight mice/sex/group received
identical treatments but were subjected to neurobehavioral tests before the start of dosing, then
again after 6 and 13 weeks of exposure. The neurobehavioral tests included undifferentiated
motor activity, fore-limb and hind-limb grip strength, thermal sensitivity, startle response, and
foot splay. Four males and four females were randomly selected from the 0, 1000 ppm extended
duration, and 10,000 ppm exposure groups for histopathological examination of the spinal cord
and tibial nerves. Animals were observed daily for signs of clinical toxicity and weighed
weekly.
A full necropsy was performed at sacrifice, weights of the major organs were recorded,
and histopathological evaluations (in control, high-dose, and extended duration groups) were
carried out at term on a variety of excised organs and tissues, including adrenal gland, brain,
bronchial lymph nodes, cecum, colon, duodenum, esophagus, gall bladder, gross lesions and
tissue masses, heart, ileum, jejunum, kidney, larynx, liver, lung and mainstem bronchi, mammary
55
-------
gland, mandibular and mesenteric lymph nodes, nasal cavity and turbinates, pancreas,
parathyroid, pituitary, rectum, salivary glands, sciatic nerve, spinal cord, spleen, sternum
including marrow, glandular and forestomach, testis with epididymis, prostate, seminal vesicles,
ovary and uterus, thymus, thyroid gland, trachea, and urinary bladder. The mandibular lymph
nodes, nasal cavity, and sternum with marrow were examined histopathologically in all groups.
The liver was examined only in the males of all exposure groups.
All of the animals in the study survived to term, although there were some signs of nasal
irritation among those animals exposed to 10,000 ppm n-hexane. However, no changes in organ
weight to body weight ratio were observed in male or female mice that could be clearly
attributed to exposure to n-hexane. Nonetheless, relative liver, kidney, and heart weights
appeared to be increased compared with controls in exposed females. The only observed
neurobehavioral deficit was a reduction in locomotor activity in females exposed to 10,000 ppm
n-hexane 6 hours/day and to 1000 ppm of n-hexane for 22 hours/day. There was an increased
incidence of paranodal axonal swelling in high-dose or extended exposure duration mice. Table
4-10 summarizes the incidence of neuropathological and respiratory tract lesions and gives the
average grade on a scale of 1-5 for the most severe lesions observed in the eight levels of nasal
cavity. The authors concluded that n-hexane caused minimal toxicity to the nervous system
and/or respiratory system at 1000 ppm and above.
Table 4-10. Incidence of nasal turbinate and neuropathological lesions in B6C3F1
mice exposed to n-hexane for 13 weeks
Site/lesion
Concentration of n-hexane (ppm)
0
500
1000
1000 (ext.)
4000
10,000
Malea
Lumen
Exudate, suppurative
Olfactory epithelium
Chronic active inflammation
Multifocal erosion
Multifocal regeneration
Metaplasia
Respiratory epithelium
Chronic active inflammation
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
2/10 (2)
2/10 (1)
0/10
0/10
0/10
0/10
4/10(1.5)
4/10 (1)
0/10
0/10
0/10
0/10
0/10
1/10(1)
0/10
10/10 (2.3)
10/10(2.1)
8/10 (3)
10/10 (2.2)
10/10 (2.8)
9/10 (1.9)
56
-------
Site/lesion
Multifocal erosion
Multifocal regeneration
Submucosa
Focal fibrosis
Paranodal swellings
Spinal cord
Tibial nerve
Teased fiber
Tibial nerve
Concentration of n-hexane (ppm)
0
0/10
0/10
0/10
0/4
0/4
0/4
500
0/10
0/10
0/10
ND
ND
ND
1000
0/10
0/10
0/10
ND
ND
ND
1000 (ext.)
0/10
0/10
0/10
0/4
0/4
3/4
4000
0/10
0/10
0/10
ND
ND
ND
10,000
2/10(1.5)
10/10 (1.4)
5/10 (1.4)
0/4
3/4
3/4
Female3
Lumen
Exudate, suppurative
Olfactory epithelium
Chronic active inflammation
Multifocal erosion
Multifocal regeneration
Metaplasia
Respiratory epithelium
Chronic active inflammation
Multifocal erosion
Multifocal regeneration
Submucosa
Focal fibrosis
Paranodal swellings
Spinal cord
Tibial nerve
Teased fiber
Tibial nerve
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/4
0/4
0/4
0/10
0/10
0/10
2/10 (1)
0/10
0/10
0/10
0/10
0/10
ND
ND
ND
0/10
0/10
0/10
1/10 (2)
1/10 (2)
1/10 (2)
0/10
0/10
0/10
ND
ND
ND
0/10
0/10
0/10
9/10 (1.7)
8/10 (1.9)
0/10
0/10
0/10
0/10
0/4
0/4
3/4
0/10
0/10
0/10
9/10 (2)
8/10 (2)
0/10
0/10
1/10 (2)
1/10 (1)
ND
ND
ND
10/10 (2.3)
7/10 (1.9)
8/10 (1.7)
10/10 (2.6)
10/10 (2.6)
5/10 (1.4)
0/10
6/10 (1.2)
9/10 (1.6)
0/4
3/4
3/4
a Numbers in parentheses represent the average grade (on a scale of 1-5) of the most severe lesions
observed in eight levels of the nasal cavity examined.
ND = No Data.
Source: Dunnicketal., 1989.
57
-------
A study by Lungarella et al. (1984) examined the effect of n-hexane in 12 male New
Zealand white rabbits/group that had been exposed to either 0 or 3000 ppm n-hexane (purity not
stated) for 8 hours/day, 5 days/week for 24 weeks. Some animals were afforded a 120-day
recovery period before sacrifice, at which point the cellular architecture of the lungs was
examined under the light and electron microscopes. Treatment-related portal-of-entry effects
included an enlargement of the air spaces in respiratory bronchioles and alveolar ducts,
pulmonary fibrosis, and papillary tumors of nonciliated bronchial epithelial cells. However,
some of these lesions were more marked in animals sacrificed immediately after the last
exposure compared with those allowed to recover for 120 days. n-Hexane exposure had no
effect on body weight gain, hematological parameters, or clinical chemistry.
4.2.2.2. Chronic Studies
No chronic studies were identified that examined the toxicological effects of n-hexane in
experimental animals by the inhalation route.
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION
4.3.1. Oral Studies
Marks et al. (1980) conducted a reproductive/developmental and teratological study in
CD-I mice in which dams were exposed to n-hexane (99% pure) in cottonseed oil by gavage on
GDs 6-15. In the first portion of this study, 4-30 mice were dosed with vehicle or increasing
concentrations of n-hexane at 260, 660, 1320, or 2200 mg/kg-day. There were no reproductive,
developmental, or teratological effects of n-hexane in mice dosed with n-hexane observed in this
portion of the study.
In the second portion of the study, 19-26 pregnant mice/group received cumulative doses
of either 0, 1830, 2170, 7920, or 9900 mg/kg-day on GDs 6-15 in the form of three separate
injections spaced throughout the day. All dams were sacrificed on GD 18, and uteri were
examined for the number of implantation sites. Live fetuses were examined for external and
visceral malformations and for skeletal variations. Dam mortality was increased in high-dose
groups. Fetal birth weight was 6.35% lower in the progeny of those dams exposed to 7920 and
9900 mg/kg-day than in controls (0.946 g for mice dosed at 7920 and 9900 mg/kg-day versus
1.011 g in controls). However, neither these nor any other fetuses in the study showed an
increased incidence of skeletal malformations or variations as a result of maternal n-hexane
treatment. The authors suggested that n-hexane is not teratogenic at concentrations associated
58
-------
with overt maternal toxicity.
Linder et al. (1992) included n-hexane in a survey of chemicals for spermatotoxic effects
in male Sprague-Dawley rats. The experimental protocol featured the oral administration of the
undiluted n-hexane either in a single dose of 20,000 mg/kg or in five consecutive daily doses of
10,000 mg/kg. The spermatotoxic tests employed included sperm head counts, sperm velocity,
sperm morphology, and the histopathology of the testis and epididymis. No change was
observed in any of these parameters in rats exposed to n-hexane.
4.3.2. Inhalation Studies
Groups of between three and eight pregnant F344 rats were exposed to air or 1000 ppm
n-hexane, 6 hours/day on GDs 8-12, 12-16, or 8-16 (Bus et al., 1979). Dams were sacrificed on
GD 22 and autopsied to evaluate reproductive parameters such as the number and position of
live, dead, and resorbed fetuses, fetal weight, and the number and type of any fetal defects,
malformations, or skeletal variations. Some pregnant females exposed on GDs 8-16 were
allowed to deliver their progeny, and total litter body weights and mortality were monitored at
weekly intervals up to 7 weeks after birth.
Exposure to n-hexane had no significant effect on the number of fetal resorptions,
external anomalies, soft tissue anomalies, or skeletal variations. Pup growth from dams exposed
on GDs 8-16 was 13.9% lower than controls for up to 3 weeks after birth, but had reached levels
similar to controls after 7 weeks. The authors concluded that n-hexane had little effect on the
reproduction and development of F344 rats.
Litton Bionetics (1979) exposed CRL:COBS CD(SD)BR rats (20 pregnant
females/group) for 6 hours/day to concentrations of 0, 100, or 400 ppm n-hexane on GDs 6-15.
Food consumption and body weights were monitored intermittently between GDs 0 and 20, at
which point the dams were sacrificed and necropsied to facilitate the evaluation of reproductive,
developmental, and teratological parameters. However, no n-hexane-related effects were
observed.
The results of a teratological study of n-hexane conducted in rats on behalf of the NTP
were reported in the published literature by Mast (1987). Mast (1987) exposed pregnant
Sprague-Dawley rats (30/group) to 0, 200, 1000, or 5000 ppm n-hexane (>99.5% pure) for 20
hours/day on GDs 6-19. Maternal toxicity was monitored throughout the experiment. Uterine,
placental, and fetal body weights were measured at sacrifice on GD 20. In addition, the number
of implantation sites and live, dead, and resorbed fetuses was reported. Live fetuses were sexed
and examined for gross, visceral, skeletal, or soft-tissue craniofacial defects.
59
-------
There was a statistically significant reduction in maternal body weight gain in the high
dose dams. There were no intrauterine deaths following exposure to n-hexane. A reduction in
fetal body weight gain was apparent in the progeny of mid- and high-dose dams (3.0-7.5% at the
mid dose and 14-15% at the high dose, respectively). This difference was statistically
significant at both 1000 and 5000 ppm in males and at 5000 ppm in females (3.48 ± 0.37, 3.54 ±
0.36, 3.27 ± 0.32, and 2.97 ± 0.38 grams for the 0, 200, 500, and 1000 ppm exposures,
respectively). The incidences of skeletal variations in individual live fetuses are shown in Table
4-11. There was a statistically significant increase in the mean percent incidence of reduced
ossification of sternebrae 1-4 per litter at 5000 ppm compared with controls (38.7 ± 23.7 versus
13.8 ± 21.6, respectively). Also, this treatment effect was correlated in a statistically significant
manner with exposure concentration. Mast (1987) concluded that the lowest n-hexane
concentration, 200 ppm, would be a NOAEL for developmental toxicity in Sprague-Dawley rats.
Table 4-11. Skeletal variations in live fetuses of pregnant Sprague-Dawley
rats exposed to n-hexane via inhalation
Subjects
Total fetuses examined
Heads examined
Skulls examined
Malformations/Variations
Dilated ureters
Renal pelvic cavitation
Supernumerary ribs
Bent ribs
Reduced Ossification
Sternebrae 1-4
Vertebral centra
Pelvis
Phalanges
Skull
Concentration of n-hexane (ppm)
0
339
170
169
200
350
157
193
1000
392
186
206
5000
408
205
203
Incidence (%)
7.4
2.4
1.2
0.0
6.9
0.0
1.7
0.0
5.1
0.8
3.1
0.3
2.9
0.5
3.7
0.0
12.4
8.3
3.2
1.2
5.9
15.4
4.6
2.0
0.6
3.1
26.3
4.8
5.4
0.3
5.3
38.5
8.8
5.1
1.7
5.9
Source: Mast, 1987.
60
-------
Mast et al. (1988a) also reported the results of an NTP-sponsored reproductive,
developmental and teratological study in mice. Pregnant CD-I mice (30/group) were exposed to
0, 200, 1000, or 5000 ppm n-hexane (99.2% pure) for 20 hours/day on GDs 6-17 (Mast et al.,
1988a). Maternal toxicity was determined at sacrifice on GD 18, and similar reproductive,
developmental, and teratological effects to those described above in Sprague-Dawley rats (Mast,
1987) were examined. Among the dams, there was a statistically significant reduction in body
weight gain and relative uterus weight in the high-dose group. Fetal weights were slightly, but
not significantly, reduced (means of litter means) for all groups compared with controls.
However, this decrease in fetal weights did correlate significantly with dose. Mean female fetal
weights were significantly reduced (approximately 6%) at 5000 ppm and exhibited a significant
correlation to increasing dose. There was also an increase in the mean percent incidence of
supernumerary ribs that appeared to be dose related (not significant). The numbers of live
fetuses per litter were reduced in all exposure groups compared with controls, although the
observed decrease was statistically significant only at 5000 ppm. The numbers of live fetuses
per litter in all exposure groups were also reduced compared with contemporary controls of the
same strain of mice. The mean percent of live implants was reduced compared with
contemporary controls at 200 and 5000 ppm but not at 1000 ppm. The mean percent of
intrauterine death (early and late resorptions combined) was greater for all exposed groups
compared with controls, but this difference was only statistically significant for the 200 ppm
exposure group and there was no dose-related trend. An increased incidence of late resorptions
was also significantly greater than in controls, following exposure to 5000 ppm n-hexane relative
to controls. The author stated that trend analysis indicated that the increased mean percent
incidence of late resorptions was positively and significantly correlated with exposure
concentration.
A number of experimental studies examined the effect of n-hexane on the male
reproductive system when administered via the inhalation route. De Martino et al. (1987)
exposed male Sprague-Dawley rats (12-39/group) to 5000 ppm n-hexane (99% pure) in either
(1) a single 24-hour exposure, (2) repeated 16-hour/day exposures for up to 8 days, or (3)
repeated 16-hour/day exposures, 6 hours/day for up to 6 weeks. The study employed two control
groups, one of which was pair-fed. Treated animals were allowed to recover for different lengths
of time after the end of treatment (from 2 days to 29 weeks, depending on the original exposure
duration). Rats exposed to 5000 ppm n-hexane displayed some evidence of neuropathy such as
paralysis, and extreme cases were sacrificed moribund and necropsied rather than being allowed
to die and undergo partial autolysis. Focal degeneration of spermatocytes and exfoliation of
61
-------
elongated spermatids was observed in rats treated with n-hexane. Early meiotic prophase
spermatocytes (leptotene and zygotene) and transitional spermatocytes as well as those
undergoing meiotic metaphase appeared to be more susceptible to the action of n-hexane than
pachytene spermatocytes. Rats receiving a single 24-hour exposure to n-hexane also showed a
measure of recovery after 2-4 weeks following the termination of exposure. By contrast, rats
exposed repeatedly to 5000 ppm n-hexane over a 6-week period showed complete atrophy of the
seminiferous tubules. The animals exposed for up to 6 weeks displayed a reduction in food
consumption and body weight gain; these effects were accompanied by signs of incipient
neuropathy. There was also a wide range of testicular lesions that did not completely resolve
during the recovery period even though body weights and clinical symptoms improved.
A detailed study by Mast et al. (1988b) examined the effects of n-hexane on sperm
morphology in B6C3F1 mice. The experimental protocol featured the exposure of 20 male
B6C3F1 mice/group to 0, 200, 1000, or 5000 ppm n-hexane, 20 hours/day for 5 consecutive
days, after which the animals were examined for gross lesions of the reproductive tract and any
disturbances to sperm morphology. There were no clinical signs of toxicity or body weight
changes in any of the dose groups, nor was there an increased incidence of aberrant sperm
characteristics such as blunt-hook, banana, amorphous, or pin-head shapes. Likewise, there was
no increased incidence of sperm with more than one head or tail.
Mast et al. (1988c) carried out a further study of the effect of n-hexane in male CD-I
mice in which 20 males/group were exposed to 0, 200, 1000, or 5000 ppm n-hexane, 20
hours/day for 5 consecutive days, then mated to unexposed virgin females. Mated females were
sacrificed 12 days after the last day of cohabitation, and their reproductive status and the
numbers and viability of their implants were recorded to assess the capacity of n-hexane to
induce male dominant lethal effects. The number of live implants was consistently greater than
10 fetuses/litter, and there was no indication of a decline in reproductive index as a result of
increasing n-hexane exposure in the males. Furthermore, there was no increase in the number of
dead implants or early resorptions as a result of the males being exposed to n-hexane prior to
mating. The study authors concluded that short-term exposure to n-hexane vapor did not result
in a male dominant lethal effect in CD-I mice.
62
-------
4.4. OTHER STUDIES
4.4.1. Acute Toxicity Data
Few data are available for a median lethal dose (LD50) for n-hexane. HSDB (2005)
presents values of 28,710, mg/kg, or 24 and 45 mL/kg (approximately 15,840 and 29,700 mg/kg)
for juvenile and adult rats. For inhalation exposure, a 4-hour LC50 of 48,000 ppm has been
reported for both rats and mice (HSDB, 2005).
A number of acute and subacute experimental studies have been carried out to assess
pulmonary toxicity following inhalation exposure to n-hexane. For example, Lungarella et al.
(1980) exposed 12 New Zealand white rabbits/group to 0 or 3000 ppm n-hexane (purity not
stated), 8 hours/day on 8 consecutive days and determined the incidence and severity of
pulmonary lesions by light and electron microscopy. Treated animals showed morphological
signs of parenchymal changes, with lung damage being most severe in the area between the
terminal bronchioles and the alveolar ducts. The lesions consisted of necrotic changes in the
bronchiolar epithelium and desquamation plus an increased number of macrophages within the
proximal alveoli. The architecture of alveolar type II cells also was affected.
The same research group measured the activities of LDH, p-glucuronidase, glucose-6-
phosphate dehydrogenase, and acid and alkaline phosphatases in lung homogenates of New
Zealand white rabbits exposed to 0 or 3000 ppm n-hexane, 8 hours/day for 8 days (Barni-
Comparini et al., 1982). As shown in Tables 4-12 and 4-13, the appearance of blood cells in
bronchial lavage and the concurrent biochemical changes observed in lung homogenates were
consistent with the morphological changes observed at necropsy in the Lungarella et al. (1980)
study. The increased levels of the lysosomal enzyme, acid phosphatase, might reflect acute
inflammation, while the increased activity of glucose-6-phosphate dehydrogenase suggests a
repair process subsequent to n-hexane-induced pulmonary disruption.
Table 4-12. Total red blood cells and nucleated cells in bronchial lavage
from n-hexane-challenged New Zealand white rabbits
Cell counts ( x 107)
Red blood cells
Total nucleated cells
Exposure group (ppm n-hexane)
0
3.5 ±0.26
0.71 ±0.06
3000
8.20±0.75a
2.06 ± 0.28a
1 Significantly different from controls (p<0.01).
63
-------
Source: Barni-Comparinietal., 1982.
Table 4-13. Enzyme activities in lung homogenates of rabbits exposed to
n-hexane
Enzyme activities (IU)
Lactate dehydrogenase
Glucose-6-phosphatedehydrogenase
Acid phosphatase
Alkaline phosphatase
p-Glucuronidase
Exposure group (ppm n-hexane)
0
75 ±3.7
5.10 ±0.29
2.31 ±0.64
1.38±0.15
0.43 ±0.09
3000
109±5.1a
9.98±1.3a
4.96 ± 0.81s
1.91 ±0.19
0.95±0.15a
a Significantly different from controls, as calculated by the authors.
Source: Barni-Comparinietal., 1982.
Sahu et al. (1982) exposed 20 male Sprague-Dawley rats/group to 476, 1149, or 1676
ppm n-hexane (purity not stated) for 6 hours/day, 5 days/week for 4 weeks. The authors
compared the composition of a cell-free supernatant from bronchial lavage in each of the
n-hexane-exposed groups to that from 10 control rats. There was a dose-dependent increase in a
number of enzyme activities and functionally relevant biochemicals (Table 4-14). While no
statistical treatment of the data was provided in the report (control versus test groups), the dose-
dependent increases in biochemical parameters and enzyme activities suggested a disruptive
effect of commercial hexane on one or more cell types in the lung. The threefold increase in
LDH activity in high-dose rats versus controls is consistent with a gross disruption of the plasma
membrane of the respiratory epithelial cells, leading to possible leakage of intracellular contents.
Table 4-14. Concentration of biochemicals and enzyme activities in bronchial
lavage fluid from male Sprague-Dawley rats exposed to n-hexane
Biochemical parameters (units)
Protein (mg/mL)
Lipid (mg/mL)
Sialic acid (ng/mL)
Acid phosphatase (^mol/hour-nig protein)
Alkaline phosphatase (^mol/hour-nig protein)
Lactate dehydrogenase (^mol/min-nig protein)
Concentration of n-hexane (ppm)
0
0.13 ±0.02
0.81 ±0.12
0.09 ± 0.02
0.21 ±0.03
0.37 ±0.05
0.16 ±0.03
476
0.14 ±0.03
1.08 ±0.11
0.12 ±0.02
0.27 ±0.04
0.41 ±0.05
0.19 ±0.02
1149
0.18 ±0.04
1.36 ±0.16
0.16 ±0.03
0.31 ±0.03
0.54 ±0.04
0.26 ± 0.02
1676
0.21 ±0.04
1.74 ±0.11
0.21 ±0.04
0.43 ±0.05
0.68 ±0.06
0.45 ±0.04
64
-------
Glucose-6-phosphate dehydrogenase
(nmol/min-mg protein)
Angiotensin-converting enzyme
(nmol/min-mg protein)
0.90 ±0.08
0.36 ±0.04
0.97 ±0.12
0.67 ± 0.08
1.35 ±0.16
0.86 ±0.07
1.72 ±0.16
1.18±0.12
Source: Sahuetal., 1982.
Ikeda et al. (1986) exposed five male Wistar rats/group continuously for 30 days to either
200 or 400 ppm n-hexane, 200 ppm n-hexane with supplemental toluene at 200 ppm, or 200 and
400 ppm toluene alone. Weight gain was significantly less in the two treatment groups with
n-hexane or toluene alone. When concentrations of norepinephrine and dopamine were
measured in various regions of the brain, exposure at 400 ppm n-hexane resulted in significant
elevations of norepinephrine levels in the thalamus (by 206%), dorsal, olfactory, and frontal
cortex (by 164%, 139%, and 157%, respectively), and cerebellum (by 170%) compared with
controls. Toluene exposure (400 ppm) led to a significant reduction in norepinephrine in the
olfactory cortex (by 82%) and in the hypothalamus (by 81%) and elevation in the ventral cortex
(by 132%). Dopamine levels were reduced in the striatum (by 91%) and elevated in the
olfactory cortex (138%) at 400 ppm toluene. Equivalent changes of this magnitude were not
seen as a result of exposure to mixtures of n-hexane and toluene, although the amount of
norepinephrine in the olfactory cortex was increased by 129%. A similar increase in the amount
of dopamine in the hippocampus was evident in animals exposed to the mixture of solvents.
Pezzoli et al. (1990) injected male VCD-1(ICR)BR mice (number not stated)
intraperitoneally with 400 mg/kg-day n-hexane, 5 days/week for 3 weeks. In addition, male
Sprague-Dawley rats were injected with a single 5 mg dose of n-hexane into the left midbrain
(substantia nigra). This exposure was repeated again 14 days later. The concentrations of
neurotransmitter-related substances of the brain, such as dopamine, homovanillic acid,
norepinephrine, serotonin, and 5-hydroxyindolacetic acid, were evaluated. Levels of dopamine
and homovanillic acid were approximately 30% lower in the brains of mice treated for 3 weeks
with n-hexane when compared with controls. A similar result was observed in response to the
single n-hexane treatment in the left brain. The authors speculated that n-hexane could possibly
alter the dopaminergic pathway; however, the mechanism was unclear.
The same researchers measured the concentrations of dopamine, homovanillic acid, and
2,5-hexanedione in the cerebellum and striatum of mice injected intraperitoneally with 1000
mg/kg n-hexane (Masotto et al., 1995). A more than twofold increase was seen in the
concentrations of 2,5-hexanedione (5.63 ± 0.4 versus 2-3 |ig/g in controls). In contrast to the
results of an earlier study (Pezzoli et al., 1990), striatal synaptosomal dopamine increased in
65
-------
mice treated with n-hexane (1000 mg/kg) either 30 or 60 minutes prior to sacrifice compared
with control mice.
Goel et al. (1982) gave six female albino rats/group daily intraperitoneal injections of
ImL/kg (660 mg/kg) n-hexane for 2 or 7 consecutive days or twice a week for 45 days. Animals
were sacrificed after 2, 7, or 45 days, and liver homogenates were prepared. Alkaline
phosphatase activity was increased in liver homogenates of all n-hexane-treated animals
regardless of duration of exposure (4.15 ± 0.48 nmol/min-mg protein after 2 days, 7.8 ± 1.5
nmol/min-mg protein after 7 days, 4.67 ± 0.58 nmol/min-mg protein after 45 days, versus 1.53 ±
0.19 nmol/min-mg protein in controls), while that of fructose-1,6-diphosphate aldolase
decreased. Some n-hexane-related changes in clinical chemistry parameters were reported,
including decreases in acetylcholinesterase, albumin, and cholesterol (in the latter, 1647 ± 45
mg/L after 2 days, 1393 ± 5 mg/L after 7 days, and 1371 ± 5 mg/L after 45 days versus 1648 ±
29 mg/L in controls). In a later study, the same researchers (Goel et al., 1987) measured 59Fe
uptake in rats receiving n-hexane via the same exposure protocol and found it to be reduced in
the bone marrow compared with controls (Goel et al., 1987). However, most hematological
parameters were unaffected by n-hexane treatment, following either intraperitoneal injection or a
7-day oral exposure regimen. Goel et al. (1988) also used the same experimental approach to
demonstrate n-hexane related reductions in hepatic total sulfhydryl content (by 15% from control
values) and the activity of such mixed function oxidases as aniline hydroxylase (by 11-58%),
benzo(a)pyrene hydroxylase (by 41-60%), and aminopyrine-N-demethylase (by 53-57%).
Bastone et al. (1987) exposed male Sprague-Dawley rats to 0 or 5000 ppm n-hexane, 16
hours/day, 6 days/week for 4 weeks. There was a treatment-related reduction of MCV, an effect
accompanied by a transient increase in plasma acetylcholinesterase activity that declined to
baseline levels 2 weeks after cessation of treatment.
Anderson and Dunham (1984) dosed five male Sprague-Dawley rats/group
intraperitoneally with 540 mg/kg n-hexane or 280 mg/kg 2,5-hexanedione daily for 35 days and
obtained conduction velocities in the sciatic and sural nerves that were slower compared with
controls (Table 4-15). The authors speculated that the electrophysiological changes may have
been related to a disruption of nerve-membrane ATPase activity.
66
-------
Table 4-15. Changes in sciatic and sural nerve action potentials induced by
n-hexane and 2,5-hexanedione
Nerve
Sciatic
Sural
Treatment
Control
n-Hexane
2,5-Hexanedione
Control
n-Hexane
2,5-Hexanedione
Amplitude (mV)
1.31±0.16
0.92 ±0.23
ND
1.04 ±0.07
0.81 ±0.30
0.90 ±0.31
Peak conduction
velocity (m/s)
24.2 ± 0.94
20.2 ± 0.92a
20.2 ± 0.46a
18.5 ±0.61
14.7±0.59a
13.5±0.29a
Duration (ms)
0.60 ± 0.04
0.78 ± 0.07a
1.72±0.11a
0.68 ±0.07
0.92 ± 0.02a
1.97±0.24a
a Statistically significant compared with controls (p<0.05).
ND = No data.
Source: Anderson and Dunham, 1984.
Khedun et al. (1996) exposed male Wistar rats for 30 days by gavage to daily doses of
0.1 mL (66 mg) n-hexane in olive oil. A Langendorff apparatus was used to measure heart rate,
coronary flow, and ventricular fibrillation threshold, the latter parameter showing a dramatic
reduction compared with controls. Concomitant reductions of magnesium and potassium were
noted in the treated group. However, exogenously replacing the magnesium and potassium did
not reverse the lowering of the ventricular fibrillation threshold.
Bio-Research Laboratories (1989) carried out an acute operant behavior study of inhaled
commercial hexane in Sprague-Dawley rats (6/sex/group) receiving a single 6-hour nose-only
exposure at either 0, 873, 2974, or 9187 ppm. Animals were tested in commercial operant
chambers that were fitted with a response lever over a feeder magazine. Prior to commencement
of the study, all animals were given up to seven training sessions of approximately 1 hour each
to learn to press the lever to get the reward (a food pellet). On the day of treatment, rats were
tested for a 30-minute session immediately following treatment. The animals in the study
showed no clinical signs of toxicity, and body weight parameters and food consumption were
unaffected by treatment. Analysis of the response data showed no effect of hexane treatment
immediately after exposure or on posttreatment days 1 or 2 on learned behavior.
4.4.2. Studies with Mixtures Containing n-Hexane
Commercial hexane is a mixture of aliphatic hydrocarbons used as a solvent for
adhesives or to clean machinery. Although the precise amount of each constituent varies,
slightly more than half (about 52%) of commercial hexane consists of n-hexane. The remaining
67
-------
portion is a mixture of isomers and structurally related chemicals of n-hexane, such as 3-
methylpentane (16%), methylcyclopentane (16%), and 2-methylpentane (13%), as well as some
minor components such as cyclohexane and 2,4-dimethylpentane.
4.4.2.1. Oral Exposure
No studies were identified that administered commercial hexane to experimental animals
via the oral route.
4.4.2.2. Inhalation Exposure
4.4.2.2.1. Subchronic studies. Duffy et al. (1991) reported (abstract only) a 13-week inhalation
toxicity study of commercial hexane in F344 rats and B6C3F1 mice, in which
10 animals/sex/group were exposed to target concentrations of 0, 900, 3000, or 9000 ppm
commercial hexane for 6 hours/day, 5 days/week for 13 weeks in a published abstract of the
report by Biodynamics (1989). There were no changes in body weight, food, and water
consumption, no treatment-related mortality at any concentration in either species, and few, if
any, clinical signs of toxicity other than lacrimation in both sexes of high-dose mice and high-
dose female rats. High-dose male and female mice and high-dose male rats displayed an
increase in absolute and relative liver weights. Adverse histopathological findings typical of
hydrocarbon nephropathy were confined to the kidneys of high-dose male rats, as described in
the experimental pathology report of the study (EPL, 1989). All male rats (controls and
exposed) showed some evidence of hyaline droplet formation and related nephropathy.
However, this effect was more severe in male rats exposed to commercial hexane compared with
controls. The kidneys of high-dose males showed mild tubular dilatation, with granular material
in the lumen and signs of epithelial regeneration compared with controls. High-dose males
displayed mild to moderate degrees of epithelial regeneration, a response that was minimal in
controls and in animals receiving commercial hexane at the intermediate concentrations.
Bio-Research Laboratories (1990) conducted a 13-week study of the effects of
commercial hexane in Sprague-Dawley rats (also reported in an abstract by Soiefer et al., 1991).
Twelve rats/sex/group were exposed to 0, 900, 3000, or 9000 ppm commercial hexane for 6
hours/day, 5 days/week for 13 weeks. The animals were evaluated in a functional observational
battery (FOB) approximately 1-2 hours after the first exposure and prior to exposure on days 1,
7, 14, 35, 63, and 91. Motor activity was tested pre-study and on days 34, 62, and 90. Six
animals/sex in the control and high-dose groups were assessed for histological signs of
68
-------
neuropathy. No n-hexane-related effects were observed in either the FOB assessment or
histologically in either sex in any treatment group.
4.4.2.2.2. Chronic exposure. API sponsored two 2-year carcinogenicity studies with
commercial hexane, one in F344 rats (Biodynamics, 1993a) and the other in B6C3F1 mice
(Biodynamics, 1993b). The principal features and key findings of these studies have been
compiled into a single research report that was published in the peer-reviewed literature
(Daughtrey et al., 1999). In each case, 50 animals/sex/group were exposed 6 hours/day, 5
days/week to a commercial hexane preparation at targeted concentrations of 0, 900, 3000, or
9000 ppm. The commercial hexane preparation used in the experiments consisted of 51.5%
n-hexane, 16% methylcyclopentane, 16.1% 3-methylpentane, 12.9% 2-methylpentane, 3.3%
cyclohexane, and trace amounts of other hydrocarbons.
There were no statistically significant differences in survival rates between control and
exposed groups of either sex. Exposed animals showed few clinical signs of toxicity in response
to exposure to commercial hexane other than lacrimation, and there were no n-hexane related
necropsy findings remote from the site-of-entry. Histopathological lesions in the respiratory
passages were noted, especially in the nasal turbinates and larynx. Specific findings consisted of
hyperplasia of epithelial and goblet cells, chronic inflammation, and increased incidence of
intracytoplasmic eosinophilic material in all groups exposed to commercial hexane. Chronic
inflammation was also seen to some extent in controls. Low-, mid-, and high-dose males and
females displayed squamous metaplasia/hyperplasia of the columnar epithelium. This effect was
not seen in controls. No treatment-related histopathological abnormalities in sciatic nerve were
observed in any group of F344 rats exposed to commercial hexane in this study. The
histopathological lesions of the respiratory tract that were evident, even in low-dose rats of both
sexes, suggest that a NOAEL cannot be derived from this study. There was no n-hexane-related
tumor formation at any tissue site in F344 rats.
There were no statistically significant differences in survival between controls and any of
the exposed mice of either sex. There were no differences in clinical signs of toxicity and
ophthalmologic or hematologic effects between the groups, and body weight changes in
commercial hexane-exposed mice were similar to those in controls. There was a statistically
significant, dose-related increase in the incidence of hepatocellular neoplasms in the livers of
high-dose females compared with controls. There was also an increased incidence of pituitary
hyperplasia, adenomas, and adenocarcinomas in exposed females (Table 4-16). Commercial
hexane was associated with decreased severity and incidence of cystic endometrial hyperplasia
69
-------
of the uterus among high-dose females compared with controls.
Table 4-16. Incidence of liver and pituitary tumors in male and female B6C3F1
mice exposed to commercial hexane for 2 years
Target organ / cellular response
Males
Adenomas
Carcinomas
Combined adenomas and carcinomas
Females
Adenomas
Carcinomas
Combined adenomas and carcinomas
Males
Hyperplasia
Adenomas
Adenocarcinomas
Total neoplasms
Females
Hyperplasia
Adenomas
Adenocarcinomas
Total neoplasms
Target concentration of commercial hexane (ppm)
0
900
3000
9000
Liver
10/49
7/49
17/49
4/50
3/50
7/50
5/50
11/50
16/50
6/50
2/50
8/50
7/50
10/50
17/50
4/49
5/49
9/49
10/50
3/50
13/50
10/50
6/50
16/50a-b
Pituitary
0/46
1/46
0/46
1/46
2/45
0/45
0/45
0/45
0/11
0/11
0/11
0/11
4/48
6/48a
0/48
6/48a
0/6
0/6
0/6
0/6
4/48
7/48c
1/48
8/48c
1/46
0/46
0/46
0/46
6/49
5/49a
0/49
5/49a
a Significantly different (p<0.05) from controls, as calculated by the authors using Fisher's Exact test.
b Significant dose-related trend; Cochrane-Armitage test, p<0.05.
0 Significantly different (p<0.0l) from controls, as calculated by the authors using Fisher's Exact test.
Sources: Daughtrey et al., 1999; Biodynamics, 1993b.
4.4.2.2.3. Reproduction/developmental studies. API sponsored two reproductive studies in
laboratory rats and mice exposed to commercial hexane (BRRC, 1989a, b). The first study was a
range-finding study in which pregnant Sprague-Dawley rats (eight/group) and CD-I mice
70
-------
(eight/group) were exposed to commercial hexane for 6 hours/day at target concentrations of 0,
900, 3000, or 9000 ppm on GDs 6-15 (BRRC, 1989a). Pregnant rats were terminated on GD 21,
pregnant mice on GD 18. Maternal body weight gain was monitored intermittently and at
termination. Uterine weights, number of ovarian corpora lutea, implantation sites, and viable
and nonviable implants were evaluated. All live fetuses were weighed, sexed, and examined for
external and visceral malformations and skeletal variations. None of the dams of either species
displayed overt maternal toxicity during the course of the experiment. There appeared to be a
slight increase in body weight gain in the high-dose rats in parallel with increased food and
water consumption in this group. The only sign of reproductive or developmental toxicity was a
reduction in fetal weights per litter in the progeny of pregnant mice exposed to 9000 ppm
commercial hexane. No treatment-related malformations or variations were observed in either
the rat or mouse fetuses.
BRRC (1989b) exposed pregnant Sprague Dawley rats (30/group) to 0, 900, 3000, or
9000 ppm commercial hexane for 6 hours/day on GDs 6-15 and sacrificed the animals on GD
21. Maternal body weights and food and water consumption were recorded on GDs 0, 6, 9, 12,
15, 18, and 21, and the weights of liver, kidney, and uterus were measured at sacrifice. As in the
range finding study, numbers of ovarian corpora lutea, implantation sites, resorptions, and live
and dead fetuses were evaluated. Fetuses were examined for external and visceral abnormalities
and for skeletal variations. There were no treatment-related effects in reproductive,
developmental, or teratological parameters in any of the groups of rats in the study. Among
maternal effects, body weight gain was reduced in high-dose dams and in the mid-dose group for
a portion of the exposure period (GDs 9-12).
In addition, pregnant CD-I mice (30/group) were exposed to the same regimen as that
described for the Sprague-Dawley rats (BRRC, 1989b). There were no n-hexane-related effects
in maternal body weight gain, no changes in food and water consumption, and no other clinical
signs of toxicity among the exposed groups compared with controls. Gestational parameters,
including the numbers of viable and nonviable implantations/litter and sex ratio, were unaffected
by exposure to commercial hexane. However, a degree of maternal toxicity was evident when
the dams were necropsied, as indicated by a dose-dependent increased incidence of discoloration
of the lungs. Dark brown foci were evident in the lungs of 4/29 high-dose and 2/25 mid-dose
dams. A NO AEL of 900 ppm would apply to the maternal effects of commercial hexane based
on this result.
Fetal body weights were unchanged among the groups, and there were no significant
changes in the incidence of individual malformations or pooled external, visceral, or skeletal
71
-------
malformations. However, there were treatment-related increased incidences of two individual
skeletal variations in high-dose pups. Comparing the incidences of these effects between
controls and high-dose groups by litter, the numbers were 0/26 versus 6/26 for bilateral bone
islands at the first lumbar arch and 20/26 versus 26/26 for all unossified intermediate phalanges
(statistically significant at/><0.05, Fisher's Exact test as calculated by the authors). A NOAEL
of 3000 ppm for these skeletal variations was identified.
BRRC (1991) carried out a two-generation reproductive/developmental toxicity study in
which, prior to breeding, 25 Sprague-Dawley rats/sex/group (FO generation) were exposed to
concentrations of 0, 900, 3000, or 9000 ppm commercial hexane for 6-hours/day, 5 days/week
for 10 weeks. The study was published in the peer reviewed literature by Daughtrey et al.
(1994a). Clinical signs of toxicity were monitored daily, and food consumption and body weight
data were recorded weekly. After 10 weeks, males and females were mated and these mating
pairs were exposed to commercial hexane at the same doses for 6 hours/day, 7 days/week for 21
days. Cohabitation was maintained only long enough for pregnancy to be achieved (copulation
plug present). For the dams, exposure was continued through GD 19, discontinued until
postnatal day (PND) 4, then reinstituted until weaning on PND 28, at which point the FO animals
were sacrificed. On PND 4, the pups were culled to 4/sex/litter, then, on PND 28, 25 Fl
rats/sex/group were randomly selected for exposure to commercial hexane for 8-11 weeks.
Subjects were then mated as described for the FO generation. All F2 rats were sacrificed on
PND 28.
Among the reproductive indices evaluated were survival, mating, fertility, gestation, live
births, and lactation. All subjects were necropsied, and excised pieces of liver, kidney, pituitary,
upper and lower respiratory tract, and any obvious lesions were examined histopathologically.
Reproductive organs and tissues taken for histopathology included the vagina, uterus, ovary,
testis, epididymis, seminal vesicles, and prostate.
In the FO generation, there were no dose-related changes in body weight gain and no
clinical signs of toxicity resulting from exposure to commercial hexane at any concentration.
However, hyaline droplet nephropathy was visible histopathologically in the high-dose FO males.
There were no changes in any of the mating indices, fertility, gestation, live pups/litter, or pup
viability at PND 28. A treatment-related effect of commercial hexane was a reduction of mean
body weight in the Fl pups of the high-dose dams, an effect that became apparent at PND 14 and
beyond. The mean body weight of the Fl pups remained lower than controls throughout their
pre-breeding period. The group-specific means were significantly decreased (by approximately
7%) on PND 21 (38.9 ± 4.0 g in high-dose pups versus 41.9 ± 3.95 g in control pups).
72
-------
There were no overt signs of clinical toxicity and no other signs of reproductive
performance deficits in the Fl generation. Similarly, no lesions in male reproductive pathology
were apparent at necropsy and histopathological examination. Hyaline droplet nephropathy was
observed in Fl high-dose males (statistically significant). The number of pups born to exposed
Fl rats were not statistically different compared with controls. F2 pup body weights in the high-
dose group were reduced by 6 to 9% compared with controls. The viability of F2 pups did not
differ between the groups.
IRDC (1986) reported a single generation reproduction/developmental toxicity study in
which Sprague-Dawley rats were exposed to 0, 100, 500, or 1503 ppm commercial hexane via
inhalation for 6 hours/day, 7 days/week. Exposure of both sexes of rat occurred for 100 days
prior to mating, through the mating period (maximum of 15 days), through GDs 1-20, and then
postnatally through weaning (PND 21).
There were some statistically significant reductions in body weight gain among the
groups, most notably in the FO females exposed to 1503 ppm. Fetal weights were reduced in the
Fl pups, especially in high-dose progeny on lactation day 4, where the reduction from control
levels was 11-13%. The body weight of high-dose Fl pups remained 8-9% lower than controls
throughout lactation. Similar body weight reductions compared with controls were also
observed in the mid-dose group throughout lactation, where the reductions were 12-17% from
control values at their greatest extent and achieved statistical significance on PNDs 14 and 21.
There were no changes in organ weight and no teratological effects in fetuses in any of the
treatment groups. Study authors considered the changes in pup weight to be of no biological
significance and assigned a NOAEL of 1503 ppm to the study.
4.4.3. Potentiation and Antagonism Studies
Altenkirch et al. (1978) exposed 22 male Wistar rats/group to 0 or 10,000 ppm n-hexane,
8900 ppm n-hexane mixed with 1100 ppm methyl ethyl ketone, or 6000 ppm methyl ethyl ketone
alone for 8 hours/day, 7 days/week for up to 19 weeks. The group exposed to methyl ethyl
ketone alone had originally been exposed to 10,000 ppm of this solvent. However, the initial
concentration had to be reduced to 6000 ppm after a few days because of severe irritation of the
upper respiratory tract. All animals exposed to solvent showed immediate clinical signs of
toxicity, such as excitation, ataxia, impaired gait, and drowsiness. The effects were more
prominent in the group exposed to the solvent mixture. Motor deficits occurred in solvent-
exposed animals, characterized by an eversion of the hind-limbs. Rats with a severe paresis
could only crawl across the floor of the cage or not move at all. These deficits occurred earlier,
73
-------
and their extent was more severe in rats exposed to n-hexane mixed with methyl ethyl ketone.
While rats exposed to pure methyl ethyl ketone did not develop any obvious motor impairment
up to week 7, all animals in this exposure group died during the eighth week. The authors
suggested that the cause of this increase in mortality was bronchopneumonia.
Serial necropsies and histopathological examinations of the peripheral nerves were
carried out on subsets of exposed rats throughout the course of the experiment. Changes, such as
multifocal, paranodal swelling of giant axons of the tibial nerves, were detectable in rats exposed
to the n-hexane/methyl ethyl ketone mixture during week 4 of exposure. However, rats exposed
to n-hexane alone did not develop such manifestations of peripheral nerve damage until week 8.
The authors described similar changes in the spinal cord in the long descending tracts at distal
sites and in the long ascending tracts at proximal sites near the medulla oblongata. Exposure to
methyl ethyl ketone alone (6000 ppm, 8 hours/day for 7 weeks) did not induce comparable
histopathological changes. Neuropathological changes were more severe and occurred earlier in
animals exposed to n-hexane mixed with methyl ethyl ketone than to n-hexane alone.
Altenkirch et al. (1982) exposed male Wistar rats (five/group) to 0, 500, or 700 ppm
n-hexane, 300 ppm n-hexane plus 200 ppm methyl ethyl ketone, 400 ppm n-hexane plus 100
ppm methyl ethyl ketone, and 500 ppm n-hexane plus 200 ppm methyl ethyl ketone for 24
hours/day, 7 days/week for up to 9 weeks. Animals were observed for clinical signs of toxicity
over the course of the experiment, and histopathological examinations of excised brain, spinal
cord, and peripheral nerves were performed at term. All exposed rats survived to term, although
some groups showed a reduction in body weight gain during the lifetime of the experiment.
Clinical signs included excessive salivation and an increase in paralysis of the hind-limbs. This
condition was thought to be indicative of peripheral neuropathy. The time for this condition to
develop was shorter in those rats exposed to the higher concentrations of n-hexane and to the
mixtures. Histopathological examinations of the peripheral nerves showed the presence of
axonal swellings, especially at the branches of the tibial and ischiatic nerves. A breakdown of
axons and myelin developed distally to the axonal swellings, with an apparent intra-axonal
accumulation of NFs. Other morphological findings included axonal swelling of the gracile tract
of the spinal cord, especially at the level of the gracile nucleus in the medulla oblongata.
A second phase of the experiment featured the exposure of male Wistar rats (five/group)
to 700 ppm n-hexane or 500 ppm n-hexane plus 200 ppm methyl ethyl ketone for 8 hours/day,
7 days/week for 40 weeks. These animals displayed neither the clinical signs of n-hexane-
induced peripheral neuropathy nor the axonal swelling and peripheral nerve fiber degeneration
that marked the histopathological responses in those animals exposed continuously for 9 weeks.
74
-------
After 40 weeks there was some evidence of nerve fiber destruction in all groups, including
controls, changes that were interpreted by the authors as being age-related. Altenkirch et al.
(1982) concluded that male Wistar rats exposed 8 hours/day for 40 weeks to either 700 ppm
n-hexane or 500 ppm n-hexane plus 200 ppm methyl ethyl ketone developed no
neuropathological or clinical signs of neuropathy. This contrasted with rats exposed to the same
concentrations for 24 hours/day in the first phase of the study. The rats in the first phase of the
study developed clinical neuropathy after 4 weeks.
Ichihara et al. (1998) carried out a series of studies on the toxicological interactions of
n-hexane and methyl ethyl ketone that was intended to resolve the apparent contradiction
between the potentiating effects of methyl ethyl ketone on n-hexane-induced neurotoxicity and
the reduced urinary levels of 2,5-hexanedione that had been observed as a result of coexposure to
methyl ethyl ketone (van Engelen et al., 1997; Shibata, 1990a, b; Altenkirch et al., 1978). Eight
male Wistar rats/group were exposed 12 hours/day, 6 days/week for 20 weeks to either filtered
air (controls), 2000 ppm n-hexane (96% purity), 2000 ppm n-hexane plus 200 ppm methyl ethyl
ketone, or 2000 ppm n-hexane plus 2000 ppm methyl ethyl ketone. MCV, DL, and urinary
2,5-hexanedione were measured every 4 weeks. A several-fold decrease in MCV and an
approximate 50% increase in DL were reported for those rats exposed to 2000 ppm methyl ethyl
ketone and 2000 ppm n-hexane compared with those exposed to 2000 ppm n-hexane alone.
These changes (decrease in MCV and increase in DL) were greater than those induced by
n-hexane alone or by a mixture of 2000 ppm n-hexane and 200 ppm methyl ethyl ketone.
Changes in urinary 2,5-hexanedione were biphasic. On the first day of exposure, coexposure
with methyl ethyl ketone decreased urinary levels of 2,5-hexanedione compared with the levels
obtained in rats exposed to 2000 ppm n-hexane. However, the urinary level of 2,5-hexanedione
in rats exposed to 2000 ppm n-hexane mixed with 2000 ppm methyl ethyl ketone increased after
4 weeks and reached twice the level seen in rats exposed to 2000 ppm n-hexane alone.
Eight male Wistar rats/group were exposed to either 100 ppm n-hexane, 100 ppm
n-hexane plus 200 ppm methyl ethyl ketone, or 200 ppm methyl ethyl ketone alone for 12
hours/day for 24 weeks (Takeuchi et al., 1983). MCV and mixed MCV were similar among
n-hexane-exposed, methyl ethyl ketone-exposed, and controls. Small, though statistically
significant, reductions in both MCV and mixed MCV were detected at various time points
during the exposure period in those rats exposed to the mixture of solvents, as compared with
rats exposed to n-hexane or methyl ethyl ketone alone. There was little change in DL among the
exposure groups (both single solvent and mixtures of solvents) throughout the experiment.
However, the results suggested that methyl ethyl ketone enhanced the subclinical
75
-------
neurophysiological effects of n-hexane at comparatively low concentrations.
Veronesi et al. (1984) assessed the neurotoxicity of combinations of n-hexane and methyl
ethyl ketone in a tissue culture system in which explanted fetal mouse spinal cord with attached
dorsal root ganglia and striated muscle were cultured for up to 56 days in media supplemented
with various solvents. Eighteen cultures/dose/solvent were examined under light and electron
microscope twice a week. Veronesi et al. (1984) developed a time-to-onset metric based on the
incubation time (in days) necessary for axonal swelling to appear in 75% of the cultures.
Cultures (eight/dose) were exposed to n-hexane in the medium at 0, 25, 50, 80, 100, or 250
|ig/mL or to methyl ethyl ketone at 0, 10, 25, 50, 200, 300, 400, or 600 |ig/mL. Cultures with
single solvent exposures were incubated for up to 49 days. Cultures with mixtures of n-hexane
(0, 25, 50, 100, 250 |ig/mL) plus methyl ethyl ketone (0, 10, 25, 50, 100 |ig/mL) were incubated
for up to 56 days.
The authors reported the development of axonal swelling, retraction of paranodal myelin,
accumulation of NFs, and peripheral displacement of neurotubules and mitochondria at n-hexane
concentrations of 100 and 250 |ig/mL. Times-to-onset for n-hexane concentrations of 100 and
250 |ig/mL were 43 and 28 days, respectively. Veronesi et al. (1984) described the cultures
incubated in methyl ethyl ketone alone as being marked by generalized cellular breakdown at the
highest dose (600 |ig/mL). Other pancytotoxic responses included intraaxonal rectilinear
inclusions that developed in several cultures treated with 200-400 |ig/mL methyl ethyl ketone.
As shown in Table 4-17, the presence of methyl ethyl ketone in neurotoxic
concentrations of n-hexane in this system reduced the time-to-onset as compared with equivalent
incubations containing n-hexane alone. Addition of methyl ethyl ketone to nonneurotoxic
concentrations of n-hexane-induced an apparently neurotoxic response.
Table 4-17. Time-to-onset for the appearance of axonal swelling in explanted
cultures of fetal mouse spinal cord incubated with mixtures of n-hexane and
methyl ethyl ketone
n-Hexane
(jig/mL)
0
25
50
100
Methyl ethyl ketone (jig/mL)a
0
NC
NC
NC
43
10
Not tested
Not tested
14
31
25
Not tested
Not tested
Not tested
Not tested
50
NC
11
12
19
100
NC
22
18
25
76
-------
250
28
Not tested
21
27
Not tested
a All values in days.
NC = No pathological change.
Source: Veronesietal., 1984.
Takeuchi et al. (1981) exposed eight male Wistar rats/group to ambient air, 1000 ppm
n-hexane, 1000 ppm toluene, and 1000 ppm n-hexane mixed with 1000 ppm toluene for 12
hours/day for 16 weeks. Electrophysiological parameters such as mixed nerve conduction
velocity, MCV, and DL were measured at the start of the experiment, after 4, 8, 12, and 16
weeks of exposure, and 4 weeks after exposure was discontinued. Marked reduction in
conduction velocities and increased DLs in the tail nerves of rats were observed following
exposure to n-hexane alone. There was slight neuropathologic impairment following exposure
to the mixture of n-hexane and toluene, while toluene exposure alone had no effect.
Ikeda et al. (1993) investigated the effects associated with exposure to either n-hexane,
toluene, or a mixture of the two solvents in Wistar rats. The rats were initially taught to respond
to a pre-signaled light flash in order to avoid an electric shock (by pressing a lever). Rats that
had an avoidance rate of over 80% (18 rats) were selected for the study and divided into three
groups of six each (toluene only, n-hexane only, and mixture of toluene and n-hexane). Controls
were exposed to air. Each exposure group was first exposed to air (as an internal control) for
one hour and then in sequence to 50, 100, 200, 400, or 800 ppm (in ascending order) of each
solvent individually or as a 50:50 mixture. The interval between exposures for each rat was 14
days and sham exposure to air was carried out every seventh day following exposure to solvent.
Shock-avoidance behavior was monitored during and after (up to an hour) exposure, and the
effects of each organic solvent were evaluated by comparing the performance in individual rats
during and after exposure with their own performance under sham exposure (i.e., to ambient air).
The highest exposure concentration of n-hexane appeared to induce a consistent increase in the
lever press rate. By contrast, the 800 ppm mixture (400 ppm n-hexane and 400 ppm toluene)
decreased lever press and avoidance rates when compared with baseline behavior.
Nylen et al. (1994) and Nylen and Hagman (1994) compared the performance of the
auditory and visual systems in rats after exposure to n-hexane or combined exposure to n-hexane
combined with either toluene or xylene. In the study involving toluene, Nylen et al. (1994)
exposed male Sprague-Dawley rats to either 1000 ppm n-hexane, 1000 ppm toluene, or a
mixture of 1000 ppm n-hexane and 1000 ppm toluene for 21 hours/day, 7 days/week for 28 days.
The auditory and visual sensitivity of the animals was measured as their BAEPs and VEPs,
77
-------
which were recorded 2 days, 3 months, and 1 year after the completion of exposure. Conduction
velocities in the nerves of the tail and compound nerve and action potentials were also compared
between controls and exposed groups. Changes in auditory brainstem responsiveness to click-
evoked stimuli were not observed in the rats exposed to n-hexane alone but were detected 2 days
after the conclusion of the exposure regimen in toluene-exposed animals and in those inhaling
the mixture of solvents. The latter group of animals showed an enhanced loss of auditory
sensitivity compared with other groups 3 months following termination of exposure. Exposure
of rats to n-hexane alone was associated with the lowering of one amplitude in the flash-evoked
potential 2 days after exposure. This change was not observed in rats exposed to the mixture of
n-hexane and toluene. There was little change in the peripheral MCV in rats 2 days and 3
months after exposure to a mixture of n-hexane and toluene, as compared with a marked
decrease in MCV in rats that had been exposed to n-hexane alone (the MCVs at the 2-day time
point were 9.2 ±1.7 m/s in n-hexane receiving animals, 15.4 ± 2.0 m/s in unexposed controls,
and 12.8 ±1.5 m/s in rats exposed to mixed n-hexane and toluene).
Nylen and Hagman (1994) exposed male Sprague-Dawley rats to 1000 ppm n-hexane,
1000 ppm xylene, or a mixture of 1000 ppm n-hexane and 1000 ppm xylene for 18 hours/day,
7 days/week for 61 days. The same neurophysiological measurements as those described by
Nylen et al. (1994) were carried out 2 days, 4 months, and 10 months after the conclusion of the
dosing regimen. For the BAEPs, exposure to the n-hexane:xylene mixture caused a persistent
loss of auditory sensitivity that, compared with controls, was statistically significant in the
7-17 dB range. For the VEPs, the latencies of the flash-evoked potentials were prolonged in the
n-hexane exposure group versus controls at the 2-day postexposure time point. Exposure to
n-hexane alone markedly reduced nerve conduction velocity, while the mixture of n-hexane and
xylene did not have much effect on this parameter. The MCVs at the 2-day time point were
12.2 ± 3.8 m/s in n-hexane receiving animals and 17.5 ± 4.0 m/s in rats exposed to mixed
n-hexane and xylene compared with 21.2 ± 2.3 m/s in unexposed controls.
Ralston et al. (1985) used an oral exposure regimen to investigate the possible
potentiation of 2,5-hexanedione-induced neurotoxicity by methyl ethyl ketone. The mixture was
administered via gavage at a concentration of 2.2 mmol/kg-day, 5 days/week for up to 90 days.
Exposure to the mixture caused a rapid onset of motor deficits in male F344 rats compared with
exposure to either chemical alone. Urinary clearance of 2,5-hexanedione was reduced, and the
area under the blood concentration time course was increased in the presence of methyl ethyl
ketone. This suggests that methyl ethyl ketone potentiates 2,5-hexanedione-induced
neurotoxicity by increasing the persistence of 2,5-hexanedione in the circulation.
78
-------
In addition to toluene, xylene, and methyl ethyl ketone, a variety of studies have shown
that acetone may affect n-hexane metabolism and potentiate n-hexane induced neurotoxicity
(Cardona et al., 1996; Ladefoged et al., 1994, 1989; Larsen et al., 1991). Specifically, these
studies have evaluated the neurotoxic effect of acetone co-exposure with the n-hexane metabolite
2,5-hexanedione.
Ladefoged et al. (1989) exposed male Wistar rats (six/group) for up to 6 weeks to either
0.5% 2,5-hexanedione, 0.5% acetone, 5.0% ethanol, 0.5% 2,5-hexanedione plus 0.5% acetone,
or 0.5% 2,5-hexanedione plus 5.0% ethanol in drinking water (w/w). Food and water
consumption were measured weekly, peripheral MCV was measured weekly from the third week
of exposure, and neurobehavioral toxicity (balance time in 30 second intervals on a moving rod)
was measured weekly. Body weight gain and water consumption were statistically significantly
reduced after two weeks following treatment with 2,5-hexanedione, 2,5-hexanedione plus
acetone, and 2,5-hexanedione plus ethanol compared with controls administered pure water.
Water consumption was also reduced in the ethanol-only exposure group following the first
week of dosing compared with controls. Neurophysiologically, rats exposed to 2,5-hexanedione
and 2,5-hexanedione plus acetone had statistically significantly reduced MCV beginning at 3
weeks exposure duration compared with controls. Acetone-only-exposed rats showed a
statistically significant reduction at 6 weeks exposure. Ethanol exposure alone did not produce
any significant changes in MCV, but coexposure with 2,5-hexanedione significantly reduced
MCV at 3 weeks exposure duration compared with controls. MCVs measured following
exposure to 2,5-hexanedione plus ethanol were greater than after exposure to 2,5-hexanedione
alone. In addition, 2,5-hexanedione plus acetone led to a greater reduction in MCV compared
with 2,5-hexanedione plus ethanol (statistically significant at week 4). Table 4-18 presents
results of this study. Acetone and ethanol alone did not affect balance time at any duration of
exposure.
Table 4-18. Effect of 2,5-hexanedione, acetone, ethanol, and mixtures of
2,5-hexanedione with acetone or ethanol in drinking water on average MCV
Dosing
week
3
4
5
Control3
29.6 (3.0)
30.3 (2.5)
29.3 (1.1)
0.5%
2,5-Hexanedionea
26.1 (2.4)b
25.8 (2.3)c
25.3 (2.9)b
0.5%
Acetone3
28.0 (2.9)
28.6(1.8)
28.8(1.2)
0.5%
2,5-Hexanedione
plus acetone3
24.8 (2.9)b
23.5 (1.8)c'e
22.3 (3.2)d
5.0%
Ethanol 3
29.2 (2.0)
29.6(1.4)
31.2(0.9)d
0.5%
2,5-Hexanedione
plus 5.0%
ethanol3
25.1(1.6)c
26.1 (2.6)b
27.1 (1.7)d
79
-------
I 6 | 31.5(2.1) | 25.1(1.8)c | 29.5 (l.l)d | 23.0 (2.8)c'e | 32.5(1.1) | 26.8 (1.7)c |
a Values are means in m/sec; numbers in parentheses are SDs.
>< 0.01; cp< 0.001; and Ap< 0.05.
e Significantly different from group receiving 0.5% 2,5-hexanedione.
Source: Ladefogedetal., 1989.
2,5-Hexandione alone reduced balance time in rats exposed for 3 and 4 weeks only.
Acetone plus 2,5-hexanedione reduced balance time to significantly lower levels than in controls
from the second week of dosing until the end of the study (6 weeks) and to significantly lower
levels than in 2,5-hexanedione-only-treated animals from the fourth week of dosing onward. A
reduction in balance time was also observed in the 2,5-hexanedione plus ethanol group compared
with controls, but there was no significant difference compared with rats given 2,5-hexanedione
alone.
Ladefoged et al. (1994) exposed male rats (20/group) to 0.5% 2,5-hexanedione, 0.5%
acetone, and 0.5% 2,5-hexanedione plus 0.5% acetone in drinking water for 6 weeks. Tap water
was given to controls. Once a week body weight and food and water consumption were
measured. In addition, the following behavioral indicators were also monitored weekly:
ambulation (crossing in an open field for 5 minutes), rearing (number of times both fore-legs
were raised from the floor in 5 minutes), balance on an accelerating rotarod, and grip strength of
fore-limbs and hind-limbs. After 6 weeks exposure, half of the rats were subjected to
histopathological analysis of nerve fibers. The other half of the rats were allowed a 10-week
recovery period followed by histological analysis of nerve fibers. Food and water consumption
and body weights were decreased in both the 2,5-hexanedione alone and 2,5-hexanedione plus
acetone groups during the entire 6 weeks of the study. Water consumption returned to normal in
the first 4 weeks of the recovery period for these rats. Body weight decrease was most
pronounced during the dosing period but remained statistically significant during the recovery
period. Acetone alone had no effect on the behavioral parameters observed. The authors stated
that the neurotoxicity of 2,5-hexanedione and 2,5-hexanedione plus acetone was demonstrated
by statistically significant changes in the performance of the dosed rats compared with controls
in the behavioral tests (data presented graphically for ambulation, rearing and balance; grip
strength data were not shown). In addition, the behavioral effects were more pronounced in the
animals dosed with both 2,5-hexanedione and acetone compared with rats dosed with
2,5-hexanedione alone. The reduction in ambulation following both exposures was reversed
within 5 weeks of the 10-week recovery period. Effects on rearing and balance on a rotarod
were reversible within 10 weeks recovery for the rats treated with only 2,5-hexanedione but not
80
-------
for rats treated with 2,5-hexanedione and acetone. The reduction in grip strength was not
reversible for either group during the 10-week recovery period.
Following 6 weeks of exposure to 2,5-hexanedione and 2,5-hexanedione plus acetone,
there was statistically significant increased giant axonal swelling in the sciatic nerve. The
median percent relative distribution of fiber area of class 9 fibers was 0.4 (0.0-0.5) and 1.6
(0.0-2.1), and that of class 10 fibers was 0.2 (0.0-0.5) and 0.9 (0.0-2.1) compared with controls
(class 9, 0.0 [0.0-0.5] and class 10, 0.0 [0.0-0.3]), respectively. The same pattern was observed
in the tibial nerve fibers. The authors stated that structural changes observed immediately after
the 6-week exposure period were greater in rats exposed to the solvent mixture compared with
those exposed to 2,5-hexanedione only. After 10 weeks recovery, the nerve tissue (sciatic and
tibial) appeared normal. The neurotoxicity observed in this study was similar to that seen
following exposure to n-hexane alone (Section 4.2.2.1).
Acetone coexposure has also been shown to affect male reproductive toxicity of
n-hexane. Larsen et al. (1991) exposed male rats (10/group) to 0, 0.13, 0.25, or 0,5%
2,5-hexanedione alone or in combination with 0.5% acetone (weight/volume) for 6 weeks in
drinking water. At week 5 of exposure, one-half of the rats were mated with unexposed females,
and the numbers of matings, pregnancies, and fetuses were recorded. The other half of the rats
were allowed a 10-week recovery period followed by mating and analysis of the same
reproductive parameters. Testis weight and morphology were also evaluated. Water and food
intake was reduced in rats receiving 2,5-hexanedione alone (dose dependent) and was slightly
further decreased with coexposure to acetone. The authors calculated average intake based on
water consumption to be 170, 270, and 440 mg/kg-day 2,5-hexanedione. Rats exposed to
2,5-hexanedione alone displayed a dose-dependent decrease in body weight gain, which was
greater with acetone exposure. Following the dosing period, food and water intake of all dosed
groups returned to control levels except in rats exposed to the high dose of 2,5-hexanedione
alone or the high dose of 2,5-hexanedione plus acetone. Body weight remained significantly
lower from the third week of dosing until the end of the study. Actual food and water intake and
body weights were not reported. The number of matings were not affected in any of the
exposure groups. A statistically significant reduction in testis weight and the number of
pregnancies and fetuses was observed in rats exposed to 0.5% 2,5-hexanedione alone and 0.25
and 0.5% 2,5-hexanedione plus acetone after 6 weeks. The highest combined treatment resulted
in infertility in the male rats. After the 10-week recovery period, the effects on the testis and
fertility persisted in the high-dose 2,5-hexanedione-only group and in the group coexposed to
acetone. In addition, following the recovery period, testicular atrophy and reduced testis tubuli
81
-------
diameters were present in all dose groups except acetone alone exposure. The authors stated that
acetone potentiated the effects of 2,5-hexanedione on testis.
4.4.4. Mode of Action Studies
Ultrastructural studies indicate that nervous system toxicity induced by n-hexane may be
the result of a sequence of events leading to degeneration of the axons (Spencer and
Schaumburg, 1977a; Schaumburg and Spencer, 1976). Sprague-Dawley rats (eight animals
total) were exposed continuously to 400-600 ppm n-hexane for 1-23 weeks. Three additional
animals received subcutaneous injections of 650-2000 mg/kg 5 days a week for up to 35 weeks.
The animals were observed for clinical signs of neuropathy (characterized by waddling gait,
hind-limb paralysis, and decreased ability to grip a rotating bar). Sciatic, tibial, and plantar
nerves were subjected to light microscopy. The authors described focal condensation of NF,
mitochondria, and smooth endoplasmic reticulum with increase in the number of NFs. The
earliest pathological indicator of peripheral nerve axonal degeneration was axonal swelling in
the distal nonterminal region of the large myelinated fibers. These axonal swellings appeared
first proximal to the nodes of Ranvier and ascended the nerve with further exposure (i.e., facing
paranodes and internodal loci). Paranodal swelling was accompanied by shrinkage and
corrugation of the adjacent distal internode. Paranodal myelin sheaths split and retracted leaving
giant axonal swellings near the nodes of Ranvier. The study authors suggested that Schwann
cells may become associated with these denuded regions and remyelinate short segments.
Remyelinated segments then mark the position of the axonal swellings that were resolved
without fiber breakdown or total internodal demyelination.
Several studies suggest that the n-hexane metabolite, 2,5-hexanedione, is the primary
toxic agent by which n-hexane brings about its neurotoxicological effects. Ladefoged et al.
(1989) exposed male Wistar rats (1 I/group) to 0, 0.5% 2,5-hexanedione in drinking water for 6
weeks. The rats were evaluated for neurobehavioral and nervous system toxicity by the rotarod
performance and measurement of MCV, respectively. MCV was significantly reduced after 3, 4,
and 5 weeks of exposure (Table 4-19). Rotarod performance was significantly reduced
(decreased average balance time) after 3 and 4 weeks exposure. In a follow-up study, Ladefoged
et al. (1994) exposed male Wistar rats (20/group) to 0 or 0.5% 2,5-hexanedione in drinking water
for 6 weeks. Statistically significant reductions in performance in neurobehavioral tests
(ambulation and rearing, rotarod, and grip strength) were noted after 3 weeks exposure to
2,5-hexanedione. In addition, the authors observed giant axonal swelling in the tibial and sciatic
nerve fibers after 6 weeks exposure to 2,5-hexanedione.
82
-------
Schaumburg and Spencer (1978) showed the rapid onset of distal axonal degeneration in
cats administered aqueous 2,5-hexanedione at low-levels (concentration and route of
administration not stated) for 60-75 days. Other symptoms typical of n-hexane-induced
peripheral neuropathy included progressive symmetrical weakness in all extremities, resulting in
paralysis. Schaumburg and Spencer (1978) demonstrated that 2,5-hexanedione also caused
widespread axonal degeneration in the mammillary body, lateral geniculate nucleus, and superior
colliculus in exposed cats. These lesions were thought to be further examples of the distal
axonopathy seen elsewhere in the peripheral and central nervous systems in humans and animals
exposed to n-hexane.
Krasavage et al. (1980) studied the relative neurotoxicity of n-hexane and its metabolites
by administering equimolar doses of each chemical by gavage to five male COBS, CD(SD) BR
rats/group for 5 days/week for 90 days (section 4.2.1.1). As judged by the time taken for
neuropathological symptoms to develop, the parent chemical and its metabolites could be ranked
in descending order of neurotoxicity as follows: 2,5-hexanedione, 5-hydroxy-2-hexane,
2,5-hexanediol, 2-hexanone, 2-hexanol, n-hexane, and practical grade hexane. 2,5-hexanedione
had approximately 38 times the neurotoxic potency of n-hexane itself on an equimolar basis
(Couri and Milks, 1982; Krasavage et al., 1980). Abou-Donia et al. (1982) observed a similar
comparative neurotoxic relationship when n-hexane, 2-hexanone, 2,5-hexanediol, and
2,5-hexanedione were administered orally or intraperitoneally to hens. Pathological examination
of treated birds showed giant paranodal axonal swelling followed by degeneration of axons and
myelin in peripheral nerves and the spinal cord. Based on the time of onset of these symptoms,
the magnitude of the clinical signs of toxicity, and the severity of the histopathological lesions,
the relative neurotoxicity of the subject chemicals was, in descending order, 2,5-hexanedione,
2,5-hexanediol, 2-hexanone, and n-hexane.
Nachtman and Couri (1984) carried out an electrophysiological study to evaluate the
comparative neurotoxicity of 2-hexanone and 2,5-hexanedione in rats. Male Wistar rats were
exposed to the chemicals at concentrations of 20 and 40 nmol/L in drinking water. Motor nerve
velocities and latencies were determined at three stimulus sites, the sciatic notch, the popliteal
space, and the plantaris tendon. Distal latency was significantly greater in animals exposed to
2,5-hexanedione (2 weeks at 40 nmol/L) than in those receiving 2-hexanone for the same
duration. 2,5-hexanedione was also shown to induce neurobehavioral deficits in male F344 rats
exposed to the chemical by gavage at a dose rate of 2.2 mmol/kg-day for 90 days (Ralston et al.,
1985). Compared with controls, 2,5-hexanedione-exposed animals performed progressively
worse in the hind-limb grasp and hind-limb place reflex tests and the balance beam and
83
-------
accelerating rotorod functional tests. Similar deficits in performance in a functional
observational battery were observed in male Long-Evans rats that were exposed to
2,5-hexanedione intraperitoneally at 0, 150, 225, and 350 mg/kg-day for 28 days (Shell et al.,
1992). These became apparent at some intermediate doses and time points, but no
neurohistopathological lesions were observed at any other exposure than the high dose after 28
days. Taken together, these studies suggested that 2,5-hexanedione-induced deficits in FOB
performance can precede the overt development of peripheral neuropathy, as exemplified by
axonal swelling, rearrangement of NFs, and regression of the myelin sheaths.
The molecular mechanisms involved in bringing about n-hexane-induced
neuropathological effects have been studied extensively. Several studies have suggested that the
mode of action involved the binding of the toxic metabolite, 2,5-hexanedione, to proteins
forming pyrrole adducts then undergo oxidation, leading to protein cross-linking. For example,
2,5-hexanedione was shown to cross-link NF proteins of spinal cord when administered to male
Sprague-Dawley rats for 180 days in drinking at a concentration of 5000 mg/L (Lapadula et al.,
1986). Spinal cords were isolated after exposure and their proteins separated using sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Separating the polypeptides
according to molecular weight by this approach revealed a reduced content of NF triplet proteins
and the additional presence of bands migrating at positions on the gel that were equivalent to
molecular weights of 138,000 and 260,000 Daltons. The latter were not observed in the
electrophoretic pattern of NFs obtained from unexposed animals. Lapadula et al. (1986) used
immunoblotting to demonstrate that lower molecular weight bands on control gels were
immunologically indistinguishable from higher molecular weight bands of gels carrying
neurofilamentous proteins of exposed animals. These findings indicate that cross-linking of
neurofilamentous proteins had taken place as a result of exposure to 2,5-hexanedione. As
reported by the authors, a number of higher molecular weight proteins were reactive with
antibodies to all three of the NF proteins under evaluation.
A substantial body of physiological and biochemical studies have explored the
mechanism by which n-hexane-derived 2,5-hexanedione binds to and cross-links proteins. For
example, DeCaprio et al. (1982) studied the covalent binding of 2,5-hexanedione to amino acids
and polypeptides in vitro. These authors showed nonacidic amino acids to be the most reactive
species when incubated with 2,5-hexanedione and 2,4-hexanedione. Polylysine also was
extremely reactive to both ketones. Using lysine with selectively t-butoxycarbonyl-blocked a-
or e-amino groups demonstrated that the e-amino group was six times more reactive than the a-
amino group. Moreover, while 2,4-hexanedione and a number of other diketones reacted with
84
-------
the e-amino group to some extent, incubation at pH 9.5 markedly enhanced the lysyl reactivity
of y-diketones such as 2,5-hexanedione. Mass spectrometric analysis of the reaction product
between the 2,5-hexanedione and the lysine e-amino group suggested that a substituted pyrrole
residue had been formed. Electrophoretic separation by charge of diketone-incubated bovine
serum albumin showed an increased mobility within the gel of 2,4- or 2,5-hexanedione-treated
protein with time. This study confirms that only diketones with y-spacing are capable of
forming pyrrole adducts, a necessary step in the neurotoxicity of alkanes.
The demonstration of pyrrole formation during 2,5-hexanedione-induced cross-linking
suggests that this may be part of the mechanism by which changes in the peripheral nerve
architecture are brought about. Sanz et al. (1995) carried out a series of in vitro assays to
quantify pyrrole adduct formation by several non-y- and/or y-diketones (such as
2,5-hexanedione). The solvents assayed were 2-hexanone; 3,4-dimethylhexane;
2,5-hexanedione; 3,4-dimethyl-2,5-hexanedione; 2-hexanol and 2,5-hexanediol as derivatives of
n-hexane; 5-methyl-3-heptanone; 6-methyl-2,4-heptanedione; 4-heptanone; and 4-heptanol as
derivatives of n-heptane. The results showed that 3,4-dimethyl-2,5-hexanedione and
2,5-hexanedione formed pyrroles at the greatest speed and to the greatest extent. This suggests
that these y-diketones may more readily form pyrroles than their non-y-diketone analogs.
Therefore, they would be expected to have the greater capacity for inducing neuropathological
effects.
In vivo evidence also supports the proposed mode of action for n-hexane-induced protein
cross-linking and pyrrole formation. Kessler et al. (1990) detected pyrrole-like substances in the
urine of a human volunteer exposed to n-hexane for 3 hours at a concentration of 146 ppm and in
the urine of male Wistar rats administered 0, 50, 100, 250, 500, 1000, or 3000 ppm n-hexane for
three 8-hour exposures. Mateus et al. (2002) detected pyrroles in the urine of male Wistar rats
exposed to 200 mg/kg or 300 mg/kg 2,5-hexanedione in the diet for up to 9 weeks.
Graham et al. (1982a) hypothesized that pyrrole derivatization of lysyl residues is central
to the development of NF aggregations. They used the 2,5-hexanedione analog, 3,4-dimethyl-
2,5-hexanedione, as a probe. The presence of the two methyl groups of the analog enhanced the
chemical's potential for pyrrole formation compared with that of 2,5-hexanedione. Rats given
0.25 mmol/kg 3,4-dimethyl-2,5-hexanedione every 8 hours developed severe limb paralysis
within 3 days. The condition was marked by axonal swelling just proximal to the first node of
Ranvier. The swellings contained masses of NFs.
Graham et al. (1982b) also demonstrated the in vitro interaction between
2,5-hexanedione and ethanolamine. Magnetic resonance spectroscopy characterized the product
85
-------
of this reaction as l-(2-hydroxyethyl)-2,5-dimethylpyrrole. The authors addressed the issue of
the nature of an orange-colored chromophore that had formed during the reaction and showed
that similar entities were formed as products of reactions between other primary amines or
proteins and 2,5-hexanedione or y-diketones such as 2,5-heptandione and 3,6-octanedione.
Graham et al. (1982b) speculated that cross-linking of NFs as a result of n-hexane exposure
likely involved some or all of the following processes: metabolism to 2,5-hexanedione,
interaction of that chemical with e-lysyl residues of proteins, cyclization to form pyrrole adducts
that undergo oxidation to electrophiles, and these electrophiles then react with protein
nucleophiles to result in covalent cross-linking of derivatized proteins to form higher molecular
weight protein aggregates.
Two further studies by Anthony et al. (1983a, b) demonstrated the enhanced
neuropathological activity of 3,4-dimethyl-2,5-hexanedione compared with that of
2,5-hexanedione. For example, in in vivo studies, five Sprague-Dawley rats/group (sex not
stated) were intraperitoneally injected five times/week for at least 7 weeks with either 0, 2.5, or
4 mmol/kg-day 2,5-hexanedione or 0, 0.0625, 0.125, or 0.25 mmol/kg-day 3,4-dimethyl-
2,5-hexanedione. Dimethyl substitution led to an acceleration of peripheral neuropathy as
judged by the lower dose and shorter time required for the onset of hind-limb paralysis. For
example, a daily dose of 0.25 mmol/kg-day 3,4-dimethyl-2,5-hexanedione produced hind-limb
paralysis after 19.6 ±1.4 days, indicative of a cumulative toxic dose of 3.5 ± 0.29 mmol/kg. By
contrast, a 16-fold higher dose of 2,5-hexanedione (4.0 mmol/kg-day) brought about hind-limb
paralysis after 35.8 days, equivalent to a cumulative toxic dose of 102 ± 7.4 mmol/kg.
Ultrastructural examination of a giant axonal swelling from the anterior root of rats exposed to
3,4-dimethyl-2,5-hexanedione showed an accumulation of NFs but comparatively few
microtubules. Light microscopy of the spinal cord showed large axonal swellings in the anterior
root, white matter, and anterior horn (Anthony et al., 1983a).
Anthony et al. (1983b) provided further evidence for increased reactivity and pyrrole-
forming capacity of 3,4-dimethyl-2,5-hexanedione compared with 2,5-hexanedione by studying
their rates of reaction with 0.02M ethanolamine and ovalbumin. Reaction products were
collected at various time points and then analyzed by electrophoresis. 3,4-Dimethyl-
2,5-hexanedione displayed a greater rate of pyrrole formation with ethanolamine than did
2,5-hexanedione, and there was a greater rate of covalent cross-linking of ovalbumin with the
dimethylated diketone. Protein cross-linking was also measured in the presence of both ketones.
The rates of polymer formed were 1-1.5 mol"1 hour"1 for 3,4-dimethyl-2,5-hexanedione
compared with 0.034-0.037 mol"1 hour"1 for 2,5-hexanedione. The increase in polymer
86
-------
formation represents an increase in the rate of protein cross-linking.
Genter et al. (1987) separated the d, I from the meso diastereomers of 3,4-dimethyl-
2,5-hexanedione, both of which form identical tetramethylpyrrol adducts in the reaction with
protein amino functions. The d,l diastereomer both formed pyrroles more rapidly and was more
neurotoxic than the meso diastereomer, strongly supporting the concept that pyrrole adduct
formation is a necessary step in the pathogenesis of y-diketone neurotoxicity. Rosenberg et al.
(1987) showed that the axonal swelling that followed intoxication with the d,l diastereomer were
demonstrably more proximal than those produced by the meso diastereomer, showing that the
rate of protein cross-linking determines the proximo-distal location of the axonal swelling.
Boekelheide (1987) carried out an in vitro study of the capacity of 2,5-hexanedione and
3,4-dimethyl-2,5-hexanedione to form cross-links in the lysine-rich polypeptide, tubulin (from
bovine brain and rat testis). The ability to form microtubules was altered in y-diketone-modified
preparations. Specifically, the maximal velocity of assembly was consistently different among
control and treated samples. A prominent decrease in the length of the nucleation phase was
observed in the presence of y-diketone. Gel filtration of the derivatized tubulin preparations
showed that dimerization had occurred in preparations exposed to 2,5-hexanedione or
3,4-dimethyl-2,5-hexanedione.
DeCaprio et al. (1988) compared the neurotoxicity and pyrrole-forming potential of
2,5-hexanedione and deuterated 2,5-hexanedione ([D10]-2,5-hexanedione) in vitro and in vivo.
The latter derivative was expected to form pyrroles at a slower rate than the native chemical
because of a primary isotope effect on the cleavage of the C-H bond. Incubation of bovine
serum albumin with 2,5-hexanedione and [D10]-2,5-hexanedione resulted in lower amounts of
pyrrole formation (Table 4-19).
Table 4-19. Pyrrole adduct formation in proteins from y-diketone-treated
rats
Treatment
Control
2,5-hexanedione
[D ! „] -2, 5 -hexanedione
Control
2,5-hexanedione
Dose level
(mg/kg-day)
-
3.5
3.5
-
2.5
Duration
(days)
17
17
17
38
38
Pyrrole concentration (nmol/mg protein)
Serum
0.2 ±0.0
7.5 ±1.8
2.8±0.8a
0.2 ±0.1
3.1 ±1.0
Brain stem
0.6 ±0.1
3.7 ±0.7
1.9±0.4a
0.6 ±0.4
1.7 ±0.6
Spinal cord
0.6 ±0.2
4.7 ±1.2
2.1±0.5b
0.3 ±0.2
2.2 ±0.7
87
-------
[D10]-2,5-hexanedione 2.5 38 1.2±0.1a 1.7 ±0.8 l.l±0.1b
'"p<0.05, significantly different from corresponding incubations with 2,5-hexanedione.
b/><0.01, significantly different from corresponding incubations with 2,5-hexanedione.
Source: DeCaprio etal., 1988.
DeCaprio et al. (1988) also exposed male Wistar rats via intraperitoneal injection of the
separate analogues. Milder symptoms of hind-limb paralysis for the perdeuterated chemical and
a lower rate of adduct formation were observed. However, tissue concentrations of each y-
diketone isomer were not markedly different. The degree of covalent cross-linking of axonal
proteins appeared to correlate with the amount of pyrrole formed (DeCaprio et al., 1988). These
findings were considered to support an absolute requirement for pyrrole formation in y-diketone
neurotoxicity.
Genter St. Clair et al. (1988) used the diketone 3-acetyl-2,5-hexanedione to show that
stable pyrrole derivatives of proteins could be formed without protein cross-linking. In vitro
incubations containing ovalbumin and 3-acetyl-2,5-hexanedione resulted in the formation of
pyrrole derivatives, with no sign of protein cross-linking. In vivo, daily intraperitoneal injection
of 3-acetyl-2,5-hexanedione in male Sprague-Dawley rats for 20 days showed no signs of hind-
limb paralysis, no axonal swelling of the peripheral nerves, and no aggregated NFs. The rate of
pyrrole formation in rats receiving 3-acetyl-2,5-hexanedione was similar to that in animals
injected with 2,5-hexanedione and 3,4-dimethyl-2,5-hexanedione. Isolated hemoglobin from rats
treated with the three y-diketones contained equivalent amounts of pyrroles. However, when the
plasma membrane protein, spectrin, was measured as an indicator of cross-linking, negative
results were obtained for 3-acetyl-2,5-hexanedione in contrast to the other two y-diketones. The
authors postulated that the electron-withdrawing acetyl group on the pyrrole formed by 3-acetyl-
2,5-hexanedione rendered the pyrrole ring less susceptible to oxidation. In the absence of
oxidation of the pyrrole ring, cross-linking of proteins could not occur. This hypothesis was
supported by the absence of peripheral neuropathy associated with the other y-diketones. This
evidence indicates that both pyrrole oxidation and protein cross-linking are necessary steps in the
pathogenesis of y-diketone neuropathy. The observation that the neurological deficits continued
to worsen for several weeks after cessation of exposure of humans to n-hexane could be
explained by the continuing oxidation of pyrrole rings and NF cross-linking.
DeCaprio and Fowke (1992) investigated the interaction of 2,5-hexanedione and spinal
cord NFs in vitro. Isolated spinal cord proteins were incubated with [14C]-2,5-hexanedione.
Incorporation of radioactivity and pyrrole formation in NFs increased linearly with
88
-------
2,5-hexanedione concentration. SDS-PAGE and fluorography showed prominent labeling of
three NF subunits (designated H, M, and L), in addition to some high molecular weight
components derived from NF-H and NF-M. Mild proteolysis permitted the isolation of the
carboxyl-terminal tail domains of NF-H and NF-M. These domains appeared to contain the
majority of the 2,5-hexanedione binding sites, suggesting the possibility for limited and selective
pyrrole adduction of NF proteins. Cyanogen bromide cleavage of 2,5-hexanedione-induced
pyrrole adducts of NF-protein "M" showed the greatest amount of 2,5-hexanedione binding in a
polypeptide fragment thought to correspond to a region at the carboxyl terminus where three
important lysine-containing sequences are situated (DeCaprio et al., 1997).
An observation consistently made between species, between adult and immature
members of the same species, and within individual humans and animals, was that longer axons
in the PNS and CNS were more vulnerable to the toxic effects of n-hexane and its metabolites
than shorter axons. The axonal swellings that initially occurred proximal to nodes of Ranvier in
the most distal internodes of the longest axons were filled with disorganized masses of NFs.
Thus, Graham et al. (1995) postulated that during repeated exposures to n-hexane the resulting
metabolism to 2,5-hexanedione resulted in progressive derivatization of protein lysyl amino
groups to form pyrrolyl adducts; oxidation of the pyrrole rings to electrophiles lead to increasing
levels of cross-linking of NFs during the proximo-distal transport of axoplasm. Furthermore, the
observations by Cavanaugh and Bennetts (1981) suggested that the constriction of axonal
diameter at nodes of Ranvier contributed to the formation of axonal swelling at these locations
by presenting obstructions to the transport of the growing masses of NFs. Additionally, they
observed that nonobstructing masses of NFs could be successfully transported to the synapse for
proteolysis. Since the rate of NF transport is 1 mm/day (Griffin et al., 1984), axonal length can
be seen to determine the period of time during which sufficient NF cross-linking must occur to
produce the threshold masses necessary to occlude transport and result in axonal swellings,
secondary myelin retraction and demyelination, and distal axonal degeneration.
89
-------
4.4.5. Genotoxicity Studies
Data from limited short-term in vitro tests provide minimal evidence of the genotoxicity
of n-hexane (Tables 4-20, 4-21, and 4-22). For example, NTP (1991) observed that n-hexane
was negative for gene reversion in Salmonella typhimurium strains T98, TA100, TA1535, or
TA1537 with or without activation with S9 (NTP, 1991; Mortelmans et al., 1986). Ishidate et al.
(1984) reported no increase in reverse mutations in the Ames test in strains TA92, TA94, TA98,
TA100, TA1535, or TA1537 treated with n-hexane (purity not stated). Houk et al. (1989)
described a spiral salmonella assay of n-hexane in strains TA98 and TA100 in which a slightly
elevated response (less than twofold over background) was reported in TA 98 without S9
activation. This finding was considered insignificant by the authors. Similarly, no DNA damage
was detected mEscherichia coll or Bacillus subtilis microsuspensions (McCarroll et al., 198 la,
b). n-Hexane provoked a marginal or weakly positive response in an in vitro test to induce
chromosome loss in Saccharomyces cerevisiae D61.M (Mayer and Goin, 1994). However, the
metabolite 2,5-hexanedione was clearly positive for chromosome loss in this system.
Table 4-20. Summary of in vitro assays on the mutagenicity/genotoxicity of
n-hexane
Test System
Cell/Strain
Results
Reference
Comments
Bacteria
S. typhimurium
E. coli
B. subtilis
TA98, TA100,
TA1535, TA1537
TA92, TA94, TA98,
TA100, TA1535,
TA1537
TA98, TA100
WP2, WP2urvA,
WP67, CM611,
WP100
W3110polA+,
P3478polA~
H17, M45
Negative (+/- S9)
Negative
(ND on S9 status)
Negative (+/- S9)
Negative (+/- S9)
Negative (+/- S9)
Mortelmans etal.,
1986
Ishidate etal.,
1984
Houk etal., 1989
McCarroll et al.,
1981a
McCarroll et al.,
1981b
Gene reversion
Spiral salmonella assay
DNA damage
Fungi
S. cerevisiae
D61.M
Borderline
positive (tested
-S9 only)
Mayer and Goin,
1994
Chromosomal loss
Mammalian cells
90
-------
Chinese hamster
fibroblasts
Human
lymphocytes
Chinese hamster
ovary
Chinese hamster
ovary
Chinese hamster
Mouse
lymphoma
CHL
V79
L5178Ytk+/-
Borderline
positive (tested
-S9 only)
Borderline
positive (-S9)
negative (+S9)
Negative (- S9)
Borderline
positive (+ S9)
Negative (+/- S9)
Negative (tested
-S9 only)
Negative (+/- S9)
Ishidate etal.,
1984
Perocco et al.,
1983
NTP, 1991
NTP, 1991
Lankas et al.,
1978
Hazleton Labs,
1992
Polyploidy
Inhibition of DNA
synthesis
SCE
CA
Induction/promotion
Forward mutations
ND = no data.
SCE = Sister chromatid exchanges.
CA = Chromosomal aberrations.
Table 4-21. Summary of in vivo assays on the mutagenicity/genotoxicity of
n-hexane
Species
Mouse
Rat
Strain
CD-I
B6C3F1
ND
Albino
Results
Negative
Negative
Negative
Negative
Positive
Positive
Comments
Dominant lethal
CAandMN
SCE
NCEandPCE
CA
CA
Reference
Litton Bionetics, 1980
Shelby and Witt, 1995
NTP, 1991
Hazleton Labs, 1992
CA = Chromosomal aberrations.
MN = Micronuclei.
NCE = Nonchromatic erythrocytes.
PCE = Polychromatic erythrocytes.
SCE = Sister chromatid exchanges.
ND = No data.
91
-------
Table 4-22. Summary of in vivo and in vitro assays on the mutagenicity/
genotoxicity of commercial hexane mixtures
Species
Strain
Results
Comments
Reference
In vitro assays
Bacteria
S. typhimurium
Mammalian Cells
Chinese Hamster
fibroblasts
TA98, TA100,
TA1535, TA1538
CHL
Negative (+/- S9)
Negative
(ND on S9 status)
Gene reversion
(in vapor phase)
CA
Microbiological
Associates, 1989
Kawachi et al.,
1980
In vivo assays
Rat
Sprague-Dawley
Negative
CA
Microbiological
Associates, 1990
CA = Chromosomal aberrations.
ND = No data.
Those in vitro cytogenic tests in mammalian cell lines that included n-hexane as a test
chemical have been generally negative, although n-hexane-induced polyploidy in Chinese
hamster lung fibroblast cells (CHL) (Ishidate et al., 1984). DNA synthesis was inhibited in
human lymphocytes in the presence of concentrations of n-hexane from 10"4-10"2 M but only at
cytotoxic concentrations (Perocco et al., 1983). NTP (1991) reported a marginally increased
incidence of sister chromatid exchanges (SCEs) in Chinese hamster ovary (CHO) cells in the
presence of S9 (not dose dependent). Similarly, n-hexane was negative for chromosomal
aberrations in CHO cells (Daughtrey et al., 1994b; NTP, 1991). n-Hexane was negative for
forward mutations in the mouse lymphoma L5178Y tk+/" assay (Hazleton Laboratories, 1992).
n-Hexane did not induce mutagenic activity in V79 Chinese hamster cells when a promoter,
methylazoxymethanol acetate, was added to the system (Lankas et al., 1978).
Tests for the genotoxic potential of n-hexane in vivo have been predominantly negative.
No dominant lethal mutations were induced following n-hexane exposure in CD-I mice (Mast
et al., 1988b; Litton Bionetics, 1980). Also, n-hexane did not induce CA and micronuclei in
bone marrow cells of B6C3F1 mice injected intraperitoneally with the chemical (Shelby and
Witt, 1995).
Hazleton Laboratories (1992) recorded a slight, but significant, increase in the number of
chromosomal mutations induced by n-hexane in albino rat bone marrow cells. Moreover, an in
vivo bone marrow cytogenetic assay found that male albino rats exposed to 150, 300, and 600
ppm of n-hexane for 5 days experienced a significant increase in CA (chromatid breaks and
markers) at all treatment levels compared with controls (Hazleton Laboratories, 1992).
92
-------
n-Hexane did not increase the incidence of SCEs in in vivo mouse bone marrow cells at
intraperitoneal doses of 500, 1000, or 2000 mg/kg (NTP, 1991). The dosed groups displayed
slight increases in CA, but this increase was not considered to be significant.
In the few studies that have addressed the genotoxicity/mutagenicity of a mixture
containing approximately 50% n-hexane, no gene reversion or chromosomal aberrations in CHO
cells (with or without activation) and chromosomal aberrations in CHL cells were seen in vitro
(Microbiological Associates, 1990; Microbiological Associates, 1989). In addition, in vivo, no
chromosomal aberrations were induced in male and female Sprague-Dawley rat bone marrow
cells after nose-only inhalation exposure to commercial hexane for 6 hours/day on 5 consecutive
days at concentrations of 876, 3249, and 8715 ppm (Microbiological Associates, 1990).
4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS AND
MODE OF ACTION—ORAL AND INHALATION
4.5.1. Oral Exposure
There are no studies that have examined the possible associations between oral exposure
to n-hexane and noncancer health effects in humans. A few studies in animals indicate that the
nervous system may be a target for the toxic effects of n-hexane following oral exposure. For
example, a 90-day gavage study in male COBS CD(SD) BR rats described the characteristic
appearance of axonal swellings in peripheral nerve in those subjects exposed to n-hexane at the
highest dose (3980 mg/kg-day) (Krasavage et al., 1980). The histopathological lesions were
accompanied by signs of hind-limb paralysis, a frequent symptom of neuropathy in experimental
animals exposed to n-hexane.
Subacute exposure to n-hexane also induced deficits in nerve conduction (Ono et al.,
1981). Specifically, exposures to approximately 811 mg/kg-day (after 2 weeks), 759 mg/kg-day
(2-4 weeks), 1047 mg/kg-day (4-6 weeks), and 2022 mg/kg-day (6-8 weeks) resulted in
statistically significant reductions in the proximal and distal MCVs of rats receiving n-hexane
compared with controls.
There are data suggesting that the principal metabolite of n-hexane, 2,5-hexanedione, is
responsible for the neurotoxicity associated with oral exposure to n-hexane. For example,
Krasavage et al. (1980) compared the neurotoxicity of n-hexane and that of its metabolites
(2,5-hexanedione, 5-hydroxy-2-hexanone, 2,5-hexanediol, 2-hexanone, and 2-hexanol) by
administering equimolar doses of each chemical by gavage to five male COBS, CD(SD)BR
rats/group for 5 days/week for 90 days. Based on the time taken by the rats to develop hind-limb
93
-------
paralysis, 2,5-hexanedione had approximately 38 times the neurotoxic potency of n-hexane itself
on an equimolar basis.
Lapadula et al. (1986) observed the formation of higher molecular weight polypeptides in
excised peripheral nerve fibers in rats exposed to 2,5-hexanedione in drinking water (5000 mg/L;
180 days) than those seen in the nerves of unexposed rats. Immunoblotting techniques provided
further evidence that cross-linking of neurofilamentous proteins may have occurred.
Neurophysiological and behavioral effects were observed in male Wistar rats exposed for
6 weeks to 5000 mg/L 2,5-hexanedione in drinking water (Ladefoged et al., 1989). Reduction of
MCV by 17% and rotarod balance time by 50% in rats orally exposed to 2,5-hexanedione
showed the capacity of the principal metabolite of n-hexane to induce subclinical symptoms of
peripheral neuropathy and motor/sensory deficits. Histopathological examination of peripheral
nerve fibers was also associated with the appearance of giant axonal swelling and a change in the
distribution of fiber area size in nerve fiber cross sections (Ladefoged et al., 1994).
The neurotoxicity of orally administered 2,5-hexanedione in male Wistar rats was linked
to the appearance of pyrrole-like substances in the urine of animals exposed to 200 or 400 mg/kg
2,5-hexanedione for 6 or 9 weeks (Mateus et al., 2002). Parallel experiments incorporated
supplemental amounts of zinc acetate (300 or 500 mg/kg) in the diets. Neurobehavioral testing
(rearing and ambulation in an open field) showed changes according to treatment, with those
animals exposed to 2,5-hexanedione alone at the higher dose performing significantly less well
than controls. The performance of animals exposed to zinc acetate plus 2,5-hexanedione was
intermediate between 2,5-hexanedione-exposed animals and controls. This suggested that zinc
may be protecting the animals from the neurotoxic effects of 2,5-hexanedione in some way,
possibly by interfering with the n-hexane-induced cross-linking of proteins.
In summary, information on the oral toxicity of n-hexane is limited to the studies of
Krasavage et al. (1980) and Ono et al. (1981). These studies provide evidence that the nervous
system is the target of toxicity following oral exposure to n-hexane. Studies indicate that oral
exposure to the primary metabolite of n-hexane, 2,5-hexanedione, results in many of the gross or
subclinical symptoms of peripheral neuropathy also observed with n-hexane.
4.5.2. Inhalation Exposure
Several studies establish associations between inhalation exposure to n-hexane and
human health effects. Specifically, occupational studies and case reports suggest that inhalation
exposure to n-hexane in humans may be associated with neurotoxicity (Section 4.1). For
example, Sanagi et al. (1980) monitored the neurophysiological performance of 14 workers
94
-------
exposed to n-hexane in the mixing and drying jobs at a factory producing tungsten carbide alloy.
The workers were examined for signs of neurological deficits compared with 14 workers who
were not exposed to any solvents in the same factory (Sanagi et al., 1980). Twenty-two
breathing zone air samples taken twice a year over a 2-year period indicated an 8-hour TWA of
58 ppm for n-hexane and 39 ppm for acetone. Compared with controls, exposed workers
reported a significantly increased occurrence of headache, hearing deficits, dysesthesia in limbs,
and muscle weakness. Exposed workers also showed an increased incidence of neurological
symptoms relating to muscle strength and reduced vibration sensation of the radial nerve.
Neurophysiological findings suggested that recovery from a slowing of motor nerve conduction
in the posterior tibial nerve was delayed after cessation of exposure.
Mutti et al. (1982a) compared MCVs in a group of 95 shoe factory workers exposed to a
mixture of hydrocarbons containing n-hexane and 52 unexposed workers from the same factory.
Exposed workers were divided into two groups based on hydrocarbon exposure. The TWA for
n-hexane of 108 breathing zone samples taken was 243 mg/m3 (69 ppm) in a mildly exposed
group and 474 mg/m3 (134 ppm) in a highly exposed group. When the severity of neurological
symptoms was compared, there was a gradation in response between the exposed groups, both of
which displayed more severe symptoms than the controls.
The groups of workers in the Sanagi et al. (1980) and Mutti et al. (1982a) studies showed
neurological symptoms as a result of n-hexane exposure. However, the subjects were also
exposed to other solvents concurrently: acetone in the tungsten alloy factory (Sanagi et al., 1980)
and cyclohexane, methyl ethyl ketone, and ethyl acetate in the shoe factory (Mutti et al., 1982a).
Some of these components may have also contributed to the neurotoxicological effects or may
have quantitatively affected the response to n-hexane to an uncertain extent. None of the
occupational exposure studies or case reports of n-hexane discussed in Section 4.1 involved
exposure to the single chemical. This suggests a limited utility of such data sets for dose-
response modeling of n-hexane.
Industrial hygiene surveys of occupationally exposed workers have shown good
correlations between the extent of occupational exposure to n-hexane and the concentration of
2,5-hexanedione in the urine (Prieto et al., 2003; Mayan et al., 2001; Cardona et al., 1996, 1993;
Mutti et al., 1993; Takeuchi, 1993; Saito et al., 1991; Ahonen and Schimberg, 1988). Therefore,
levels of this metabolite in the urine may be a useful, indirect means of monitoring exposure to
n-hexane in the workplace. For example, when Governa et al. (1987) investigated the
correlation between ENM changes indicative of polyneuropathy and urinary excretion of
metabolites indicative of exposure to n-hexane, they identified a value of 7.5 mg/L
2,5-hexanedione as representing a threshold to the occurrence of abnormalities. However, some
95
-------
variation from this relationship was apparent, because Governa et al. (1987) identified three
workers with 2,5-hexanedione urinary concentrations of 3.0, 3.3, and 4.5 mg/L, all of whom
displayed ENM changes. To protect against the onset of subclinical and clinical
neuropathological symptoms of n-hexane exposure, ACGIH proposed a BEI of 0.4 mg/L as an
acceptable concentration of 2,5-hexanedione in urine at the end of shift on the last day of a
workweek (ACGIH, 2003).
Other responses in humans occupationally exposed to solvents containing n-hexane
included the possible impairment of color vision (Gobba and Cavalleri, 2003; Iregren et al.,
2002; Issever et al., 2002; Seppalainen et al., 1979; Raitta et al., 1978) and the onset of some
parkinsonism-type neurological symptoms (Canesi et al., 2003; Vanacore et al., 2000; Hageman
et al., 1999; Pezzoli et al., 1996, 1995, 1989).
Several studies in laboratory animals support the human data on nervous system effects
following inhalation exposure to n-hexane. Huang et al. (1989) reported dose-dependent
reductions in MCV in male Wistar rats (eight/group) exposed to 0, 500, 1200, or 3000 ppm
n-hexane (>99% pure) for 12 hours/day, 7 days/week, for 16 weeks. Additionally, there were
some behavioral deficits in high- and mid-dose rats, including a reduction in grip strength and a
comparative slowness of motion indications of neurological impairment. No hind-limb paralysis
was observed. Histologically there was an increased incidence of paranodal swellings, some
evidence of demyelination, and remyelination was present in peripheral nerves. There were
statistically significant dose-dependent reductions in nervous system specific proteins,
particularly the P-S100 protein from tail nerve fibers. The neurophysiological deficits and
histopathology indicate a NOAEL of 500 ppm. Effects on hind-limb grip strength typically
preceded electrophysiological alterations in the progression of nerve degeneration following
exposure to neurotoxic chemicals (Harry et al., 1998). For example, studies using a chemical
with a similar mode of action as n-hexane, such as carbon disulfide, suggest that changes in
nerve conduction velocity are usually seen in later stages of nerve degeneration following
chemical exposure compared to behavioral alterations in hind-limb and fore-limb grip strength
(Harry et al., 1998; Sills et al., 1998).
Other studies have also observed neurological symptoms in experimental animals
exposed subchronically to n-hexane via inhalation. Ono et al. (1982) established a LOAEL of
200 ppm for histopathological effects characterized by axonal swelling and degeneration of the
myelinated axons in Wistar rats subchronically exposed to 200 and 500 ppm n-hexane.
Howd et al. (1983), Pryor et al. (1983), and Ichihara et al. (1998) used n-hexane
concentrations in the 1000-2000 ppm range to induce neurophysiological deficits and/or
behavioral changes in laboratory animals. Data from the Chemical Industry Institute of
96
-------
Toxicology's (CUT) 13-week toxicological study in F344 rats exposed to n-hexane confirmed
the neuropathological responses based on the dose-dependent appearance of paranodal swellings
of the tibial nerves in high- and mid-dose males (Cavender et al., 1984a, b).
The NTP-sponsored study of n-hexane in B6C3F1 mice induced neurobehavioral deficits
in high-dose mice (10,000 ppm) exposed for 6 hours/day, 5 days/week for 90 days and in another
group exposed to 1000 ppm n-hexane for 22 hours/day (NTP, 1991; Dunnick et al., 1989). The
authors concluded that the data in this study indicated an exposure concentration of 500 ppm
n-hexane as a NOAEL.
Huang et al. (1992) exposed male Wistar rats to 2000 ppm (99% pure) n-hexane for 12
hours/day, 6 days/week for a total of 24 weeks. Effects of treatment included an overall
reduction in MCV after 8 weeks and an increase in DL after 12 weeks. There was a reduction in
the activity or amount of neuron-specific enolase (y-enolase), creatine kinase-B, and the P-S100
protein. The onset of neurophysiological deficits was most evident in the distal segment of the
sciatic nerve (near the knee). Other sections of the central and peripheral nervous systems were
comparatively unaffected.
The reproductive/developmental toxicity of n-hexane has been investigated in a number
of studies in experimental animals exposed via the inhalation route. For example, Bus et al.
(1979) exposed pregnant F344 rats to 0 or 1000 ppm on GDs 8-12, 12-16, or 8-16. Progeny of
exposed dams had birth weights that were approximately 14% lower than controls for up to 3
weeks after birth. Litton Bionetics (1979) did not observe any n-hexane related effects in
reproductive, developmental, or teratological parameters when CRL:COBS CD(SD)BR rats were
exposed for 6 hours/day to concentrations of 0, 100, and 400 ppm n-hexane on GDs 6-15.
There was an n-hexane-related reduction in body weight gain in the fetuses of mid- and
high-dose dams (3.0-7.5% at the mid dose and 14-15% at the high dose in females and males),
when 30 pregnant Sprague-Dawley rats/group were exposed to 0, 200, 1000, or 5000 ppm
n-hexane (>99.5% pure) for 20 hours/day on GDs 6-19 (Mast, 1987). Examination of the
fetuses revealed some potentially n-hexane-related variations, including the incidence of
supernumerary ribs and reduced skeletal ossification. A companion study in CD-I mice (Mast et
al., 1988a) observed signs of maternal toxicity, such as reduced body weight gain and relative
uterus weight in high-dose dams. Fetal resorptions were evident in all exposure groups.
There was some evidence that n-hexane can induce toxicological effects in the male
reproductive system. Abnormal sperm and varying degrees of severity in the histopathology of
the testis were observed following inhalation exposure to n-hexane (Nylen et al., 1989;
DeMartino et al., 1987). However, exposing male B6C3F1 mice to n-hexane at concentrations
of up to 5000 ppm did not result in any changes of sperm morphology or in the architecture of
97
-------
the male reproductive system (Mast et al., 1988c). Similar treatments to male CD-I mice did not
induce dominant lethal effects when n-hexane-receiving animals were mated with unexposed
females (Mast et al., 1988b).
There is no clear evidence of other systemic effects resulting from inhalation exposure to
n-hexane. Data from CIIT's 13-week toxicological study in F344 rats exposed to n-hexane have
indicated that the kidney may be a target organ of n-hexane, at least in this test species
(Cavender et al., 1984a, b). However, this response may be related to the well-described a2u-
globulin-related hyaline droplet nephropathy that is characteristic of some strains of male rat but
not of humans. When mode of action evidence convincingly demonstrates that an effect is
secondary to a2u-globulin accumulation, the data are not to be used in the assessment of human
health risk (U.S. EPA, 1991). The criteria for demonstrating this mode of action for risk
assessment purposes have been clearly defined (U.S. EPA, 1991). Three core criteria must be
met: (1) increase in hyaline droplets in the renal proximal tubule cells; (2) determination that the
accumulating protein in the droplets is a2u-globulin; and (3) additional pathological lesions
associated with a2u-globulin are also present. In addition, a number of mechanistic studies can
be used to further support conclusions regarding the role of a2u-globulin. Data suggest that
kidney effects following inhalation exposure to n-hexane may be due to the accumulation of a2u-
globulin. However, the evidence for this mode of action is equivocal.
There is some evidence that exposure of experimental animals to high concentrations of
n-hexane via inhalation may result in portal-of-entry effects. For example, the NTP-sponsored
study of inhalation exposure of B6C3F1 mice to n-hexane reported signs of irritation in the
respiratory tract, such as inflammation, erosion, and regeneration of the olfactory epithelium
with fibrosis of the submucosa (NTP, 1991; Dunnick et al., 1989).
A substantial number of toxicological studies were carried out on various formulations of
a mixture containing approximately 50% n-hexane (commercial hexane) that provided limited
support for n-hexane-induced health effects. There were no signs of nervous system
degeneration when Sprague-Dawley rats were exposed subchronically to up to approximately
125 ppm commercial hexane (Biodynamics, 1978). However, 500 ppm commercial hexane was
effective in inducing some signs of neuropathological degeneration in male Sprague-Dawley rats
exposed in a similar exposure regimen (IRDC, 1992a, b).
Soiefer et al. (1991) reported in an abstract that Sprague-Dawley rats subchronically
exposed to commercial hexane via inhalation at concentrations of 9000 ppm showed no
neuropathological responses or altered performance in an FOB.
A single toxicological study has addressed the possible impacts of commercial hexane
when administered to experimental animals using a chronic dosing regimen (Daughtrey et al.,
98
-------
1999; Biodynamics, 1993a, b). The study exposed F344 rats and B6C3F1 mice to concentrations
of up to 9000 ppm commercial hexane for 2 years. Rats displayed a variety of histopathological
lesions in the respiratory epithelium at all doses in males and in mid- and high-dose females.
However, no treatment-related histopathological abnormalities in sciatic nerve were observed in
any group of F344 rats in this study.
Reproductive and developmental toxicological effects of commercial hexane have been
studied in experimental animals. Bushy Run Research Center (1989a, b) conducted a range-
finding study and a follow-up developmental toxicity study in rats and mice. BRRC (1989a, b)
exposed pregnant Sprague-Dawley rats (25/group) and pregnant CD-I mice (8-30/group) to 0,
900, 3000, or 9000 ppm commercial hexane for 6 hours/day on GDs 6-15. In the range-finding
study, developmental toxicity was observed in the progeny of mice exposed to 9000 ppm
commercial hexane only. Specifically, there was a reduction (per litter) in fetal weights in
progeny of the high-dose dams.
In a follow-up study, body weight gain was reduced in high- and mid-dose rat dams for
part of the exposure period lasting from GDs 9-12. The only developmental effects observed
were an increased incidence of two individual skeletal variations in high-dose pups. The
incidences of these lesions were 0/26 versus 6/26 (control versus high dose) for bilateral bone
islands at the first lumbar arch and 20/26 versus 26/26 (control versus high dose) for all
intermediate phalanges unossified.
Sprague-Dawley rats were exposed to commercial hexane at concentrations up to 9000
ppm in a two-generation reproductive/developmental toxicological study (Daughtrey et al.,
1994a; BRRC, 1991). There was a reduction of mean body weight in the Fl pups of the high-
dose dams at PND 14 and beyond. Reduced fetal weight was also observed in a single-
generation reproductive/developmental study of commercial hexane in Sprague-Dawley rats
(IRDC, 1986). The animals were exposed to 0, 100, 500, or 1503 ppm commercial hexane via
inhalation for 6 hours/day, 7 days/week for 100 days prior to mating, through the mating period
(maximum of 15 days), through GDs 1-20, and then postnatally through weaning (PND 21).
Fetal weights were reduced in the Fl pups, especially in high-dose progeny on lactation day 4,
where the reduction from control levels was 11-13%. The body weights of high-dose Fl pups
remained 8-9% lower than those of controls throughout lactation. Similar body weight
reductions compared with controls were also observed in the mid-dose group throughout
lactation, where the reductions were 12-17% from control values at their greatest extent and
achieved statistical significance on PNDs 14 and 21. The authors considered that the changes in
pup weight were of questionable biological significance and assigned aNOAEL of 1503 ppm to
the study.
99
-------
Some signs of systemic kidney toxicity of commercial hexane were suggested by
histopathologic lesions seen in high-dose male F344 rats exposed to 0, 900, 3000, or 9000 ppm
commercial hexane for 6 hours/day, 5 days/week for 13 weeks (Duffy et al., 1991; Biodynamics,
1989), although the relevance of this finding to human toxicity is uncertain, as discussed above.
A summary of toxicological studies of n-hexane in experimental animals exposed via the
inhalation route is provided in Table 4-23.
100
-------
Table 4-23. Toxicity findings in inhalation studies for n-hexane
Reference
Biodynamics
(1978)
Bus et al.
(1979)
Cavender et
al. (1984a, b)
Howd et al.
(1983)
Strain/species
12 Sprague-
Dawley
rats/sex/group
3 to 8 Pregnant
F344 rats
F344 Rats
Male F344 rats
Doses (ppm)
Phase I
0
6
26
129
(6 h/d, 5 d/w)
Phase II
0
5
27
126
(21 h/d, 7 d/w)
0
1000
(6 h/d)
0
3000
6500
10,000
(6 h/d, 5 d/w)
0
1000
(24 h/d, 6 d/w)
Duration
adjusted doses
(mg/m3)
0
3.8
16.4
81.2
0
15.4
83.3
388
0
881
0
1888
4091
6294
0
3021
Duration
26 Weeks
26 Weeks
CDs 8-12
CDs 12-16
CDs 8-16
13 Weeks
11 Weeks
Response a
No effects
No effects
Fetal weights 1 (with
partial recovery)
Organ/weight t ;
PNS histopathology
Hind-limb paralysis;
MCVI
NOAEL
(mg/m3)
81.2
388
None
1888
None
LOAEL
(mg/m3)
None
None
881
4091
3021
(continued on next page)
101
-------
Reference
Huang et al.
(1989)
Huang et al.
(1992)
Ichihara et al.
(1998)
IRDC (1992a,
b)
Litton
Bionetics
(1979)
Lungarella et
al. (1984)
Strain/species
8 Wistar
rats/sex/group
7 Male Wistar
rats
7 Male Wistar
rats
Male Sprague-
Dawley rats
20 Pregnant
CRL:COBS
CD(SD) BR
rats
12 Male New
Zealand rabbits
Doses (ppm)
0
500
1200
3000
(12 h/d, 7 d/w)
0
2000
(12 h/d, 6 d/w)
0
2000
(12 h/d, 6 d/w)
0
125
500
(22 h/d, 7 d/w)
0
100
400
(6 h/d)
0
3000
(8 h/d, 5 d/w)
Duration
adjusted doses
(mg/m3)
0
881
2115
5287
0
7058
0
3021
0
403
1615
0
88.1
352.5
0
2517
Duration
16 Weeks
24 weeks
20 weeks
6 months
CDs 6-15
24 weeks
Response a
Neurological FOB 1 ;
PNS histopathology;
MCVI
MCVI at 8 weeks
DL t at 12 weeks
1 Activity of
y-enolase, creatine
kinase-B, P-S100
MCVI,DLt
Abnormal gait and
peripheral nerve
histopathology (HD)
Liver pathology (LD)
No effects
Lung histopathology
(with partial
recovery)
NOAEL
(mg/m3)
881
None
None
403
None
352.5
None
LOAEL
(mg/m3)
2115
7058
3021
1615 (NS)
403 (liver)
None
2517
(continued on next page)
102
-------
Reference
Mast (1987)
Mastetal.,
(1988a)
NTP (1991);
Dunnick et al.
(1989)
Ono et al.
(1982)
Pryor et al.
(1983)
Strain/species
30 Pregnant
Sprague-
Dawley rats
35 Pregnant
CD-I mice
10B6C3F1
mice/sex/group
8 Male Wistar
rats/group
Male F344 rats
Doses (ppm)
0
200
1000
5000
(20 h/d)
0
200
1000
5000
(20 h/d)
0
500
1000
4000
10,000
(6 h/d, 5 d/w)
1000
(22 h/d, 5 d/w)
0
200
500
(12 h/d, 7 d/w)
0
2000
(14 h/d)
Duration
adjusted doses
(mg/m3)
0
587.5
2937
14,686
0
587.5
2937
14,686
315
629
2518
6294
2307
0
352
881
0
4112
Duration
CDs 6-19
CDs 6-17
13 Weeks
24 Weeks
14 Weeks
Response a
Fetal weights 1
Some ossification
changes
Relative uterus
weight, body weight
gain I (HD)
Fetal resorptions
Nasal irritation;
Relative liver, kidney
and heart weights t;
Neurobehavioral
deficits and
histopathology
MCVI,DL t
Axon degeneration
FOB effects
NOAEL
(mg/m3)
587.5
None
2937
None
315
None
None
LOAEL
(mg/m3)
2937
587.5
14686
587.5
6294
(and 2307)
352
4112
Increase (t); decrease (I); HD = high dose; LD = low dose.
103
-------
4.5.3. Mode of Action Information
Peripheral neuropathy following inhalation exposure to n-hexane demonstrates that the
nervous system is the target of toxicity for n-hexane. Further, inhalation exposure to n-hexane
has been shown to decrease MCV and SCV (with increased DL) in both humans and laboratory
animals. Myelin sheath thickness determines nerve conduction velocity and is proportional to
the diameter of the axon and internodal length (ffrench-Constant et al., 2004; Michailov et al.,
2004). Histopathological examination of the nerves of laboratory animals suggests that the mode
of action of n-hexane-induced neurotoxicity may involve a sequence of events including
accumulation of NF-filled axonal swellings with secondary demyelination that could lead to a
decrease in MCV and SCV (Spencer and Schaumburg, 1977a). Specifically, in the nerves of rats
exposed to 400-600 ppm n-hexane for up to 35 weeks, there was an increase in axonal swelling
in the distal region of large myelinated fibers (Spencer and Schaumburg, 1977a). As axonal
swelling progressed, subsequent localized demyelination and remyelination produced axonal
degeneration in a distal retrograde manner following further exposure to n-hexane. Shrinkage of
the internode, accumulation of NFs, and myelin sheath degeneration accompanied this axonal
swelling.
Studies in laboratory animals suggest that the n-hexane metabolite, 2,5-hexanedione, is
the primary toxic agent leading to neurological effects following exposure to n-hexane (Section
4.4.4). Administration of 2,5-hexanedione has been shown to result in axonal swelling
accompanied by axonal and secondary myelin degeneration in the PNS in laboratory animals
(Nachtman and Couri, 1984; Abou-Donia et al., 1982; Krasavage et al., 1980). The time to onset
of these symptoms, severity of the lesions, and magnitude of the neurotoxicity indicate that
2,5-hexanedione is more toxic than n-hexane itself or any its other metabolites. In vivo and in
vitro studies indicate that the mode of action of 2,5-hexanedione may involve covalent cross-
linking of NF proteins in peripheral nerve and spinal cord (Mateus et al., 2002; Sanz et al., 1995;
Kessler et al., 1990; Lapadula et al., 1986; DeCaprio et al., 1982). Specifically, evidence
suggests that 2,5-hexanedione may react with lysine residues to form pyrrole adducts. Genter St.
Clair et al. (1988) demonstrated that pyrrole derivatization is required for neuropathy to develop
but that it is not sufficient alone. In addition, oxidation of the pyrrole is also necessary for cross-
linking of NFs. Formation of pyrrole adducts, followed by oxidation of the pyrrole rings and
cross-linking of NFs, has been hypothesized to be responsible for the accumulation of NFs
observed in the distal axonal swellings of the peripheral nerves following n-hexane exposure
(Graham et al., 1982a, b, 1995).
Studies with chemicals structurally related to 2,5-hexanedione provide further support for
the neurotoxic mode of action of 2,5-hexanedione. Treatment with other y-diketones (such as 3-
104
-------
methyl-2,5-hexanedione, 3,4-dimethyl-2,5-hexanedione, and 1,2-diacetylbenzene) is associated
with axonal swelling and NF accumulation (Kim et al., 2002, 2001; Spencer et al., 2002; Graham
et al., 1995; Anthony et al., 1983a, b). These structurally related chemicals have also been
utilized to provide insight into the process of axonal swelling following exposure to y-diketones.
Neurofilaments are believed to accumulate in axonal swellings due to the pyrrolization and
cross-linking of their proteins as they are transported (anterograde) in the nerve. Studies indicate
that the extent and location of axonal swelling depend on the reactivity of n-hexane with NF
protein and the neurotoxicity associated with the n-hexane (Kim et al., 2002, 2001; Spencer et
al., 2002; Graham et al., 1995; Genter St. Clair et al., 1988; Anthony et al., 1983a, b). For
example, 3,4-dimethyl-2,5-hexanedione forms pyrroles faster than 2,5-hexanedione and therefore
the axonal swellings occur at mid-level and proximal nerve locations compared with distal
locations following exposure to 2,5-hexanedione. Further studies detailing the neurotoxicity of
the d,l and meso diastereomers of 3,4-dimethyl-2,5-hexanedione (Genter et al., 1987) and of 3-
acetyl-2,5-hexanedione (Genter St. Clair et al., 1988) demonstrate that both pyrrole formation
and subsequent oxidation of the pyrrole rings resulting in NF cross-linking are necessary steps in
the pathogenesis of n-hexane neurotoxicity.
4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER CHARACTERIZATION
4.6.1. Summary of Overall Weight-of-Evidence
Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), there is
inadequate information to assess the carcinogenic potential of n-hexane. Specifically, there are
no animal carcinogenicity studies available that examine exposure to n-hexane, and there is a
single human study (Beall et al., 2001) where workers were chronically exposed to mixtures
containing n-hexane along with other chemicals. A 2-year carcinogenicity bioassay in mice and
rats exposed to a mixture containing various hydrocarbons, including n-hexane, showed an
increased incidence of liver tumors in female mice (Daughtrey et al., 1999; Biodynamics, 1993a,
b). Daughtrey et al. (1999) observed an increased incidence of combined hepatocellular
adenomas and carcinomas in female mice exposed to the highest dose of a mixture containing
n-hexane (commercial hexane). In addition, the study authors identified a statistically significant
trend for increased incidence of pituitary adenomas in female mice exposed to commercial
hexane. Studies indicate that n-hexane is mostly nongenotoxic in short-term testing protocols.
n-Hexane showed a minimal response in S. cerevisiae D61.M (Mayer and Goin, 1994) and
induced an increased incidence in the number of chromosomal mutations in albino rat bone
marrow cells (Hazleton Laboratories, 1992). Also, the low pKa of exocyclic amino functional
105
-------
groups of DNA (<5) would preclude reaction with 2,5-hexanedione to yield pyrrole adducts.
Thus, these data suggest a lack of mutagenic potential of n-hexane. The available studies in
humans, as well as laboratory animals, are inadequate for cancer risk assessment. The previous
IRIS assessment (1990) did not contain a characterization of the carcinogenic potential of
n-hexane in humans.
4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence
Only one of the occupational exposure studies on n-hexane has inferred a possible
association between n-hexane and increased cancer incidence. Beall et al. (2001) conducted a
nested case control study evaluating the relationship between the occurrence of intracranial
tumors among employees at a petrochemical plant and exposure to chemicals, including ionizing
radiation, methylene chloride, acrylonitrile, vinyl chloride, formaldehyde, n-hexane, and various
other chlorinated, halogenated, volative, and aromatic hydrocarbons and nitroso compounds.
The workers were also exposed to organometallic and elemental metallic catalysts. The study
authors selected subjects from approximately 2595 plant workers. The workers were mailed
questionnaires that evaluated work history in the plant, and a total of 12 cases of intracranial
tumors, which developed after hire dates at the plant, were identified from the respondents. All
cases were confirmed by review of medical records and pathology specimens by four
neuropathologists. Six of these cases, all of which were men, had primary brain cancers or
gliomas (two astrocytomas, two glioblastomas, and two oligodendrogliomas). Six cases had
benign intracranial tumors, of which two were diagnosed as vestibular schwannomas (observed
in one man and one woman), two as meningiomas (both in men), and two pituitary adenomas
(observed in one man and one woman). Ten healthy controls were matched to each case by age,
gender, birth year, race and an initiation date for work in the building complex that was prior to
the tumor diagnosis date for the case. The median length of employment at the facility was 16.8
years for cases and 10.9 years for controls.
Work histories were obtained from company records or interviews, the latter providing
information about complete work history, exposures encountered, extent of hands-on work at
each job, and incidence of certain other nonoccupational factors that may be related to risk of
occurrence of brain cancers and intracranial tumors (exposure to diagnostic irradiation, use of
anticonvulsant and ototoxic drugs, history of head trauma, seizures, meningitis, use of cellular
phones and radiation badges, amateur radio operation, pesticide application, furniture refmishing,
and history of hearing loss). Exposure information was obtained from company accounting
records that detailed hours worked on projects during each year of employment and self-reported
workplace exposure to chemicals of interest. The authors compared cases and controls with
106
-------
respect to self-reported exposure to chemicals of interest, project-based work histories indicating
the potential use of chemicals of interest, and self-reported exposure to any of the other
nonoccupational factors that may be related to the risk of brain cancers. Conditional regression
was used and maximum likelihood estimates of odds ratios with a 95% confidence interval were
reported.
The authors showed that the OR for self-reported exposure to n-hexane was statistically
significantly elevated (OR, infinity), with a CI of 1.4 to infinity (6 cases and 26 controls
evaluated) for gliomas. The OR for potential exposure to n-hexane based on job-related
exposure estimates was 2.3 (CI, 0.4 to 13.7; 4 cases and 26 controls evaluated) for gliomas.
Analyses by duration indicated a statistically significantly elevated OR of 16.2 (CI, 1.1 to 227.6;
two cases and two controls evaluated) for potential long-term use of n-hexane (> 48 months) for
gliomas. No relationship was found between exposure to n-hexane and the occurrence of
intracranial tumors. While the results of this study indicated that exposure to n-hexane may have
contributed to the occurrence of brain tumors, specifically gliomas, the small number of cases,
large number of chemicals to which the employees were potentially exposed, and high
correlation between some of the parameters because of coexposure to several other chemicals,
the results do not permit a conclusion about carcinogenicity from exposure to n-hexane alone.
In laboratory animals exposed for 2 years via inhalation to a commercial hexane mixture
containing n-hexane (0, 900, 3000, or 9000 ppm), there was a statistically significant increase in
hepatocellular combined adenomas and carcinomas (7/50, 8/50, 9/49, or 16/50, respectively) in
female B6C3F1 mice (Daughtrey et al., 1999; Biodynamics, 1993b). This increase was not
observed in male mice or in either sex of F344 rats exposed to commercial hexane under the
same conditions.
Because commercial hexane is a variable mixture of hydrocarbons of which only about
52% is n-hexane, the use of commercial hexane as a toxicological surrogate for the qualitative
and quantitative effects of pure n-hexane may be unjustified.
n-Hexane has shown little evidence of mutagenic activity in a number of short-term test
systems. In vitro tests showed that n-hexane was not genotoxic in the salmonella (Ames) assay
(with or without activation), did not cause DNA damage of E. coli or B. subtilis, and was
negative for chromosomal aberrations in CHO cells and forward mutations in the mouse
lymphomaL5178 tk+/~ assay (Daughtrey et al., 1994b; Hazleton Laboratories, 1992; NTP, 1991;
Houk et al., 1989; Mortelmans et al., 1986; Ishidate et al., 1984; McCarroll et al., 1981a, b).
n-Hexane was marginal for inducing chromosome loss in the DNA of S. cerevisiae D61.M
(Mayer and Goin, 1994). In in vivo tests, n-hexane was negative for inducing dominant lethal
mutations in CD-I mice (Mast et al., 1988b; Litton Bionetics, 1980). Furthermore, n-hexane did
107
-------
not induce CA and micronuclei in bone marrow cells of B6C3F1 mice injected intraperitoneally
(Shelby and Witt, 1995). n-Hexane did not increase the incidence of SCEs in in vivo mouse
bone marrow cells (NTP, 1991). Hazleton Laboratories (1992) recorded a slight, but significant,
increase in the number of chromosomal mutations due to n-hexane exposure in albino rat bone
marrow cells.
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.7.1. Possible Childhood Susceptibility
No studies were available regarding adverse effects of n-hexane exposure in children. A
single study by Howd et al. (1983) provided data indicating adult rats may be more susceptible to
n-hexane-induced toxicity than weanling rats. Specifically, adult rats showed effects of greater
magnitude and earlier onset compared with weanling rats. The authors suggested that this
difference in neurotoxicity may be due to the decreased size of axons and greater rate of growth
and repair of peripheral nerves in weanling compared with adult rats.
Metabolism may vary between children and adults due to differences in the development
and maturity of phase I and phase II enzymes (Johnsrud et al., 2003; Ginsberg et al., 2002).
Studies indicate that the mode of action of n-hexane toxicity may involve metabolism to more
toxic metabolites. Several enzymes, such as CYP2E1, may be involved in n-hexane metabolism.
Studies with human liver microsomes collected from postmortem GD 8 to PND 18 samples
indicate that the amounts of CYP2E1 increased by more than two orders of magnitude with age.
Because CYP2E1 has been implicated in the transformation of n-hexane to its toxic metabolite,
2,5-hexanedione, these data suggest that the lower level of CYP2E1 in children may result in
altered responses (i.e., decreased susceptibility) to the toxic effects of n-hexane exposure.
4.7.2. Possible Gender Differences
The available data provide equivocal evidence suggesting gender differences in toxicity
following n-hexane exposure. Some apparent sex-specific neoplastic effects occurred in B6C3F1
mice in response to inhalation exposure to commercial hexane over a 2-year period (Daughtrey
et al., 1999; Biodynamics, 1993b). These include a dose-dependent appearance of hepatocellular
combined adenomas and carcinomas in female mice that was significantly different to controls at
the highest dose (9000 ppm) and some possibly n-hexane-related increase in the formation of
adenocarcinomas of the pituitary gland in female mice compared with controls. In these cases,
there are no obvious biochemical or physiological mechanisms underlying this apparent disparity
of response between the sexes. Accordingly, it could be argued that the difference in response
108
-------
may be accounted for by either sex differences, the influence of other hydrocarbons in the
mixture, or chance. The historical background rate of hepatocellular tumor formation in this
strain of mouse ranges from 11-70% with a mean of 42.1% in males and 3-54% with a mean of
25.2% in females (Daughtrey et al., 1999). Therefore, the apparent differences in incidence of
hepatocellular combined adenomas and carcinomas in exposed and control females may have
been due to an unusually low incidence of these lesions in the concurrent female controls.
109
-------
5. DOSE RESPONSE ASSESSMENT
5.1. ORAL REFERENCE DOSE
No epidemiology or case report studies examining health effects in humans or any
chronic laboratory studies evaluating potential health effects in animals following oral exposure
to n-hexane are available. The only oral n-hexane exposure study (Krasavage et al., 1980)
identified was of subchronic duration, utilized gavage exposure, and evaluated a small number
(five/group) of animals. Several animals died in each dose group (two in the mid-dose and one
in the high-dose group) during the course of the study. In this study, Krasavage et al. (1980)
exposed five male COBS CD(SD) BR rats/group to doses of 0, 6.6, 13.2, and 46.2 mmol/kg (570
mg/kg) n-hexane by gavage 5 days/week for 90 days. The period of treatment and observation
was extended to 120 days for those animals receiving 46.2 mmol/kg n-hexane to ensure that an
overt neuropathological endpoint was detected. The onset of neuropathy was assessed by the
initial appearance of hind-limb paralysis, at which point animals were sacrificed and examined
histopathologically. Hind-limb paralysis was observed in 3/4 animals exposed to the high dose
of n-hexane. Giant axonal swellings were present in the nerves of 4/4 animals in the high-dose
group.
An RfD for n-hexane cannot be derived in the absence of a suitable oral study of
sufficient duration that evaluates an array of endpoints. The previous IRIS assessment (1990)
for n-hexane also did not contain a derivation of an oral RfD. A route-to-route extrapolation
using available inhalation data is currently not possible since limited PBTK models are available
for n-hexane (Fisher et al., 1997; Perbellini et al., 1986). The Fisher et al. (1997) lactational
transfer model was developed using rodent tissue solubility and allometrically-scaled metabolic
rate constants from the published literature (in abstract form only) to estimate human tissue
metabolic parameters. The authors also suggested that the absence of exposure and toxicokinetic
data on lactational transfer of chemicals to nursing infants was a disadvantage of this model.
The PBTK model by Perbellini et al. (1986) is also inappropriate for use in route-to-route
extrapolation. The dose metric for the critical effect in this model is a function of the
concentration of 2,5-hexanedione in circulation. The concentration-duration-response function
for 2,5-hexanedione is unknown. In addition, the oral dose of n-hexane necessary to yield the
same blood-concentration-time profile for 2,5-hexanedione, taking into account gastrointestinal
uptake of n-hexane, is not accounted for by Perbellini et al. (1986). Furthermore, studies
indicate that the major metabolite of n-hexane in humans is 2,5-hexanedione but in laboratory
animals it is 2-hexanol. Thus, using a PBTK model based on information from laboratory
animal studies may not be appropriate.
110
-------
5.2. INHALATION REFERENCE CONCENTRATION
5.2.1. Choice of Principal Study and Critical Effect — with Rationale and Justification
Many occupational and experimental exposure studies in humans have investigated the
health effects following inhalation exposure to n-hexane. These studies indicate that the nervous
system is the target of toxicity of n-hexane (Section 4.1.2.2). Specifically, these human studies
show decreased MCV following exposure to n-hexane in the range of approximately 50-2500
ppm. However, all of the human occupational n-hexane exposure studies indicate coexposure to
a variety of other chemicals known to potentiate n-hexane-induced neurotoxicity (Section 4.4.3).
The most well-conducted n-hexane human occupational exposure study is by Sanagi et al.
(1980), which compared the neurological functions of workers in a tungsten carbide alloy factory
who were exposed to n-hexane (and other solvents) to workers in the same factory who were not
exposed to n-hexane. The authors determined that the breathing zone TWA concentration of
n-hexane in air was 58 ppm and coexposure to acetone was 39 ppm (two measurements per year
over a 2-year period). No other solvent concentrations were reported by the study authors.
Exposed workers had decreased MCV and showed clinical signs of peripheral neuropathy
identified by dysesthesia in limbs, muscle weakness, and reduced vibration sensation. This study
had been chosen as the principal study in the 1990 IRIS assessment. The principal limitation of
Sanagi et al. (1980) is that the workers were coexposed to n-hexane and acetone, which raises
the question of the role of acetone coexposure in the observed neurological changes.
Several studies indicate that acetone may affect n-hexane metabolism, neurotoxicity, and
reproductive toxicity (Zhao et al., 1998; Cardona et al., 1996; Ladefoged et al., 1994, 1989; Lam
et al., 1991; Larsen et al., 1991; Ladefoged and Perbellini, 1986). A study in humans showed
that acetone concentrations in the workplace significantly correlated with the ratio of urinary
n-hexane metabolites (specifically 2,5-hexanedione) to air n-hexane concentrations (Cardona et
al., 1996). It has been suggested that acetone may potentiate neurotoxicity by decreasing the
elimination of 2,5-hexanedione. For example, studies in rodents and rabbits have shown that
coexposure to acetone and 2,5-hexanedione decreases elimination and increases the
concentration of 2,5-hexanedione in serum and sciatic nerve compared with administration of
2,5-hexanedione alone (Zhao et al., 1998; Ladefoged and Perbellini, 1986). In addition, acetone
has been shown to induce CYP2E1, one of the enzymes shown to be involved in the metabolism
of n-hexane to its toxic metabolite 2,5-hexanedione in rats (Patten et al., 1986) (see also Section
3.3). Thus, coexposure to acetone may induce CYP450 enzymes and increase the production of
the neurotoxic metabolite 2,5-hexanedione.
Oral coexposure studies in rats further support acetone potentiation of n-hexane
111
-------
neurotoxicity (Section 4.4.3). Ladefoged et al. (1994, 1989) exposed male rats to
2,5-hexanedione alone and 2,5-hexanedione plus acetone in drinking water for 6 weeks and
evaluated neurological and behavioral endpoints. Rats exposed to 2,5-hexanedione alone and
2,5-hexanedione plus acetone showed decreased balance time on a rotating rod, altered behavior
(ambulation, grip strength, and rearing), decreased MCV, and increased giant axonal swelling of
the sciatic nerve. The authors stated that these effects were greater in severity in the rats
coexposed to 2,5-hexanedione plus acetone compared with those exposed to 2,5-hexanedione
alone. Lam et al. (1991) showed that coexposure of male Wistar rats to 0.5% n-hexane and 0.5%
acetone resulted in inhibition of acquisition (not performance) of spatial learning skills as
measured by the radial eight-arm maze. In addition, Larsen et al. (1991) suggested that
coexposure to acetone and 2,5-hexanedione may contribute to irreversible damage to the testis
and male infertility in rats. Taken together, the data suggest that acetone may alter n-hexane
metabolism and potentiate n-hexane-induced neurotoxicity and reproductive toxicity. Thus,
Sanagi et al. (1980) was not retained as the principal study for the derivation of the RfC, based
on coexposure of study subjects to n-hexane and acetone.
Several additional human occupational exposure studies (see Section 4.1.2.2) support the
n-hexane-induced neurotoxicity identified in humans by Sanagi et al. (1980), but all contain
insufficient data on the duration or concentration of n-hexane exposure. In addition, all available
human inhalation exposure studies indicate the potential for coexposure to other solvents, most
of which have been shown to potentiate n-hexane-induced toxicity (Section 4.4.3). Therefore,
reliable effect levels cannot be identified from the available reports of human experimental and
occupational exposure. For example, Mutti et al. (1982a) showed decreased MCV in shoe
factory workers exposed to mixtures containing n-hexane, cyclohexane, methyl ethyl ketone, and
ethyl acetate. The authors reported that these effects occurred at TWA n-hexane concentrations
in breathing zone air of 69 ppm (mildly exposed) and 134 ppm (highly exposed). Chang et al.
(1992) evaluated the neurological effects of exposure to offset machine workers in a printing
factory. These workers were exposed to lead and mercury in the printing inks and cleaning
solutions containing mainly (14-20%) n-hexane along with a variety of other solvents (including
toluene). TWA air concentrations of n-hexane were 63 ppm for the general air concentration and
134 ppm for breathing zone air for offset printing areas. The authors observed clinical
symptoms of paresthesia, weakness, and numbness in the extremities and electrophysical deficits
in MCV, SNAP, and MAP in exposed workers. Due to the known coexposures to other
chemicals and the lack of data on duration of exposure, the available human studies were not
considered further for the choice of the principal study.
As described in Section 4.2.2, the toxic effects in laboratory animals following inhalation
112
-------
exposure to n-hexane support the nervous system as the primary target of toxicity. A number of
studies identified a variety of effects on the nervous system, kidney, liver, and developing fetus
at doses between 125 and 500 ppm (IRDC, 1992a, b; NTP, 1991; Dunnick et al., 1989; Huang et
al., 1989; Mast et al., 1988a, b; Mast, 1987; Ono et al., 1982). These studies were considered for
the selection of the principal study and are described below. Benchmark dose (BMD) modeling,
where the data were amenable, was performed and is discussed in detail in Section 5.2.2 and
Appendix B.
Neurological deficits and respiratory lesions (mild epithelial lesions) were observed when
B6C3F1 mice were exposed subchronically to 0, 500, 1000, 4000, and 10,000 ppm n-hexane for
6 hours/day, 5 days/week for 90 days or to 1000 ppm n-hexane for 22 hours/day or 6 hours/day,
5 days/week for 90 days (NTP, 1991; Dunnick et al., 1989). Dunnick et al. (1989) reported
decreased locomotor activity and increased axonal swellings in the paranodal nerve in the 1000
ppm continuous exposure group (22 hours/day) and the 10,000 ppm exposure group (6
hours/day). Histopathology of the spinal cord and tibial nerve was evaluated in four animals/sex
from the control, 1000 ppm continuous exposure, and 10,000 ppm exposure groups only. The
NOAEL (500 ppm) was based on the appearance of mild epithelial lesions in the nasal cavity.
The authors suggested that this effect was more severe in the 1000 ppm continuous exposure
group (22 hours/day) than the 4000 ppm exposure group (6 hours/day). They also considered
these effects to be nonspecific and indicative of inflammatory and regenerative changes
secondary to the effects of the inhaled irritant. The authors were unclear as to whether the
altered morphology was due to inflammation or direct action of n-hexane. Thus, the study
authors stated that the nasal irritation was most likely secondary to the inhaled irritant. In
addition, the absence of sufficient neuropathological information from the mid-concentration
groups (i.e., 500, 1000, 4000 ppm for 6 hours/day) is considered to represent a significant
deficiency in the interpretation of the Dunnick et al. (1989) study. Therefore, the NTP
(1991)/Dunnick et al. (1989) study was not selected as the principal study for the derivation of
the RfC.
IRDC (1992a) exposed male Sprague-Dawley rats to 0, 125, and 500 ppm n-hexane
subchronically for 6 months (22 hours/day, 7 days/week). n-Hexane exposure resulted in a
significant decrease in mean absolute and relative liver and kidney weights at both doses. These
changes in organ weights were not accompanied by any histopathological evidence of liver or
kidney toxicity. In the second phase of this study, IRDC (1992b) demonstrated an increased
incidence of chronic nephritis in 6/11 controls and 10/10 rats exposed to 500 ppm n-hexane.
This response is considered equivocal due to the high incidence of kidney nephropathy in the
control animals. Axonal degeneration and muscle atrophy were also observed but only at the
113
-------
high dose. The data on axonal degeneration and muscle atrophy are not suitable for BMD
modeling since each effect lacks an adequate dose-response for modeling (i.e., effects were seen
at only the high dose). For example, 0/10, 0/10, and 7/10 animals showed tibial/sciatic nerve
axonal degeneration and 0/10, 0/10, and 9/10 animals showed skeletal muscle atrophy at 0, 125,
and 500 ppm, respectively. Finally, the results of this study are potentially compromised by
possible coexposure to a phthalate ester-type compound. The authors indicated that during
exposure a brown oily material collected on the glass beads of the inhalation system for each
exposure group. Samples of this brown material were subjected to infrared spectroscopy, which
confirmed the presence of a phthalate ester-type compound. While the observed axonal
degeneration at the high dose could constitute a LOAEL, the noted contamination compromises
the results. Therefore, IRDC (1992b) was not selected as the principal study for the derivation of
the RfC.
Ono et al. (1982) observed subchronic effects of n-hexane on the nervous system in male
Wistar rats (eight/group) exposed to 0, 200, and 500 ppm n-hexane for 12 hours/day for 24
weeks. Only one animal from each group was examined histopathologically in an attempt to link
any functional deficits to morphological changes that may have taken place over the duration of
the experiment. The authors stated that they did not observe any definite clinical signs of
neuropathy in any of the exposed groups. MCV and mixed MCVs (distal and both proximal and
distal combined) were statistically significantly decreased in rats exposed to n-hexane at both
200 and 500 ppm. Distal latency and proximal mixed MCV were statistically significantly
decreased at the low dose but not the high dose. Degeneration of the myelinated axons was
evident in the peripheral nerves at both exposures (histopathology in one animal). While the
observed decreases in MCV could constitute a LOAEL, the lack of observed clinical neuropathy
and failure to evaluate nerve histopathology on a larger number of animals are limitations of this
study. In addition, BMD modeling of the data produced poor goodness of fit values estimated
from the data (Section 5.2.2). Therefore, the Ono et al. (1982) study was not selected as the
principal study for the derivation of the RfC.
Mast et al. (1988a) exposed pregnant CD-I mice (30/group) to 0, 200, 1000, and 5000
ppm n-hexane for 20 hours/day on GDs 6-17. The authors reported a significant increase in the
number of late resorptions in mice exposed to 5000 ppm n-hexane. The effects noted are at only
the high dose. The Mast et al. (1988a) study was not selected as the principal study for the
derivation of the RfC since effects were noted only at a dose higher than doses where effects
were observed in other studies.
Mast (1987) exposed pregnant Sprague-Dawley rats (30/group) to 0, 200, 1000, or 5000
ppm n-hexane for 20 hours/day on GDs 6-19. The authors observed a statistically significant
114
-------
reduction in fetal body weight gain in males at 1000 and 5000 ppm n-hexane exposure. A
statistically significant increase in the incidence of reduced ossification of sternebrae 1-4 was
also observed at 5000 ppm. This study identified a developmental NOAEL of 200 ppm for these
effects, but the range between the NOAEL and the next higher dose (1000 ppm) was
considerable. This uncertainty in the dose-response makes the selection of this study as the
principal study questionable. Several additional studies have evaluated the effect of n-hexane
exposure on the reproductive system and the developing fetus (Linder et al., 1992; Mast et al.,
1988b, c; De Martino et al., 1987; Marks et al., 1980; Bus et al., 1979; Litton Bionetics, 1979).
In contrast to the studies by Mast (1987) and Mast et al. (1988a), these studies did not indicate
that n-hexane exposure produced adverse reproductive or developmental effects. Nevertheless,
BMD modeling was performed on the Mast (1987) data set. The results of the BMD modeling
can be found in Section 5.2.2 and Appendix B.
Huang et al. (1989) exposed Wistar rats (eight/group) via inhalation to 0, 500, 1200, or
3000 ppm (0, 1762, 4230, or 10,574 mg/m3) n-hexane for 12 hours/day, 7 days/week for 16
weeks. Statistically significant, group-specific, dose-dependent changes in MCV were obtained
in the mid- and high-concentration groups but not in the low-concentration group.
Histopathological changes to the peripheral nerves were marked by paranodal swelling and
demyelination. These changes were most apparent in high-dose rats but occurred in mid-dose
animals as well. Rats exposed to mid- and high-concentrations of n-hexane in the Huang et al.
(1989) study also showed some signs of behavioral deficits, including a reduction in grip
strength and slowness of motion. This study was considered further for selection as the principal
study for the derivation of the RfC. The data for changes in MCV were subjected to BMD
modeling (Section 5.2.2 and Appendix B).
The Huang et al. (1989) study was selected as the principal study with peripheral
neuropathy (decreased MCV) as the critical effect. The available human and animal n-hexane
inhalation exposure data suggest that the nervous system is the primary target of n-hexane
toxicity (Sections 4.1.2 and 4.2.1). Most of the reproductive and developmental studies suggest
that n-hexane does not adversely affect these endpoints. For this reason and due to the
uncertainty in the dose-response, the Mast (1987) study that evaluated developmental effects was
considered but not selected as the principal study for the derivation of the RfC. In addition,
Huang et al. (1989) evaluated a comprehensive array of neurological endpoints, an adequate
number of animals and exposure groups, and the study was of the appropriate quality for the
derivation of the RfC.
Several studies provide support for the selection of Huang et al. (1989) as the principal
115
-------
study and peripheral neuropathy as the critical effect. Specifically, studies in humans exposed to
n-hexane levels in the workplace to a range of approximately 30-200 ppm (130-690 mg/m3)
n-hexane, showed effects associated with peripheral neuropathy, such as decreased MCV
(Yucesoy et al., 1999; Karakaya et al., 1996; Chang et al., 1992; Huang et al., 1991; Yokoyama
et al., 1990; Huang and Chu, 1989; Mutti et al., 1982a, b; Sanagi et al., 1980). Studies in
animals also provide support for the selection of Huang et al. (1989) as the principal study. In a
follow-up study, Huang et al. (1992) observed an overall reduction in MCV in rats exposed to
2000 ppm n-hexane for 12 hours/day, 6 days/week for a total of 24 weeks, with the onset of
neurophysiological deficits most evident in the distal segment of the sciatic nerve. Other
sections of the central and peripheral nervous systems were comparatively unaffected. Howd et
al. (1983), Pryor et al. (1983), and Ichihara et al. (1998) all used single concentrations of
n-hexane in the 1000-2000 ppm range to induce neurophysiological deficits and/or behavioral
changes in F344 or Wistar rats exposed to n-hexane. Data from CIIT's 13-week toxicological
study in F344 rats exposed to n-hexane (0, 3000, 6500, or 10,000 ppm) confirmed the
neuropathological responses to n-hexane based on the appearance of paranodal swelling of the
tibial nerves in mid- and high-dose males (Cavender et al., 1984a, b).
5.2.2. Methods of Analysis
As described in Section 4.2.2, the toxic effects in laboratory animals following inhalation
exposure to n-hexane support the nervous system as the primary target of toxicity. A number of
studies identified a variety of effects on the nervous system, kidney, liver, and developing fetus
at relatively low doses (Huang et al., 1989; Mast, 1987; Ono et al., 1982). These studies were
considered for the selection of the principal study and are described below.
Endpoints selected that are continuous variables (fetal body weight and MCV) were
modeled with available continuous models (linear, polynomial, power, and Hill). The hybrid
model software in the Benchmark dose software (BMDS) is still undergoing beta-testing and was
not used because it was not considered to be sufficiently validated for use in quantitative dose-
response assessment. The hybrid modeling approach defines the benchmark response (BMR) in
terms of change in the mean. The BMR was defined as a change of 10% for quantal endpoints
such as developmental abnormalities in the fetus and a change of one standard deviation (1SD)
from the control mean for continuous endpoints such as fetal body weight and MCV (U.S. EPA,
2000c). This BMR was selected because there was no clear biological rationale for selecting an
alternative BMR level (U.S. EPA, 2000c).
The 95% lower confidence limit on the benchmark concentration (BMCL) estimates for
116
-------
the various studies are presented for the best fitting models in Table 5-1, and detailed discussion
of the modeling for each study is presented in Appendix B. For each model, the software
performed residual and overall chi-squared goodness-of-fit tests and determined the Akaike's
Information Criterion (AIC). The chi-squared p-va\ue is a measure of the closeness between the
observed data and the data predicted by the model fit. Models with chi-squared ^-values >0.1
were considered adequate fits. The AIC is a measure of the model fit based on the log-likelihood
at the maximum likelihood estimates for the parameters. Models with lower AIC values among
those with adequate chi-squared ^-values were identified. The BMCL estimates varied quite a
bit, depending on the endpoint, model selected, and parameter hypotheses and constraints. The
"best model" selection criteria are presented in detail in Appendix B. Huang et al. (1989) was
selected as the principal study (Section 5.2.1). Based on the criteria described in Benchmark
Dose Technical Guidance Document (U.S. EPA, 2000c), decreased MCV data (12 weeks) in
male rats provided the best fit using the Hill model. Output from the software for the Hill model
run (of the male rat MCV data) follows in Appendix B.
117
-------
Table 5-1. BMD results of n-hexane inhalation toxicity studies for selection of the principal study
Reference
Mast, 1987
Mast, 1987
Huang etal., 1989
Huang etal., 1989
Huang etal., 1989
Ono et al., 1982
Onoetal., 1982
Ono etal., 1982
Endpoint
Reduced
ossification of
sternebrae 1-4
Decreased fetal
body weight gain
MCV 8 weeks
MCV 12 weeks
MCV 16 weeks
MCV
Mixed MCV
(total)
Mixed MCV
(distal)
Dose
groups
4
4
4
4
4
3
3
3
Model
Nested
logistic
Quadratic
Hill
Hill
Hill
Power
Power
Power
Fixed
parameters
n>l
n >1
rho = 0
n=2
rho = 0
n=l
rho = 0
n=2
rho = 3.5
n = 0.75
rho = 2
n=0.5
rho = 2
n=0.5
Goodness of
fit/7-valuec
O.0001
(group:
0.1834)d
0.2517
0.789
0.313
0.779
0.3745
< 0.0001
< 0.0001
AIC
1433
-104.4
27.22
27.35
34.88
92.84
78.39
65.82
BMCab
(ppm)
1571
1540
198
156
367
120
60
51
BMCL"
(ppm)
943
848
143
122
321
85.4
33
28
BMCLHEC
(mg/m3)
2770
1494
252
215
566
150.5
58.1
49.3
a BMC = Benchmark concentration.
b Calculated using a BMR of 10% for quanta! endpoints, and a BMR of 1 standard deviation for continuous endpoints.
0 The nested module also estimates p-value for goodness-of-fit by applying a chi-squared test to data grouped by the strata or levels of these
covariates as well as by dose. As shown here, the chi-squared goodness-of-fit test may give different results than the usual methods used
elsewhere in this table.
d All models were adjusted for total litter size.
118
-------
5.2.2.1. Adjustment to a Human Equivalent Exposure Concentration
Because the RfC is a standard applicable to continuous lifetime human exposure but
derived from animal studies featuring intermittent, less-than-lifetime exposures, EPA guidance
(U.S. EPA, 1994b) provides mechanisms for (1) adjusting experimental exposure concentrations
to a value reflecting continuous exposure duration (ADJ) and (2) determining a human
equivalent concentration (HEC) from the animal exposure data. The former employs an inverse
concentration-time relationship to derive a health-protective duration adjustment to time-weight
the intermittent exposures used in the principal study. The BMCL (1SD) of 122 ppm
(430 mg/m3) for decreased MCV in rats exposed to n-hexane for 12 weeks exposure
(12 hours/day, 7 days/week) as reported by Huang et al. (1989) is adjusted to continuous
exposure (12 hours/day to 24 hours/day) as follows:
BMCL ADJ = BMCL (mg/m3) x 12 hours/day - 24 hours/day
= 430 mg/m3 x 12 hours/day + 24 hours/day
= 215 mg/m3
The RfC methodology provides a mechanism for deriving a human equivalent
concentration from the duration-adjusted point of departure (BMCLADJ) determined from the
animal data. The approach takes into account the extrarespiratory nature of the toxicological
responses and accommodates species differences by considering blood:air partition coefficients
for n-hexane in the laboratory animal (rat) and humans. According to the RfC guidelines (U.S.
EPA, 1994b), n-hexane is a category 3 vapor because it is largely inactive in the respiratory tract,
is rapidly transferred between the lungs and blood, and has toxicological effects that are
extrarespiratory. Therefore, the duration adjusted BMCL^j is multiplied with the ratio of the
animal/human blood:air partition coefficients (LA/LH). As set forth in Section 3.1, values
reported in the literature for these parameters include an LA of 2.29 in F344 rats (Gargas et al.,
1989) and an LH in humans of 0.8 (Perbellini et al., 1985). By default, because an LA/LH greater
than unity would be derived from these values, a value of unity is adopted for this ratio. This
allows a BMCL to be derived as follows:
= BMCLADJ (mg/m3) x 1 (interspecies conversion)
= 215 mg/m3 x 1
= 215 mg/m3
The BMCLjjEc value of 215 mg/m3 for reduced MCV in rats was used to derive the RfC
119
-------
for n-hexane.
5.2.3. RfC Derivation—Including Application of Uncertainty Factors (UFs)
The BMCLjjEc of 215 mg/m3 for reduced MCV in Wistar rats exposed to n-hexane for 12
weeks duration (12 hours/day, 7 days/week, 16 weeks total duration) was used as a point of
departure for calculating the RfC for n-hexane (Huang et al., 1989).
A total UF of 300 was applied to the point of departure of 215 mg/m3: 10 for intraspecies
variation (UFH: human variability); 3 for interspecies differences (UFA); 3 to extrapolate from
less than lifetime to chronic exposure (UFS); and 3 to account for database deficiencies (UFD).
A UFH of 10 was applied to account for variations in susceptible subpopulations. One
animal study suggests that weanling rats may be less susceptible to n-hexane-induced
neurotoxicity than adult rats (Howd et al., 1983). Howd et al. (1983) compared the neurotoxicity
of n-hexane in weanling versus young adult F344 rats, which were exposed to 0 or 1000 ppm
n-hexane (95% pure) 24 hours/day, 6 days/week for 11 weeks. The authors observed
significantly decreased grip strength and increased incidence of hind-limb paralysis in both
weanling and adult rats. However, both endpoints appeared earlier and were of greater severity
in adults compared with weanlings. The authors suggested that these differences in n-hexane-
induced neurotoxicity may be due to smaller diameter and shorter axons in weanling compared
with adult rats.
The CYP2E1 enzyme is responsible for metabolism of various aliphatic and aromatic
hydrocarbons, solvents, and industrial monomers, including n-hexane and acetone.
Polymorphism of CYP2E1 could possibly lead to interindividual differences in the toxicity of
chemicals metabolized by this enzyme. n-Hexane-induced neurotoxic effects are believed to be
the result of its metabolism to the toxic metabolite, 2,5-hexanedione, by CYP2E1. In addition,
differences in the development and maturity of phase I and phase II metabolic enzymes
(specifically CYP2E1) between adults and children have been shown in several studies
(Johnsrud et al., 2003; Ginsberg et al., 2002). Taken together, these data suggest that differences
in metabolism of n-hexane may exist within the human population and between adults and
children.
Only one study with one dose group is available that directly observed susceptibility
differences between adult and weanling animals (Howd et al., 1983). Several mode of action
studies provide some evidence supporting the hypothesis that this increased susceptibility is due
to differences in axonal length between adults and weanling rats. These studies did not directly
observe effects of n-hexane on NFs in weanling or young animals. Given the paucity of studies
120
-------
directly observing susceptibility differences between weanling and adult animals and the
possibility of altered metabolic enzyme activity among individual humans and between adults
and children, a UFH of 10 was applied to account for variations in susceptible subpopulations.
A UFA of 3 was applied to account for uncertainty in extrapolating from laboratory
animals to humans. This value is adopted by convention where an adjustment from an animal-
specific EMCL^j to a BMCLjjEc has been incorporated. Application of a full uncertainty factor
of 10 would depend on two areas of uncertainty (i.e., toxicokinetic and toxicodynamic
uncertainties). In this assessment, the toxicokinetic component is mostly addressed by the
determination of a human equivalent concentration as described in the RfC methodology (U.S.
EPA, 1994b). The toxicodynamic uncertainty is also accounted for to a certain degree by the use
of the applied dosimetry method.
A UFS of 3 was applied to extrapolate from subchronic to chronic exposure. A
subchronic (16 weeks) study was used for the derivation of the RfC. However, 16 weeks is half
of the time required for a newly synthesized NF protein to be transported from the neuronal cell
body to the axon terminal in the longest axons of the CNS and PNS of an adult rat (Griffin et al.,
1984). The rate of NF transport down an adult rat axon is 1 mm/day. The longest axons extend
from the lumbar spinal cord to the hind foot and measure no more than 22 cm in the adult rat.
Thus, transport for the full length of the axon would take approximately 32 weeks in an adult rat.
Since the lifetime of NFs (target of toxicity of n-hexane) is shorter than the lifetime of an adult
rat, extrapolation from subchronic to chronic exposure is not necessary and a UFS of 3 was
applied.
A UFD of 3 was applied to account for database deficiencies. The database includes
many human occupational exposure studies (all with coexposure to other potentially neurotoxic
chemicals), subchronic animal studies in rats and mice, neurotoxicity studies in both humans and
laboratory animals, and developmental studies in rats and mice following inhalation exposure to
pure n-hexane. The database lacks a developmental neurotoxicity study and a multigeneration
reproductive and developmental toxicity study following inhalation exposure to pure n-hexane
alone. Prenatal exposure to pure n-hexane-induced skeletal anomalies, decreased fetal body
weight, and increased resorptions, suggesting that the fetus may be affected by n-hexane
inhalation exposure (Mast et al., 1988a; Mast, 1987; Bus et al., 1979). One of these studies
indicated a developmental NOAEL of 200 ppm for reduced fetal body weight gain (Mast, 1987).
However, it remains unclear whether these developmental effects occur at doses lower than those
that cause neurotoxicity. Studies investigating the reproductive and developmental effects of
commercial hexane, a mixture containing approximately 50% n-hexane, are also available (see
Section 4.4.2.2.3 ). These studies evaluated reproductive and developmental effects following
121
-------
exposure to doses of > 500 ppm commercial hexane and resulted in marginal decreases in pup
body weights and increased skeletal variations (BRRC, 1989a, b). Given the lack of
multigeneration reproductive and developmental studies following exposure to pure n-hexane
and the uncertainty associated with low-dose developmental effects of exposure to n-hexane, a
UFD of 3 was applied.
A UF to account for the extrapolation from a LOAEL to a NOAEL was not applied
because BMD modeling was used to determine the point of departure for derivation of the RfC.
Therefore, the RfC from the Huang et al. (1989) data is calculated as follows:
RfC = BMCLjjEc -H UF
= 215 mg/m3 -300
= 0.72 mg/m3
= 7 x 10'1 mg/m3
5.2.4. Previous RfC
The previous IRIS assessment for n-hexane contained an RfC of 2 x 10"1 mg/m3 that was
derived from the Sanagi et al. (1980) occupational exposure study in which group-specific
behavioral deficits and neurophysiological changes were considered to be associated with a
LOAEL of 58 ppm. However, the subjects of the study were also exposed to acetone at a mean
concentration of 39 ppm. More recent data suggest that coexposure to acetone potentiates
n-hexane metabolism and n-hexane-induced neurotoxicity (Cardona et al., 1996; Ladefoged et
al., 1994, 1989; Larsen et al., 1991). Therefore, it is possible that the incidence or severity of the
neurological changes observed by Sanagi et al. (1980) may have been a result of coexposure to
both solvents. Thus, Sanagi et al. (1980) was not retained as the principal study for the
derivation of the RfC, based on coexposure of study subjects to n-hexane and acetone in the
current assessment.
The subchronic NTP (1991) study (published in the literature as Dunnick et al., 1989) in
which B6C3F1 mice were exposed to 0, 500, 1000, 4000, and 10,000 ppm 6 hours/day, 5
days/week or 1000 ppm 22 hours/day, 5 days/ week n-hexane via inhalation for 13 weeks was
used as a co-principal study. The critical effect in the subchronic study was epithelial lesions in
the nasal cavity.
122
-------
The chosen NOAEL (500 ppm) in the coprincipal study was based on the appearance of
mild epithelial lesions in the nasal cavity. The authors suggested that this effect was more severe
in the 1000 ppm continuous exposure (22 hours/day) than in the 4000 ppm exposure group (6
hours/day). They also considered these effects to be nonspecific and indicative of inflammatory
and regenerative changes secondary to the effects of the inhaled irritant. The authors were
unclear as to whether the altered morphology was due to inflammation or direct action of
n-hexane.
Histopathology of the spinal cord and tibial nerve was reported for four males and four
females from the 0, 1000 ppm continuous exposure, and the 10,000 ppm exposure groups only.
Dunnick et al. (1989) reported decreased locomotor activity and increased axonal swelling in the
paranodal nerve in the 1000 ppm continuous exposure (22 hours/day) and the 10,000 ppm
exposure (6 hours/day). The study authors did not perform neurological histopathology at the
mid concentrations (500, 1000, or 4000 ppm for 6 hours/day). The lack of histopathology is
considered to be a significant deficiency in the Dunnick et al. (1989) study, since the nervous
system appears to be the primary target of n-hexane-induced neurotoxicity (Section 4.5.2).
Thus, Dunnick et al. (1989) was not retained as the principal study for the derivation of the RfC
in the current assessment.
The available human and laboratory animal inhalation studies for n-hexane indicate that
the nervous system is a target of toxicity (Sections 4.1.2.2 and 4.2.2.1). Thus, the absence of
sufficient neuropathological information from among the mid-concentration groups (i.e., 500,
1000, 4000 ppm for 6 hours/day) is considered to represent a significant data gap in the Dunnick
et al. (1989) study, bounding the selected NOAEL with greater than acceptable uncertainty.
Therefore, Dunnick et al. (1989) and Sanagi et al. (1980) were replaced as the principal studies
by that of Huang et al. (1989).
5.3. CANCER ASSESSMENT
As discussed in Section 4.6.1., the available database for n-hexane contains inadequate
information to assess carcinogenic potential according to the U.S. EPA's Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005a). Genotoxicity evidence has been largely
negative for n-hexane.
A 2-year inhalation study of commercial hexane, a mixture containing n-hexane and
other structurally related hydrocarbons, showed a statistically significant increase in combined
hepatocellular adenomas and carcinomas in female B6C3F1 mice (Daughtrey et al., 1999;
Biodynamics, 1993a, b). However, the increased tumor incidence was of borderline statistical
123
-------
significance and was not present in treated male mice nor in either sex of F344 rats exposed to
commercial hexane under the same conditions. Additionally, because commercial hexane is a
variable mixture of hydrocarbons of which only about 52% is n-hexane, its use as a toxicological
surrogate for the qualitative and quantitative effects of pure n-hexane would be unjustified.
124
-------
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
n-Hexane (CASRN 110-54-3) is a solvent that has many uses in the chemical and food
industries, either in pure form or as a component of commercial hexane. The latter is a mixture
that contains approximately 52% n-hexane; the balance is made up of structural analogs and
related chemicals such as methylpentane and methylcyclopentane.
Highly purified n-hexane is used as a reagent for chemical or chromatographic
separations. Other grades of n-hexane are used as solvents for extracting edible fats and oils in
the food industry and as a cleaning agent in the textile, furniture, and printing manufacturing
industries. Hexane is the solvent base for many commercial products, such as glues, cements,
paint thinners, and degreasers. n-Hexane is a minor constituent of crude oil and natural gas and
occurs in different petroleum distillates.
No data are available regarding the potential toxicity of n-hexane in humans orally
exposed to n-hexane. However, as might be expected for a chemical with such wide application,
the potential exists for persons to be environmentally and/or occupationally exposed to n-hexane
via other routes of exposure. In fact, a considerable number of epidemiological studies
(summarized in Section 4.1) have been reported on n-hexane, the majority of which have shown
an association between inhalation exposure to n-hexane and neurological symptoms in
occupationally exposed individuals. However, the extent of exposure to n-hexane in many, if
not all, of the occupational studies is imprecise, and subjects were likely exposed concurrently to
other solvents. This means that the data in these studies are inappropriate for dose-response
modeling.
Animal data also indicate that the nervous system may be the primary target of toxicity
following inhalation exposure. The principal study of Huang et al. (1989) identified behavioral,
neurophysiological, and neuropathological effects in Wistar rats. This study has been used to
derive an RfC of 7 x 10'1 mg/m3.
Compelling mode of action evidence has focused on the capacity of n-hexane to undergo
metabolism to 2,5-hexanedione. This y-diketone appears to have the ability to interact with
specific proteins on the NFs. While some of the details remain to be worked out, a
preponderance of evidence suggests that pyrrole formation is critical for the induction of
neurotoxicity by y-diketones, with pyrrole oxidation a necessary further step to initiate NF cross-
linking. The importance of y-diketone-generated pyrrole formation also indicates the unique
125
-------
nature of n-hexane-generated neuropathy compared with the relative benign effects of such
structural analogues as n-heptane and pentane. The inability of the latter chemicals to form y-
diketones does not permit their generation of the neurotoxicological impacts manifest by
n-hexane.
Data on the toxicity of n-hexane via the oral route are poor as only one study was
identified in which experimental animals were exposed to n-hexane for an adequate duration
(Krasavage et al., 1980). However, as discussed in Section 5.1, this study is inadequate for the
development of an oral toxicity value for n-hexane.
Under the Guidelines for Carcinogenic Risk Assessment (U.S. EPA, 2005a), data are
inadequate for an assessment of the human carcinogenic potential of n-hexane. This descriptor
is chosen because no human or animal studies are available. Carcinogenicity data are available
for the tumor-inducing capacity of commercial hexane (of which about 52% is n-hexane).
However, the relevance of this study to the identification of the carcinogenic potential of
n-hexane is unclear due to the unknown toxicity contribution of the other components of the
mixture and uncertainty as to whether the apparent carcinogenic response in female mice was
truly treatment related.
6.2. DOSE RESPONSE
6.2.1. Noncancer
The database for oral exposure to n-hexane is limited to two subchronic gavage studies,
both of which were unsuitable for the calculation of an RfD.
The RfC of 7 x 10"1 mg/m3 was derived from reduced MCV in male Wistar rats following
inhalation exposure to n-hexane (Huang et al., 1989). Treatment-related changes included
reduced peripheral nerve conduction, some behavioral perturbation, and histological changes
indicative of peripheral neuropathy. There is sufficient evidence from other studies in
experimental animals to confirm that the nervous system is the primary target for the
toxicological effects of n-hexane (Huang et al., 1992; NTP, 1991; Dunnick et al., 1989;
Cavender et al., 1984a, b; Howd et al., 1983; Pryor et al., 1983). Subclinical and overt
symptoms of peripheral neuropathy have been described in persons exposed to n-hexane in the
workplace.
The overall confidence in this RfC assessment is medium. Confidence in the principal
study (Huang et al., 1989) is medium; it involves a comparatively low but acceptable number of
animals per group (eight/sex) and reports behavioral deficits, neurophysiological changes, and
126
-------
neuropathological effects within a dose-range in which both a NOAEL and LOAEL could be
identified. Animal studies in a second species (mice) corroborate the primacy of the
neurological endpoint and confirm the validity of the critical effect for peripheral neuropathy.
Confidence in the database is medium. The database lacks chronic exposure information on pure
n-hexane via any route of exposure and a multigenerational developmental and reproductive
toxicity study and a developmental neurotoxicity study. The sub chronic inhalation study of
Huang et al. (1989) satisfies the minimum inhalation database requirements for deriving an RfC
for n-hexane. Reflecting medium confidence in the principal study and medium confidence in
the database, confidence in the RfC is medium.
6.2.2. Cancer
Under the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), the database
for n-hexane is inadequate to assess human carcinogenic potential. As such, data are unavailable
to calculate quantitative cancer risk estimates.
127
-------
7. REFERENCES
Abbritti, G; Siracusa, A; Cianchetti, C; et al. (1976) Shoe-makers' polyneuropathy in Italy: the aetiological problem.
BrJIndMed 33:92-99.
Abou-Donia, MB; Makkawy, HA; Graham, D. (1982) The relative neurotoxicities of n-hexane, methyl n-butyl
ketone, 2,5-hexanediol, and 2,5-hexanedione following oral or intraperitoneal administration in hens. Toxicol Appl
Pharmacol 62:369-389.
ACGIH (American Conference of Governmental Industrial Hygienists). (2001) Documentation of the Threshold
Limit Values and Biological Exposure Indices. Cincinnati, OH: American Conference of Governmental Industrial
Hygienists.
ACGIH. (2003) TLVs and BEIs based on the documentation of the Threshold Limit Values for chemical substances
and physical agents and biological exposure indices. Cincinnati, OH: American Conference of Governmental
Industrial Hygienists.
Ahonen, I; Schimberg, RW. (1988) 2,5-hexanedione excretion after occupational exposure to n-hexane. Br J Ind
Med 45:133-136.
Allen, N; Mendell, J; Billmaier, D; et al. (1975) Toxic polyneuropathy due to methyl n-butyl ketone. An industrial
outbreak. ArchNeurol 32:209-218.
Altenkirch, H; Stoltenburg, G; Wagner, H. (1978) Experimental studies on hydrocarbon neuropathies induced by
methyl-ethyl-ketone (MEK). J Neurol 219:159-170.
Altenkirch, H; Wagner, H; Stoltenburg, G; et al. (1982) Nervous system responses of rats to subchronic inhalation of
N-hexane and N-hexane + methyl-ethyl-ketone mixtures. J Neurol Sci 57:209-219.
Anderson, RJ; Dunham, CB. (1984) Electrophysiologic deficits in peripheral nerve as a discriminator of early
hexacarbon neurotoxicity. J Toxicol Environ Health 13:835-843.
Anthony, DC; Boekelheide, K; Graham, D. (1983a) The effect of 3,4-dimethyl substitution on the neurotoxicity of
2,5-hexanedione. I. Accelerated clinical neuropathy is accompanied by more proximal axonal swellings. Toxicol
Appl Pharmacol 71:362-371.
Anthony, DC; Boekelheide, K; Anderson, CW; et al. (1983b) The effect of 3,4-dimethyl substitution on the
neurotoxicity of 2,5-hexanedione. II. Dimethyl substitution accelerates pyrrole formation and protein cross-linking.
Toxicol Appl Pharmacol 71:372-382.
ATSDR (Agency for Toxic Substances and Disease Registry). (1993a) Toxicological profile for automotive
gasoline. Available from ATSDR, Public Health Service, U.S. Department of Health and Human Services, Atlanta,
GA. Available from: .
ATSDR. (1993b) Toxicological profile for jet fuels JP-4 and JP-7. Available from ATSDR, Public Health Service,
U.S. Department of Health and Human Services, Atlanta, GA. Available from:
.
ATSDR. (1999) Toxicological profile for n-hexane. Available from ATSDR, Public Health Service, U.S.
Department of Health and Human Services, Atlanta, GA. Available from: .
Bachman, MO; De Beer, Z; Myers, JE. (1993) n-hexane neuorotoxicity in metal can manufacturing workers. Occup
Med 43:149-154.
128
-------
Baelum, J; Molhave, L; Hansen, SH; et al. (1998) Metabolic interaction between toluene, trichloroethylene and
n-hexane in humans. Scand J Work Environ Health 24:30-37.
Baker, T; Rickert, D. (1981) Dose-dependent uptake, distribution, and elimination of inhaled n-hexane in the
Fischer-344 rat. Toxicol Appl Pharmacol 61:414-422.
Barni-Comparini, I; Lungarella, G; Fonzi, L. (1982) Response of lung enzyme activities in rabbits following short-
term exposure to n-hexane: correlation between morphological and biochemical changes. Agents Actions 12:737-
742.
Bastone, A; Frontali, N; Mallozzi, C; et al. (1987) Cholinesterases in blood plasma and tissues of rats treated with
n-hexane or with its neurotoxic metabolite 2,5-hexanedione. Arch Toxicol 61:138-144.
Beall, C; Delzell, E; Rodu, B; et al. (2001) Case-control study of intracranial tumors among employees at a
petrochemical research facility. Journal of Occupational and Environmental Medicine 43:1103-1113.
Biodynamics Inc. (1978) 26-Week inhalation toxicity study of n-hexane in the rat. EPA Document No. FYI-AX-
1081-0137; NTIS No. OTS0000137-0.
Biodynamics Inc. (1989) A thirteen week inhalation toxicity study of commercial hexane in rat and mouse (final
report) with cover letter dated 122789. EPA Document No. FYI-AX-1189-0714; NTIS No. OTS0000714-3.
Biodynamics Inc. (1993a) Letter from American Petroleum Institute to U.S. EPA regarding an inhalation
oncogenicity study of commercial hexane in rats and mice: Part 1 (rats) with attachments, dated 04/16/93. Submitted
under Section 4 of TSCA. EPA Document No. 42084 L5-2; NTIS No. OTS0572989.
Biodynamics Inc. (1993b) Letter from American Petroleum Institute to U.S. EPA regarding an inhalation
oncogenicity study of commercial hexane in rats and mice: Part II (mice) with attachments, dated 06/03/93.
Submitted under Section 4 of TSCA. EPA Document No. 42084 L6-2; NTIS No. OTS0572994.
Bio-Research Laboratories. (1989) An acute operant behavior study of inhaled commercial hexane in the albino rat
(draft) with attachments and cover letter dated 122789. EPA Document No. FYI-AX-0190-0733; NTIS No.
OTS0000733.
Bio-Research Laboratories. (1990) A thirteen week inhalation toxicity study of commercial hexane on behavior and
neuromorphology in rats (final report) with attached reports, appendices & letter dated 021790. Submitted under
Section 4 of TSCA. EPA Document No. 40-9089428; NTIS No. OTS0524324.
Boekelheide, K. (1987) 2,5-hexanedione alters microtubule assembly. II. Enhanced polymerization of cross-linked
tubulin. Toxicol Appl Pharmacol 88:383-396.
BRRC (Bushy Run Research Center). (1989a) Developmental toxicity studies of commercial hexane vapor in CD
(Sprague-Dawley) rats and commercial hexane vapor in CD-I mice (final reports) with attachments & cover letter
111789. Submitted under Section 4 of TSCA. EPA Document No. 40-8989413; NTIS No. OTS0524323.
BRRC. (1989b) Draft results of the developmental toxicity exposure range-finding studies of commercial hexane
vapor in mice and rats with cover letter dated 033089. EPA Document No. FYI-AX-0489-0459; NTIS No.
OTS0000459-2.
BRRC. (1991) Two-generation reproduction study of inhaled commercial hexane in CD (Sprague-Dawley) rats
(final report) with attachments and cover letter dated 041791. Submitted under Section 4 of TSCA. EPA Document
No. 40-9189447; NTIS No. OTS0532897.
129
-------
Bus, J; White, E; Tyler, R; et al. (1979) Perinatal toxicity and metabolism of n-hexane in Fischer-344 rats after
inhalation exposure during gestation. Toxicol Appl Pharmacol 51:295-302.
Bus, J; White, E; Gillies, P; et al. (1981) Tissue distribution of n-hexane, methyl n-butyl ketone, and 2,5-
hexanedione in rats after single or repeated inhalation exposure to n-hexane. Drug Metab Dispos 9:386-387.
Bus, J; Demo, D; Cox, M. (1982) Dose-dependent disposition of n-hexane in F344 rats after inhalation exposure.
Fundam Appl Toxicol 2:226-229.
Canesi, M; Perbellini, L; Maestros, L; et al. (2003) Poor metabolization of n-hexane in Parkinson's disease. J Neurol
250:556-560.
Cardona, A; Marhuenda, D; Marti, J; et al. (1993) Biological monitoring of occupational exposure to n-hexane by
measurement of urinary 2,5-hexanedione. Int Arch Occup Environ Health 65:71-74.
Cardona, A; Marhuenda, D; Prieto, MJ; et al. (1996) Behaviour of urinary 2,5-hexanedione in occupational
coexposure to n-hexane and acetone. Int Arch Occup Environ Health 68:88-93.
Cavanaugh, JB, Bennetts, BJ. (1981) On the pattern of changes in the rat nervous system produced by
2,5-hexanediol. Brain 104: 297-318.
Cavender, FL; Casey, HW; Salem, H; et al. (1984a) A 13-week vapor inhalation study of n-hexane in rats with
empasis on neurotoxic effects. Fundam Appl Toxicol 4(2 Pt 1): 191-201.
Cavender, FL; Casey, HW; Gralla, EJ; et al. (1984b) The subchronic inhalation toxicity of n-hexane and methyl
ethyl ketone. Advances in Modern Environmental Toxicology 6:215-231.
Chang, YC. (1990) Patients with n-hexane-induced poly neuropathy: a clinical follow up. Br J Ind Med 47:485-489.
Chang, YC. (1991) An electrophysiological follow up of patients with n-hexane polyneuropathy. Br J Ind Med
48:12-17.
Chang, YC; Yip, PK. (1987) n-hexane-induced electroneurographic changes and early detection of n-hexane
intoxication. J FormosonMed Assoc 86:194-200.
Chang, CM; Yu, CW; Fong, KY; et al. (1992) N-hexane neuropathy in offset printers. Journal of Neurology,
Neurosurgery, and Psychiatry 56:538-542.
Cianchetti, C; Abbritti, G; Perticoni, G; et al. (1976) Toxic polyneuropathy of shoe-industry workers: a study of 122
cases. Journal of Neurology, Neurosurgery, and Psychiatry 39:1151-1161.
Couri, D; Milks, M. (1982) Toxicity and metabolism of the neurotoxic hexacarbons n-hexane, 2- hexanone, and
2,5-hexanedione. Annu Rev Pharmacol Toxicol 22:145-166.
Crosbie, SJ; Williams, FM; Blain, PG. (1994) The metabolism of n-hexane by rat skeletal muscle, liver, lung and
brain microsomes. Biochem Soc Trans 22:1308.
Crosbie, SJ; Blain, PG; Williams, FM. (1997) Metabolism of n-hexane by rat liver and extrahepatic tissues and the
effect of cytochrome. Human & Experimental Toxicology 16:131-137.
Daughtrey, WC; Neeper-Bradley, T; Duffy, J; et al. (1994a) Two-generation reproduction study on commercial
hexane solvent. J Appl Toxicol 14:387-393.
130
-------
Daughtrey, WC; Putman, DL; Duffy, J; et al. (1994b) Cytogenetic studies on commercial hexane solvent. J Appl
Toxicoll4:161-165.
Daughtrey, W; Newton, P; Rhoden, R; et al. (1999) Chronic inhalation carcinogenicity study of commercial hexane
solvent in F344 rats and B6C3F1 mice. Toxicol Sci 48:21-29.
DeCaprio, AP; Fowke, JH. (1992) Limited and selective adduction of carboxyl-terminal lysines in the high
molecular weight neurofilament proteins by 2,5-hexanedione in vitro. Brain Res 586:219-228.
DeCaprio, AP; Olajos, EJ; Weber, P. (1982) Covalent binding of a neurotoxic n-hexane metabolite: conversion of
primary amines to substituted pyrrole adducts by 2,5-hexanedione. Toxicol Appl Pharmacol 65:440-450.
DeCaprio, AP; Briggs, RG; Jackowski, SJ; et al. (1988) Comparative neurotoxicity and pyrrole-forming potential of
2,5-hexanedione and perdeuterio-2,5-hexanedione in the rat. Toxicol Appl Pharmacol 92:75-85.
DeCaprio, AP; Kinney, EA; Fowke, JH. (1997) Regioselective binding of 2,5-hexanedione to high-molecular-weight
rat neurofilament proteins in vitro. Toxicol Appl Pharmacol 145:211-217.
DeMartino, C; Malorni, W; Amantini, MC; et al. (1987) Effects of respiratory treatment with N-hexane on rat testis
morphology. I. A light microscopic study. Exp Mol Pathol 46:199-216.
Donate, R. (1999) Functional roles of S100 proteins, calcium-binding proteins of the EF-hand type. Biochim
Biophys Acta 1450(3): 191-231.
Donate, R. (2001) SI 00: a multigenic family of calcium-modulated proteins of the EF-hand type with intracellular
and extracellular functional roles. Int J Biochem Cell Biol 33: 637-668.
dos Santos, CR; Meyer Passarelli, MM; de Souza Nascimento, E. (2002) Evaluation of 2,5-hexanedione in urine of
workers exposed to n-hexane in Brazilian shoe factories. J Chromatogr B Analyt Technol Biomed Life Sci 778:237-
244.
Duffy, J; Newton, P; Cockrell, B; etal. (1991) A thirteen week inhalation toxicity study of commercial hexane in the
rat and mouse. Toxicologist 11:315.
Dunnick, JK; Graham, D; Yang, RS; et al. (1989) Thirteen-week toxicity study of n-hexane inB6C3Fl mice after
inhalation exposure. Toxicology 57:163-172.
EPL (Experimental Pathology Laboratories Inc.). (1989) A thirteen week inhalation toxicity study of commercial
hexane in the rat and mouse (pathology report) with attachment and cover letter dated 112289. EPA Document No.
FYI-AX-1189-0714; NTIS No. OTS0000714-1.
EPL. (1992) Partial study: 6-month continuous inhalation exposure of rats to hexane mixtures (re-evaluation of renal
pathology). EPA Document No. FYI-0892-0901; NTIS No. OTS0000901L6.
Fano, G; Biocca, S; Fulle, S; et al. (1995) The S-100: a protein family in search of a function. Prog Neurobiol 46:71-
82.
Fedtke, N; Bolt, HM. (1986) Detection of 2,5-hexanedione in the urine of persons not exposed to n-hexane. Int Arch
Occup Environ Health 57:143-148.
Fedtke, N; Bolt, HM. (1987) The relevance of 4,5-dihydroxy-2-hexanone in the excretion kinetics of n-hexane. Arch
Toxicol 61:131-137.
131
-------
ffrench-Constant, C; Colognato, H; Franklin RJM. (2004) Neuroscience: the mysteries of myelin unwrapped.
Science 304:688-689.
Fisher, J; Mahle, D; Bankston, L; et al. (1997) Lactational transfer of volatile chemicals in breast milk. Am Ind Hyg
Assoc 158:425-431.
Frontali, N; Amantini, MC; Spagnolo, A; et al. (1981) Experimental neurotoxicity and urinary metabolites of the C5-
C7 aliphatic hydrocarbons used as glue solvents in shoe manufacture. Clin Toxicol 18:1357-1367.
Gargas, M; Burgess, R; Voisard, D. (1989) Partition coefficients of low-molecular-weight volatile chemicals in
various liquids and tissues. Toxicol Appl Pharmacol 98:87-89.
Gelmont, D; Stein, RA; Mead, JF. (1981) Isoprene-the main hydrocarbon in human breath. Biochem Biophys Res
Commun 99: 1456-1460.
Center, MB, Szakal-Quin, G, Anderson, CW; et al. (1987) Evidence that pyrrole formation is a pathogenetic step in
y-diketone neuropathy. Toxicol Appl Pharmacol 87: 351-362.
Center St. Clair, MB; Amarnath, V; Moody, MA; et al. (1988) Pyrrole oxidation and protein cross-linking as
necessary steps in the development of gamma-diketone neuropathy. ChemRes Toxicol 1:179-185.
Ginsberg, G; Hattis, D; Sonawane, B; et al. (2002) Evaluation of child/adult pharmacokinetic differences from a
database derived from the therapeutic drug literature. Toxicol Sci 66:185-200.
Gobba, F; Cavalleri, A. (2003) Color vision impairment in workers exposed to neurotoxic chemicals.
Neurotoxicology 24:693-702.
Goel, SK; Rao, GS; Pandya, KP. (1982) Toxicity of n-hexane and n-heptane: some biochemical changes in liver and
serum. Toxicol Lett 14:169-174.
Goel, SK; Rao, GS; Krishnamurti, CR. (1987) Toxicity of n-hexane and 2,5-hexanediol to the hemopoietic system: a
preliminary report. Xenobiotica 17:217-222.
Goel, SK; Rao, GS; Pandya, KP. (1988) Hepatotoxic effects elicited by n-hexane or n-heptane. J Appl Toxicol 8:81-
84.
Governa, M; Calisti, R; Coppa, G; et al. (1987) Urinary excretion of 2,5-hexanedione and peripheral
polyneuropathies in workers exposed to hexane. J Toxicol Environ Health 20:219-228.
Graham, D; Anthony, DC; Boekelheide, K; et al. (1982a) Studies of the molecular pathogenesis of hexane
neuropathy. II. Evidence that pyrrole derivatization of lysyl residues leads to protein cross-linking. Toxicol Appl
Pharmacol 64:415-422.
Graham, D; Anthony, DC; Boekelheide, K. (1982b) In vitro and in vivo studies of the molecular pathogenesis of
n-hexane neuropathy. Neurobehav Toxicol Teratol 4:629-634.
Graham, D; Amarnath, V; Valentine, WM; et al. (1995) Pathogenetic studies of hexane and carbon disulfide
neurotoxicity. CritRev Toxicol 25:91-112.
Griffin, JW, Anthony, DC, Fahnestock, KE; et al. (1984) 3,4-Dimethyl-2,5-hexanedione impairs the axonal transport
of neurofilament proteings. JNeurosci4: 1516-1526.
Hageman, G; van der, HJ; van Hout, M; et al. (1999) Parkinsonism, pyramidal signs, polyneuropathy, and cognitive
132
-------
decline after long-term occupational solvent exposure. J Neurol 246:198-206.
Harry, GJ, Graham, DG, Valentine WM, et al. (1998) Carbon disulfide neurotoxicity in rats: VIII. Summary.
Neurotoxicology 19: 159-161.
Hazleton Laboratories. (1992) Initial submission: in vivo and in vitro mutagenicity studies n-hexane (hexane UV)
(final report) with attachments and cover letter dated 020592. Submitted under Section 8ECP of TSCA. EPA
Document No. 88-920000955; NTIS No. OTS0535721.
Houk, V; Schalkowsky, S; Claxton, L. (1989) Development and validation of the spiral Salmonella assay: an
automated approach to bacterial mutagenicity testing. Mutat Res 223:49-64.
Howd, R; Rebert, C; Dickinson, J; et al. (1983) Comparison of the rates of development of functional hexane
neuropathy in weanling and young adult rats. Neurobehav Toxicol Teratol 5:63-68.
HSDB (Hazardous Substances Data Bank). (2005) n-Hexane. National Library of Medicine, National Institutes of
Health, U.S. Department of Health and Human Services, Bethesda, MD. Available from:
.
Huang, CC; Chu, NS. (1989) Evoked potentials in chronic n-hexane intoxication. Clin Electroencephalogr 20:162-
168.
Huang, J; Kato, K; Shibata, E; et al. (1989) Effects of chronic n-hexane exposure on nervous system-specific and
muscle-specific proteins. Arch Toxicol 63:381-385.
Huang, CC; Shih, TS; Cheng, SY; et al. (1991) n-hexane polyneuropathy in a ball-manufacturing factory. J Occup
Med 33:139-142.
Huang, J; Shibata, E; Kato, K; et al. (1992) Chronic exposure to n-hexane induces changes in nerve-specific marker
proteins in the distal peripheral nerve of the rat. Human & Experimental Toxicology 11:3 23-3 27.
Iba, MM; Fung, J; Gonzalez, FJ. (2000) Functional Cyp2el is required for substantial in vivo formation of 2,5-
hexanedione from n-hexane in the mouse. Arch Toxicol 74:582-586.
Ichihara, G; Saito, I; Kamijima, M; et al. (1998) Urinary 2,5-hexanedione increases with potentiation of
neurotoxicity in chronic coexposure to n-hexane and methyl ethyl ketone. Int Arch Occup Environ Health 71:100-
104.
Ikeda, M; Koizumi, A; Kasahara, M; et al. (1986) Combined effects of n-hexane and toluene on norepinephrine and
dopamine levels in rat brain tissues after long-term exposures. Bull Environ Contam Toxicol 36:510-517.
Ikeda, T; Katakura, Y; Kishi, R; et al. (1993) Acute neurobehavioral effects of coinhalation of toluene and n-hexane
on schedule-controlled behavior in rats. Environ Res 63:70-81.
Imbriani, M; Ghittori, S; Pezzagno, G; et al. (1984) n-hexane urine elimination and weighted exposure
concentration. Int Arch Occup Environ Health 55:33-41.
IRDC (International Research and Development Corporation). (1986) Single-generation inhalation
reproductive/fertility study on a commercial hexane (Volume I-II). Submitted under Section 4 of TSCA. EPA
Document No. 40-8689172; NTIS No. OTS0524106.
IRDC. (1992a) 6-Month continuous inhalation exposures of rats to hexane mixtures - Phase I. EPA Document No.
FYI-0892-0901; NTIS No. OTS0000901K6.
133
-------
IRDC. (1992b) 6-Month continuous inhalation exposures of rats to hexane mixtures - Phase II. EPA Document No.
FYI-0892-0901; NTIS No. OTS0000901K5.
Iregren, A; Andersson, M; Nylen, P. (2002) Color vision and occupational chemical exposures: I. An overview of
tests and effects. Neurotoxicology 23:719-733.
Ishidate, M; Sofuni, T; Yoshikawa, K; et al. (1984) Primary mutagenicity screening of food additives currently used
in Japan. Food Chem Toxicol 22:623-636.
Issever, H; Malat, G; Sabuncu, HH; et al. (2002) Impairment of colour vision in patients with n-hexane exposure-
dependent toxic poly neuropathy. Occup Med 52:183-186.
Iwata, M; Takeuchi, Y; Hisanaga, N; etal. (1983) Changes of n-hexane metabolites in urine of rats exposed to
various concentrations of n-hexane and to its mixture with toluene or MEK. Int Arch Occup Environ Health 53:1-8.
Johnsrud, EK; Koukouritaki, SB; Divakaran, K; et al. (2003) Human hepatic CYP2E1 expression during
development. J Pharmacol Exp Ther 307:402-407.
Karakaya, A; Yucesoy, B; Burgaz, S; et al. (1996) Some immunological parameters in workers occupationally
exposed to n-hexane. Hum Exp Toxicol 15:56-58.
Kawachi, T; Yahagi, T; Kada, T; et al. (1980) Cooperative program on short-term assays for carcinogenicity in
Japan. IARC Sci Publ 217:323-330.
Kessler, W; Heilmaier, H; Kreuzer, P; et al. (1990) Spectrophotometric determination of pyrrole-like substances in
urine of rat and man: an assay for the evaluation of 2,5-hexanedione formed from n-hexane. Arch Toxicol 64:242-
246.
Khedun, SM; Maharaj, B; Naicker, T. (1996) Hexane cardiotoxicity-an experimental study. Israel Journal of
Medical Sciences 32:123-128.
Kim, M-S; Sabri, MI; Miller, VH; et al. (2001) 1,2 -Diacetylbenzene, the neurotoxic metabolite of a chromogenic
aromatic solvent, induces proximal axonopathy. Toxicol Appl Pharmacol 177: 121-131.
Kim, M-S; Hashemi, SB; Spencer PS; et al. (2002) Amino acid and protein targets of 1, 2-diacetylbenzene, a potent
aromatic y-diketone that induces proximal neurofilamentous axonopathy. Toxicol Appl Pharmacol 183: 55-65.
Kivits, GA; Ganguli-Swarttiywm MA; Christ, EJ. (1981) The composition of alkanes in exhaled air of rats as a result
of lipid peroxidation in vivo. Effects of dietary fatty acids, vitamins, and selenium. Biochim Biophys Acta 665: 559-
570.
Krasavage, WJ; O'Donoghue, JL; DiVincenzo, GD; et al. (1980) The relative neurotoxicity of methyl-n-butyl
ketone, n-hexane and their metabolites. Toxicol Appl Pharmacol 52:433-441.
Ladefoged, O; Perbellini, L. (1986) Acetone-induced changes in the toxicokinetics of 2,5-hexanedione in rabbits.
Scand J Work Environ Health 12: 627-629.
Ladefoged, O; Hass, U; Simonsen, L. (1989) Neurophysiological and behavioural effects of combined exposure to
2,5-hexanedione and acetone or ethanol in rats. Pharmacol Toxicol 65:372-375.
Ladefoged, O; Roswall, K; Larsen, JJ. (1994) Acetone potentiation and influence on the reversibility of
2,5-hexanedione-induced neurotoxicity studied with behavioural and morphometric methods in rats. Pharmacol
Toxicol 74:294-299.
134
-------
Lam, HR, Larsen, JJ, Ladefoged, O; et al. (1991) Effects of 2,5-hexanedione alone and in combination with acetone
on radial arm maze behavior, the "brain swelling" reaction and synaptosomal functions. Neurotoxicol Teratol 13:
407-412.
Lankas, G; Baxter, C; Christian, R. (1978) Effect of alkane tumor-promoting agents on chemically induced
mutagenesis in cultured V79 Chinese hamster cells. J Toxicol Environ Health 4:37-42.
Lapadula, DM; Irwin, RD; Suwita, E; et al. (1986) Cross-linking of neurofilament proteins of rat spinal cord in vivo
after administration of 2,5-hexanedione. Journal of Neurochemistry 46:1843-1850.
Larsen, JJ; Lykkegaard, M; Ladefoged, O. (1991) Infertility in rats induced by 2,5-hexanedione in combination with
acetone. Pharmacol Toxicol 69: 43-46.
Linder, RE; Strader, LF; Slott, VL; et al. (1992) Endpoints of spermatotoxicity in the rat after short duration
exposures to fourteen reproductive toxicants. Reprod Toxicol 6:491-505.
Litton Bionetics Inc. (1979) Teratology study in rats using n-hexane (final report) with cover letter dated 011587.
Submitted under Section 4 of TSCA. EPA Document No. 40-8789144; NTIS No. OTS0511634.
Litton Bionetics Inc. (1980) Mutagenicity evaluation of n-hexane in the mouse dominant lethal assay [final report].
EPA Document No. FYI-AX-1081-0137; NTIS No. OTS0000137-0.
Lungarella, G; Fonzi, L; Centini, F. (1980) Respiratory tract lesions induced in rabbits by short-term exposure to
n-hexane. Res Commun Chem Pathol Pharmacol 29:129-139.
Lungarella, G; Barni-Comparini, I; Fonzi, L. (1984) Pulmonary changes induced in rabbits by long-term exposure to
n-hexane. Arch Toxicol 55:224-228.
Mapleson, W. (1973) Circulation-time models of the uptake of inhaled anaesthetics and data for quantifying them.
Brit J Anaesth 45:319-334.
Marks, TA; Fisher, PW; Staples, RE. (1980) Influence of n-hexane on embryo and fetal development in mice. Drug
Chem Toxicol 3:393-406.
Masotto, C; Bisiani, C; Camisasca, C; et al. (1995) Effects of acute n-hexane and 2,5-hexanedione treatment on the
striatal dopaminergic system in mice. J Neural Transm Suppl 45:281-285.
Mast, T. (1987) Inhalation developmental toxicology studies: Teratology study of n-hexane in rats [final report].
Public Health Service, U.S. Department of Health and Human Services; TER90082. Prepared by the Pacific
Northwest Laboratory, Richland, WA, for the National Toxicology Program, National Institute for Environmental
Health Services, Research Triangle Park, NC; PNL-6453.
Mast, T; Decker, J; Stoney, K; et al. (1988a) Inhalation developmental toxicology studies: Teratology study of
n-hexane in mice [final report]. Public Health Service, U.S. Department of Health and Human Services; NTP
TER90083. Prepared by the Pacific Northwest Laboratory, Richland, WA, for the National Toxicology Program,
National Institute for Environmental Health Services, Research Triangle Park, NC; PNL-6590.
Mast, T; Hackett, P; Decker, J; et al. (1988b) Inhalation reproductive toxicology studies: sperm morphology study of
n-hexane in B6C3F1 mice. Prepared by the Pacific Northwest Laboratory Richland, WA, for the National
Toxicology Program, National Institute for Environmental Health Services, Research Triangle Park, NC; PNL-6672.
Available from: National Technical Information Service, Springfield, VA; NTIS No. DE89000262.
Mast, T; Rommerein, R; Evanoff, J; et al. (1988c) Inhalation reproductive toxicology studies: male dominant lethal
135
-------
study of n-hexane in Swiss (CD-I) mice. Prepared by the Pacific Northwest Laboratory Richland, WA, for the
National Toxicology Program, National Institute for Environmental Health Services, Research Triangle Park, NC;
PNL-6679. Available from: National Technical Information Service, Springfield, VA; NTIS No. DE89000271.
Mateus, ML; dos Santos Ana, PM; Batoreu, MC. (2002) Evidence of zinc protection against 2,5-hexanedione
neurotoxicity: correlation of neurobehavioral testing with biomarkers of excretion. Neurotoxicology 23:747-754.
Mayan, O; Teixeira, JP; Pires, AF. (2001) Biological monitoring of n-hexane exposure in Portuguese shoe
manufacturing workers. Appl Occup Environ Hyg 16:736-741.
Mayan, O; Teixeira, JP; Alves, S; et al. (2002) Urinary 2,5 hexanedione as abiomarker of n-hexane exposure.
Biomarkers 7:299-305.
Mayer, V; Coin, C. (1994) Induction of chromosome loss in yeast by combined treatment with neurotoxic
hexacarbons and monoketones. Mutat Res 341:83-91.
McCarroll, N; Keech, B; Piper, C. (1981a) A micro suspension adaptation of the Bacillus subtilis "rec" assay.
Environ Mutagen 3:607-616.
McCarroll, N; Piper, C; Keech, B. (1981b) AnE. coli microsuspension assay for the detection of DNA damage
induced by direct-acting agents and promutagens. Environ Mutagen 3:429-444.
Michailov, GV; Sereda, MW; Brinkmann, BG. (2004) Axonal Neuregulin-1 regulates myelin sheath thickness.
Science 304: 700-703.
Microbiological Associates. (1989) Salmonella/mammalian microsome mutagenicity assay of the vapor phase of
commercial hexane using the desiccator methodology (final report) with attached appendices. Submitted under
Section 4 of TSCA. EPA Document No. 40-8989391; NTIS No. OTS0524322.
Microbiological Associates. (1990) Subchronic in vivo cytogenetics assay in rats using nose-only inhalation
exposure to commercial hexane (final report) with attachments and cover letter dated 061790. Submitted under
Section 4 of TSCA. EPA Document No. 40-9089437; NTIS No. OTS0532896.
Mortelmans, K; Haworth, S; Lawlor, T; et al. (1986) Salmonella mutagenicity tests: II. Results from the testing of
270 chemicals. Environ Mutagen 8(Suppl 7): 1-119.
Murata, K; Araki, S; Yokoyama, K; et al. (1994) Changes in autonomic function as determined by ECG R-R interval
variability in sandal, shoe and leather workers exposed to n-hexane, xylene and toluene. Neurotoxicology 15:867-
875.
Mutti, A; Cavatorta, A; Lommi, G; et al. (1982a) Neurophysiological effects of long-term exposure to hydrocarbon
mixtures. ArchToxicol 5(Suppl): 120-124).
Mutti, A; Ferri, F; Lommi, G; et al. (1982b) n-hexane-induced changes in nerve conduction velocities and
somatosensory evoked potentials. Int Arch Occup Environ Health 51:45-54.
Mutti, A; Falzoi, M; Lucertini, S; et al. (1984) n-hexane metabolism in occupationally exposed workers. Br J Ind
Med 41:533-538.
Mutti, A; Bergamaschi, E; Ghittori, S; et al. (1993) On the need of a sampling strategy in biological monitoring: the
example of hexane exposure. Int Arch Occup Environ Health 65:8171-8175.
Nachtman, JP; Couri, D. (1984) An electrophysiological study of 2-hexanone and 2,5-hexanedione neurotoxicity in
136
-------
rats. Toxicol Lett 23:141-145.
Nakajima, T; Elovaara, E; Park, SS; et al. (1991) Immunochemical detection of cytochrome P450 isozymes induced
in rat liver by n-hexane, 2-hexanone and acetyl acetone. Arch Toxicol 65:542-547.
National Research Council (NRC). (1983) Risk assessment in the federal government: managing the process.
Washington, DC: National Academy Press.
National Safety Council (NSC). (2003) n-hexane. Chemical Backgrounder. National Safety Council, Itasca, IL.
Available from: .
National Toxicology Program (NTP). (1991) Toxicity studies of n-hexane inB6C3Fl mice (inhalation studies).
Public Health Service, U.S. Department of Health and Human Services; TOX-2. Available from: National Institute
for Environmental Health Services, Research Triangle Park, NC.
Nylen, P; Hagman, M. (1994) Function of the auditory and visual systems, and of peripheral nerve, in rats after long-
term combined exposure to n-hexane and methylated benzene derivatives. II. Xylene. Pharmacology & Toxicology
74:124-129.
Nylen, P; Ebendal, T; Eriksdotter-Nilsson, M; et al. (1989) Testicular atrophy and loss of nerve growth factor-
immunoreactive germ cell line in rats exposed to n-hexane and a protective effect of simultaneous exposure to
toluene or xylene. Arch Toxicol 63:296-307.
Nylen, P; Hagman, M; Johnson, AC. (1994) Function of the auditory and visual systems, and of peripheral nerve, in
rats after long-term combined exposure to n-hexane and methylated benzene derivatives. I. Toluene. Pharmacology
& Toxicology 74:116-123.
Ono, Y; Takeuchi, Y; Hisanaga, N. (1981) A comparative study on the toxicity of n-hexane and its isomers on the
peripheral nerve. Int Arch Occup Environ Health 48:289-294.
Ono, Y; Takeuchi, Y; Hisanaga, N; et al. (1982) Neurotoxicity of petroleum benzine compared with n-hexane. Int
Arch Occup Environ Health 50:219-229.
Passero, S; Battistini, N; Cioni, R; et al. (1983) Toxic polyneuropathy of shoe workers in Italy. A clinical,
neurophysiological and follow-up study. Italian Journal of Neurological Sciences 4:463-472.
Pastore, C; Marhuenda, D; Marti, J; et al. (1994) Early diagnosis of n-hexane-caused neuropathy. Muscle Nerve
17:981-986.
Patten, CJ; Ning, SM; Lu, AYH; et al. (1986) Acetone-inducible cytochrome P-450: purification, catalytic activity,
and interaction with cytochrome b5. Archives of Biochemistry and Biophysics 251:629-638.
Paulson, GW; Waylonis, GW. (1976) Polyneuropathy due to n-hexane. Arch Intern Med 136:880-882.
Perbellini, L; Brugnone, F; Faggionato, G. (1981) Urinary excretion of the metabolites of n-hexane and its isomers
during occupational exposure. Brit J Indust Med 38:20-26.
Perbellini, L; Leone, R; Fracasso, ME; et al. (1982) Metabolic interaction between n-hexane and toluene in vivo and
in vitro. Int Arch Occup Environ Health 50:351-358.
Perbellini, L; Brugnone, F; Caretta, D; et al. (1985) Partition coefficients of some industrial aliphatic hydrocarbons
(C5-C7) in blood and human tissues. Br J Ind Med 42:162-167.
137
-------
Perbellini, L; Mozzo, P; Brugnone, F; et al. (1986) Physiologicomathematical model for studying human exposure to
organic solvents: kinetics of blood/tissue n-hexane concentrations and of 2,5-hexanedione in urine. Br J Ind Med
43:760-768.
Perbellini, L; Mozzo, P; Olivato, D; et al. (1990) "Dynamic" biological exposure indexes for n-hexane and
2,5-hexanedione, suggested by a physiologically based pharmacokinetic model. Am Ind Hyg Assoc J 51:356-362.
Perocco, P; Bolognesi, S; Alberghini, W. (1983) Toxic activity of seventeen industrial solvents and halogenated
compounds of human lymphocytes cultured in vitro. Toxicol Lett 16:69-75.
Pezzoli, G; Barbieri, S; Ferrante, C; et al. (1989) Parkinsonism due to n-hexane exposure. Lancet 2:874.
Pezzoli, G; Ricciardi, S; Masotto, C; et al. (1990) n-hexane induces parkinsonism in rodents. Brain Res 531:355-
357.
Pezzoli, G; Antonini, A; Barbieri, S; et al. (1995) n-hexane-induced parkinsonism: pathogenetic hypotheses.
Movement Disorders 10:279-282.
Pezzoli, G; Strada, O; Silani, V; et al. (1996) Clinical and pathological features in hydrocarbon-induced
parkinsonism. AnnNeurol 40:922-925.
Prieto, MJ; Marhuenda, D; Roel, J; et al. (2003) Free and total 2,5-hexanedione in biological monitoring of workers
exposed to n-hexane in the shoe industry. Toxicol Lett 145:249-260.
Pryor, G; Dickinson, J; Howd, R; et al. (1983) Neurobehavioral effects of subchronic exposure of weanling rats.
Neurobehav Toxicol Teratol 5:47-52.
Raitta, C; Seppalainen, AN; Huuskonen, MS. (1978) N-hexane maculopathy in industrial workers. Albrecht Von
Graefes Arch Klin Exp Ophthalmol 209:99-110.
Ralston, WH; Hilderbrand, RL; Uddin, DE; et al. (1985) Potentiation of 2,5-hexanedione neurotoxicity by methyl
ethyl ketone. Toxicol Appl Pharmacol 81:319-327.
Ramsey, J; Andersen, ME. (1984) A physiologically based toxicokinetic description of the inhalation
pharmacokinetics of styrene in rats and humans. Toxicol Appl Pharmacol 73:159-175.
Robertson, P; White, E; Bus, J. (1989) Effects of methyl ethyl ketone pretreatment on hepatic mixed-function
oxidase activity and on in vivo metabolism of n-hexane. Xenobiotica 19:721-729.
Rosenberg, CK, Center, MB, Szakal-Quin, G; etal. (1987). dl-versus meso-3,4-dimethyl-2,5-hexanedione: a
morphometric study of the proximo-distal distribution of axonal swellings in the anterior root of the rat. Toxicol
Appl Pharmacol. 87: 363-373.
Sahu, SC; Lowther, DK; Jones, SL. (1982) Biochemical response of rat lung to inhaled n-hexane. Toxicol Lett
12:13-17.
Saito, I; Shibata, E; Huang, J; et al. (1991) Determination of urinary 2,5-hexanedione concentration by an improved
analytical method as an index of exposure to n-hexane. Br J Ind Med 48:568-574.
Sanagi, S; Seki, Y; Sugimoto, K; et al. (1980) Peripheral nervous system functions of workers exposed to n-hexane
at a low level. Int Arch Occup Environ Health 47:69-79.
Sanz, P; Flores, I; Soriano, T; et al. (1995) In vitro quantitative structure-activity relationship assessment of pyrrole
138
-------
adducts production by Y-diketone-forming neurotoxic solvents. Toxicol in Vitro 9:783-787.
Scelsi, R; Poggi, P; Fera, L; et al. (1980) Toxic polyneuropathy due to n-hexane. J Neurol Sci 47:7-19.
Scelsi, R; Poggi, P; Fera, L; et al. (1981) Industrial neuropathy due to n-hexane. Clinical and morphological findings
in three cases. Clin Toxicol 18:1387-1393.
Schaumburg, HH; Spencer, PS. (1976) Degeneration in central and peripheral nervous systems produced by pure
n-hexane: an experimental study. Brain 99: 183-192.
Schaumburg, H; Spencer, P. (1978) Environmental hydrocarbons produce degeneration in cat hypothalamus and
optic tract. Science 199:199-200.
Seppalainen, A; Raitta, C; Huuskonen, MS. (1979) n-hexane-induced changes in visual evoked potentials and
electroretinograms of industrial workers. Electroencephalogr Clin Neurophysiol 47:492-498.
Shelby, M; Witt, K. (1995) Comparison of results from mouse bone marrow chromosome aberration and
micronucleus tests. EnvironMol Mutagen 25:302-313.
Shell, L; Rozum, M; Jortner, B; et al. (1992) Neurotoxicity of acrylamide and 2,5-hexanedione in rats evaluated
using a functional observational battery and pathological examination. Neurotoxicol Teratol 14:273-283.
Shibata, E; Huang, J; Ono, Y; et al. (1990a) Changes in urinary n-hexane metabolites by coexposure to various
concentrations of methyl ethyl ketone and fixed n-hexane levels. Arch Toxicol 64:165-168.
Shibata, E; Huang, J; Hisanaga, N; et al. (1990b) Effects of MEK on kinetics of n-hexane metabolites in serum. Arch
Toxicol 64:247-250.
Sills, RC; Harry, GJ; Lowrey, KB; et al. (1998) Carbon disulfide neurotoxicity in rats: V. Morphology of axonal
swelling in the muscular branch of the posterior tibial nerve and spinal cord. Neurotoxicology 19:117-127.
Smith, AG; Albers, JW. (1997) n-hexane neuropathy due to rubber cement sniffing. Muscle & Nerve 20:1445-1450.
Sobue, I; lida, M; Yamamura, Y; et al. (1978) N-hexane polyneuropathy. Int J Neurol 11:317-330.
Soiefer, A; Robinson, K; Broxup, B; et al. (1991) A subchronic neurotoxicity study of commercial hexane vapor in
the rat. Toxicologist 11:315.
Soriano, T; Menendez, M; Sanz, P; etal. (1996) Method for the simultaneous quantification of n-hexane metabolites:
application to n-hexane metabolism determination. Human & Experimental Toxicology 15:497-503.
Spencer, PS; Schaumburg, HH. (1977a) Ultrastructural studies of the dying-back process III. The evolution of
experimental peripheral giant axonal degeneration. J Neuropathol Exp Neurol 36(2):276-299.
Spencer, PS; Schaumburg, HH. (1977b) Ultrastructural studies of the dying-back process IV. Differential
vulnerability of PNS and CNS fibers in experimental central-peripheral distal axonopathies. J Neuropathol Exp
Neurol 36(2):300-320.
Spencer, PS; Kim, M-S; Sabri, MI. (2002) Aromatic as well as aliphatic hydrocarbon solvent axonopathy. Int J Hyg
Environ Health 205:131-136.
Takeuchi, Y; Ono, Y; Hisanaga, N. (1981) An experimental study on the combined effects of n-hexane and toluene
on the peripheral nerve of the rat. Br JIndMed 38:14-19.
139
-------
Takeuchi, Y; Ono, Y; Hisanaga, N; et al. (1983) An experimental study of the combined effects of n-hexane and
methyl ethyl ketone. Br J Ind Med 40:199-203.
Takeuchi, Y. (1993) n-hexane polyneuropathy in Japan: a review of n-hexane poisoning and its preventive measures.
Environ Res 62:76-80.
Toftgard, R; Haaparanta, T; Eng, L; et al. (1986) Rat lung and liver microsomal cytochrome P-450 isozymes
involved in the hydroxylation of n-hexane. Biochemical Pharmacology 35:3733-3738.
U.S. EPA (Environmental Protection Agency). (1988) Recommendations for and documentation of biological values
for use in risk assessment. EPA 600/6-87/008. Available from: National Technical Information Service, Springfield,
VA; PB88-179874/AS.
U.S. EPA. (1991) Guidelines for developmental toxicity risk assessment. Federal Register 56(234):63798-63826.
U.S. EPA. (1994a) Interim policy for particle size and limit concentration issues in inhalation toxicity studies.
Federal Register 59(206):53799.
U.S. EPA. (1994b) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry. EPA/600/8-90/066F. Available from: National Technical Information Service, Springfield, VA; PB2000-
500023, and .
U.S. EPA. (1995) Use of the benchmark dose approach in health risk assessment. U.S. Environmental Protection
Agency. EPA/630/R-94/007. Available from: National Technical Information Service (NTIS), Springfield, VA;
PB95-213765, and .
U.S. EPA. (1996) Guidelines for reproductive toxicity risk assessment. Federal Register 61(212):56274-56322.
U.S. EPA. (1998a) Guidelines for neurotoxicity risk assessment. Federal Register 63(93):26926-26954.
U.S. EPA. (1998b) Science policy council handbook: peer review. Prepared by the Office of Science Policy, Office
of Research and Development, Washington, DC. EPA 100-B-98-001. Available from: National Technical
Information Service, Springfield, VA; PB98-140726, and .
U.S. EPA (1999) Guidelines for carcinogen risk assessment [review draft]. Risk Assessment Forum, Washington,
DC; NCEA-F-0644. Available from: .
U.S. EPA. (2000a) Science policy council handbook: peer review. 2nd edition. Prepared by the Office of Science
Policy, Office of Research and Development, Washington, DC. EPA 100-B-OO-OOl. Available from:
.
U.S. EPA. (2000b) Science policy council handbook: risk characterization. Prepared by the Office of Science Policy,
Office of Research and Development, Washington, DC. EPA 100-B-00-002. Available from:
.
U.S. EPA (2000c) Benchmark dose technical guidance document [external review draft]. EPA/630/R-00/001.
Available from: .
U.S. EPA. (2002) A review of the reference dose and reference concentration processes. Risk Assessment Forum,
Washington, DC; EPA/630/P-02/0002F. Available from: .
U.S.EPA. (2005a) Guidelines for carcinogen risk assessment. Risk Assessment Forum, Washington, DC;
EPA/630/P-03/001F. Available from: .
140
-------
U.S. EPA. (2005b) Supplementary guidance for assessing susceptibility from early-life exposure to carcinogens.
Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available from:
.
Valentino, M. (1996) Residual electroneurographic modifications in subjects with n-hexane-induced
poly neuropathy: a follow-up study. La Medicina Del Lavoro 87:289-296.
van Engelen, JG; Rebel-de Haan, W; Opdam, JJ; et al. (1997) Effect of coexposure to methyl ethyl ketone (MEK) on
n-hexane toxicokinetics in human volunteers. Toxicol Appl Pharmacol 144:385-395.
Vanacore, N; Gasparini, M; Brusa, L; et al. (2000) A possible association between exposure to n-hexane and
Parkinsonism. Neurological Sciences 21:49-52.
Vaz, AD; Coon, MJ. (1987) Hydrocarbon formation in the reductive cleavage of hydroperoxides by cytochrome
P-450. Proc Natl Acad Sci USA 84: 1172-1176.
Veronesi, B; Lington, A; Spencer, P. (1984) A tissue culture model of methyl ethyl ketone's potentiation of n-hexane
neurotoxicity. Neurotoxicology 5:43-52.
Veulemans, H; Van Vlem, E; Janssens, H; et al. (1982) Experimental human exposure to n-hexane. Study of the
respiratory uptake and elimination, and of n-hexane concentrations in peripheral venous blood. Int Arch Occup
Environ Health 49:251-263.
Wang, JD; Chang, YC; Kao, KP; et al. (1986) An outbreak of n-hexane-induced polyneuropathy among press
proofing workers in Taipei. Am J Ind Med 10:111-118.
Yamamura, Y. (1969) n-hexane polyneuropathy. Folia Psychiatr Neurol Jap 23:45-57.
Yokoyama, K; Feldman, RG; Sax, DS; et al. (1990) Relation of distribution of conduction velocities to nerve biopsy
findings in n-hexane poisoning. Muscle Nerve 13:314-320.
Yokoyama, K; Araki, S; Murata, K; et al. (1997) Postural sway frequency analysis in workers exposed to n-hexane,
xylene, and toluene: assessment of subclinical cerebellar dysfunction. Environ Res 74:110-115.
Yucesoy, B; Yucel, A; Erdem, O; et al. (1999) Effects of occupational chronic coexposure to n-hexane, toluene, and
methyl ethyl ketone on NK cell activity and some immunoregulatory cytokine levels in shoe workers. Human &
Experimental Toxicology 18:541-546.
Zhao, W; Misumi, J; Yasui, T; et al. (1998) Effects of methyl ethyl ketone, acetone, or toluene coadministration on
2,5-hexanedione in the sciatic nerve, serum, and urine of rats. Int Arch Occup Environ Health 71:236-244.
141
-------
APPENDIX A
SUMMARY OF EXTERNAL PANEL PEER REVIEW AND PUBLIC COMMENTS
AND DISPOSITION
The support document and IRIS summary for n-hexane have undergone both Agency
review performed by scientists within EPA and a more formal external peer review performed by
scientists in accordance with EPA guidance on peer review (U.S. EPA, 1998b, 2000a).
Comments made by the Agency reviewers were addressed prior to submitting the documents for
external peer review and are not part of this appendix. The five external peer reviewers were
tasked with providing written answers to general questions on the overall assessment and on
chemical-specific questions in areas of scientific controversy or uncertainty. A summary of
significant comments made by the external reviewers and the public and EPA's response to these
comments follows.
EXTERNAL PEER REVIEW PANEL COMMENTS
The reviewers made several editorial suggestions to clarify specific portions of the text.
These changes were incorporated in the document as appropriate and are not discussed further.
1) Oral reference dose (RfD) for n-hexane
No oral RfD has been derived. Have the rationale and justification for not deriving
an RfD been transparently described in the documents? Are there additional
studies that should be considered in this decision?
COMMENT: All of the reviewers felt that the rationale and justification for not deriving an
RfD had been transparently described in the assessment, and no additional studies were available
for consideration in this decision.
RESPONSE: No response.
2) Inhalation reference concentration (RfC) for n-hexane
A-l
-------
a) Have the rationale and justification for deriving an RfC been transparently
described in the documents? Are there additional studies that should be considered
in this decision?
COMMENT: One reviewer commented that some of the information presented in the document
was repeated in multiple sections. This reviewer specifically referred to Altenkirch et al. (1982)
and Ichihara et al. (1998). This reviewer, along with another reviewer, also suggested more
synthesis of the literature presented in the document.
In addition, one reviewer noted that Tables 3-2 and 3-3 showed unexplained differences of
n-hexane tissue levels following the same exposure scenario. This reviewer was also unclear as
to why exposure units were changed from ppm to mg/m3 in Table 4-22.
All of the remaining reviewers indicated that the rationale and justification for deriving the RfC
was adequate and that no additional studies should have been considered in the derivation of the
RfC.
RESPONSE: Sections of the document where studies were described multiple times were
revised so that descriptions were not repeated. The literature is extensively synthesized and
interpreted in Sections 4.5 and 5.2.1.
The discrepancies between Tables 3-2 and 3-3 have been clarified. Table 3-2 presents n-hexane
tissue levels measured 72 hours after a 6-hour exposure and Table 3-3 presents n-hexane tissue
levels measured immediately following a 6-hour exposure. Finally, Table 4-22 presents
exposure concentrations in both ppm and mg/m3.
b) The 1990 IRIS assessment for n-hexane used a human occupational exposure
study by Sanagi et al. (1980) for the derivation of the RfC. The draft reassessment
for n-hexane uses a subchronic rat study by Huang et al. (1989) for the derivation of
the RfC. The workers evaluated in the Sanagi et al. (1980) study had coexposure to
acetone and n-hexane. Data were identified that indicate n-hexane metabolism and
n-hexane-induced neurotoxicity are potentiated by coexposure to acetone. Thus,
this study was not selected for the derivation of the RfC in the current assessment.
A-2
-------
The rationale supporting selection of the Huang et al. (1989) study versus the Sanagi
et al. (1980) study as the principal study in the derivation of the RfC is presented in
Sections 5.2.1 and 5.2.4 of the Toxicological Review. Is the Huang et al. (1989) study
the most appropriate selection for the principal study (i.e., best study upon which to
determine the point of departure)? Has the rationale for this choice been
transparently and objectively described? Is the selection of Huang et al. (1989) as
the principal study scientifically objective? Is the exclusion of Sanagi et al. (1980) as
the principal study based on coexposure to acetone justified? Should the Huang et
al. (1989) study and the Sanagi et al. (1980) study be considered as coprincipal
studies in the derivation of the RfC?
COMMENT: Three reviewers felt that the selection of the Huang et al. (1989) study as the
principal study was appropriate and that the rationale for the selection of this study as the
principal study was adequately described in the document. These reviewers did agree that the
human study by Sanagi et al. (1980) provided information useful to the assessment and should be
discussed and integrated into the discussion of principal study selection to a greater extent.
One reviewer felt that both the animal study (Huang et al., 1989) and the human study (Sanagi et
al., 1980) should be considered as coprincipal studies in the derivation of the RfC. This reviewer
noted that the coexposure to acetone in the human study (Sanagi et al., 1980) did confound the
results of this study. The reviewer stated that the coexposure to acetone in the human study
would lead to the derivation of a more conservative RfC. Further this reviewer agreed that the
58 ppm exposure presented in the human study represents a LOAEL.
A single reviewer indicated that the selection of the human study (Sanagi et al., 1980) as the
principal study for the derivation of the RfC was appropriate in this assessment. The reviewer
also stated that the coexposure to acetone in the human study would lead to the derivation of a
more conservative RfC. This reviewer suggested that the animal study (Huang et al., 1989) be
presented as supporting data.
RESPONSE: EPA has determined that the Huang et al. (1989) animal study is the most
appropriate study upon which to determine the point of departure for the derivation of the RfC
for n-hexane. EPA has considered the reviewers' comments collectively and decided that the
exclusion of the human study by Sanagi et al. (1980) as the principal study is scientifically
A-3
-------
justified based on the coexposure of the study subjects to acetone and n-hexane. Specifically,
more recent data indicate that acetone increases the metabolism of n-hexane to the toxic
metabolite 2,5-hexanedione and decreases elimination of 2,5-hexanedione (see Section 5.2.1). In
addition, acetone has been shown to exacerbate n-hexane-induced neurotoxicity and
reproductive toxicity (see Sections 4.4.3 and 5.2.1). Given the uncertainty associated with the
contribution of acetone to the neurological effects observed in humans by Sanagi et al., (1980),
the Huang et al. (1989) animal study is most appropriate to be used in the derivation of the RfC
for n-hexane. The Sanagi et al. (1980) human study is used as supporting data in the selection of
the principal study and critical effect for the derivation of the RfC for n-hexane.
c) Has the most appropriate critical effect (decreased motor nerve conduction
velocity in male rats following 12 weeks n-hexane exposure) been selected? Has the
rationale and justification for this effect been transparently described? Is the
selection of the critical effect scientifically justified?
COMMENT: All of the reviewers agreed that the selection of decreased MCV as the critical
effect was the most appropriate endpoint for the derivation of the RfC. The reviewers felt that
the rationale and justification were transparently presented and that the selection of this endpoint
was scientifically justified. One reviewer suggested that a brief statement of the sensitivity of
this endpoint compared with decreased hind-limb grip strength be added to the assessment.
RESPONSE: A statement indicating the sensitivity of measurement of motor nerve conduction
velocity compared with hind-limb grip strength has been added to the document (see Section
4.5).
d) An RfC has been derived utilizing benchmark dose modeling to define the point
of departure. Is benchmark dose modeling the best approach for determining the
point of departure? Has the benchmark dose modeling been accurately and
transparently described? In the absence of a biological rationale for choosing an
appropriate effect level, a point of departure corresponding to a change in the mean
equal to one control standard deviation from the control mean has been used. Is
this the best approach for determining the effect level? Has the most appropriate
model been utilized? Please comment on the model choice (and the values utilized
for the model parameters) as well as the approach.
A-4
-------
COMMENT: All reviewers agreed with the use of benchmark dose modeling for the derivation
of the RfC for n-hexane. Two reviewers requested further explanation of the benchmark
response level (one standard deviation versus 1, 5, or 10%) and suggested perhaps a lower point
of departure should be used to derive the RfC to be more health protective. One of these
reviewers suggested the application of an additional uncertainty factor to account for severity of
effect.
One reviewer indicated that in Section 5.2.2 and/or Appendix B, the connection between the
benchmark dose method and software (i.e., what input options and parameters were used) be
added to the discussion and that the actual mathematical equation and an explanation for
selection of the various models be included. This reviewer also requested clarification of the
parameters used in modeling the Huang et al. (1989) 8- and 12-week exposure data.
One reviewer noted that the BMD modeling method used (fetal-based analysis) for the reduced
fetal body weight data from Mast (1987) was incorrect. Specifically, this reviewer indicated that
a litter-based analysis would be more appropriate. This reviewer also suggested that the data
from the Mast et al. (1988a) study should be modeled.
RESPONSE: The U.S. EPA Benchmark Dose Technical Guidance Document (U.S. EPA,
2000c) states that in the absence of any data indicating the level of response that is considered to
be biologically significant or adverse, a change in the mean equal to one control standard
deviation from the control mean can be used. This gives an excess risk of approximately 10%
for the proportion of individuals below the 2nd percentile and above the 98th percentile of controls
for normally distributed effects. For the Huang et al., (1989) data set, an approximate 7%
change in the mean MCV value corresponds to the point of departure. This percentage is an
indication of the severity of the critical effect. The U.S. EPA does not have guidance specific to
the application of an effect level extrapolation factor to take into account severity of effect.
Thus, this factor is not generally applied in IRIS assessments and was not applied to the point of
departure for the derivation of the RfC for n-hexane. Based on the severity of effect and the
recommendations of the U.S. EPA Benchmark Dose Technical Guidance Document (U.S. EPA,
2000c) it was determined that a BMR of one control standard deviation from the control mean is
appropriate for the derivation of the RfC for n-hexane.
Significant text has been added to Appendix B describing the input options and parameters used
A-5
-------
in modeling of the results from the Huang et al. (1989), Ono et al. (1982), and Mast (1987)
studies. Where appropriate, the mathematical equations and accompanying explanations for
selection of one model over another have also been added to the text in Appendix B. Finally,
clarification of parameters used in the BMD modeling of the Huang et al. (1989) data has been
added to the text in Appendix B.
The reduced fetal body weight data from Mast (1987) has been modeled using a litter-based
analysis rather than a fetal-based analysis to address correlation among the pups. The revised
modeling results and output are presented in Section 5.2.2 and Appendix B. The Mast et al.
(1988a) study showed a statistically significant decreased fetal body weight and increased
incidence of late resorptions at the highest dose. In addition, the only other statistically
significant finding was an increased mean percent intrauterine death (increased early and late
resorptions combined) at the low dose, with no dose-related trend. Thus, this study was not
considered further for the selection of the principal study.
e) Are the uncertainty factors applied to the point of departure for the derivation of
the RfC scientifically justified and transparently and objectively described in the
Toxicological Review?
COMMENT: Two reviewers suggested that an effect level extrapolation factor be applied to the
point of departure to account for use of a BMR of 1 standard deviation.
Two reviewers felt that the uncertainty factor applied for extrapolation from a subchronic to a
chronic study was inappropriate. One reviewer indicated that the length of duration of the
principal study (16 weeks) is half the time required for a newly synthesized NF protein to be
transported from the neuronal cell body to the axon terminal in the longest PNS and CNS axons
of an adult rat. Thus, this reviewer felt that there is no justification to extrapolate to a lifetime of
the rat since the lifetime (maximum 32 weeks) of the target tissue, the NF, is shorter. The
second reviewer stated that "given what is known about the toxic mechanism," the full
uncertainty factor should not be applied.
Two reviewers suggested that the intraspecies uncertainty factor applied to the point of departure
for the derivation of the RfC for n-hexane was inappropriate. Specifically, one reviewer noted in
one study younger rats were observed to be less susceptible to the effects of n-hexane-induced
A-6
-------
neurotoxicity compared with adult rats. This reviewer suggested that this difference was due to
shorter length axons in the CNS and PNS in young compared with adult rats. Thus, since the
RfC was derived from data in adult rats, any variation seen between young and adult rats would
be observed as a decreased sensitivity to the effects of n-hexane-induced neurotoxicity in young
rats. The reviewer further commented that no data exist that report differences in absorption,
distribution, metabolism, or elimination. This reviewer noted that if concerns regarding genetic
polymorphisms were included, then the UF of 10 for intraspecies differences was appropriate.
The second reviewer provided no specific rationale for the reduction of the intraspecies
uncertainty factor.
All of the reviewers suggested that the application of a database uncertainty factor of 3 was
reasonable, but several of them stated that the justification for this determination was inadequate.
Specifically, concerns about increased susceptibility in the fetus are not supported by the data.
RESPONSE: The uncertainty factor for extrapolation from subchronic to chronic exposure has
been reduced from 10 to 3 due to considerations of NF biology. Specifically, 16 weeks (the
duration of the principal study by Huang et al., 1989) is half the time required for a newly
synthesized NF protein to be transported from the neuronal cell body to the axon terminal of the
longest axon in and adult rat. The longest axon of an adult rat stretches from the lumbar spinal
cord to the hind foot (approximately 22 cm) and NFs are transported axonally in adult rats at a
rate of 1 mm/day. Complete transport of the newly synthesized NF protein the full length of the
axon takes approximately 32 weeks. Thus, since the lifetime of the target of n-hexane-induced
toxicity is much less than the lifetime of the adult rat, extrapolation from subchronic to chronic
exposure is not justifiable. Additional text has been added to Section 5.2.2 to justify this
reduction in the uncertainty factor for extrapolation from a subchronic to a chronic exposure.
Information provided by reviewers concerning differences in susceptibility between adult versus
young rat due to axonal length has also been added to the document (see Sections 4.4.4 and
5.2.3). A single study in rats with one dose group reported susceptibility differences between
adult and weanling rats (adults showed earlier and more severe neurotoxic effects). Mode of
action studies supporting the hypothesis that the possible increased susceptibility to n-hexane-
induced neurotoxic effects in adults is due to differences in axonal length exist (Michailov et al.,
2004; ffrench-Constant et al., 2004; Graham et al., 1995; Griffin et al., 1984; Cavanagh and
Bennetts, 1981; Spencer and Schaumburg, 1977a, b; Schaumburg and Spencer, 1976). However,
A-7
-------
these studies do not report data for weanling animals. Given the potential for altered metabolism
of n-hexane in individuals with polymorphisms in the CYP2E1 enzyme and the uncertainties
associated with susceptibility differences between weanling and adult rats, a UF of 10 for
intraspecies variability was retained.
f) The database for n-hexane is lacking a developmental neurotoxicity study. Given
the potential increased susceptibility of the developing fetus to n-hexane-induced
toxicity and the increased neurotoxicity in humans and animals following n-hexane
exposure, a UFDB of 3 was applied. Has the rationale and justification for the UFDB
been transparently described? Is the application of this UF appropriate?
COMMENT: All of the reviewers indicated that the database uncertainty factor that was applied
for the derivation of the RfC for n-hexane was appropriate and reasonable. Several reviewers
asked for increased justification and clarification. Specifically, several reviewers had concerns
that increased susceptibility in the fetus is not supported by the data.
RESPONSE: Clarification and justification for the database uncertainty factor have been added
to the document. Specifically, information has been added to Section 5.2.3 describing in more
detail the reproductive studies available with exposures to pure n-hexane and the multigeneration
reproductive and developmental toxicity studies with a mixture containing n-hexane. The
database for n-hexane includes human occupational exposure studies (all with coexposure to
other potentially neurotoxic chemicals), subchronic animal studies, and developmental studies in
animals. The database lacks a multigeneration reproductive and developmental toxicity study.
Prenatal exposure to pure n-hexane was shown to induce skeletal anomalies, decreased fetal
body weight, and increased resorptions (Mast et al., 1988a; Mast, 1987; Bus et al., 1979). In
addition, the multigeneration reproductive and toxicity studies with the commercial hexane
mixture containing n-hexane suggested that n-hexane may induce reductions in fetal body weight
and increased skeletal anomalies (Daughtrey et al., 1994a; BRRC, 1991, 1989a, b; IRDC, 1986).
Taken together, these studies indicate that n-hexane may have effects on the reproductive system
or developing fetus. However, the data do not elucidate whether low doses of n-hexane (i.e.,
doses that have been shown to lead to neurotoxicity in humans and animals) cause reproductive
and developmental effects. Thus, due to the lack of multigeneration reproductive and
developmental toxicity studies evaluating the effects of low doses of exposure to pure n-hexane,
a UFDB of 3 was applied.
A-8
-------
3) Carcinogenicity of n-hexane
Under EPA's Draft Revised Guidelines for Carcinogen Risk Assessment (U.S. EPA,
1999) (www.epa.gov/ncea), data are inadequate for an assessment of the human
carcinogenic potential of n-hexane. Do the available data support this statement?
Are there additional studies that should be considered in this decision?
How well were the data from individual studies characterized and are the
conclusions that are drawn from each study valid? How well are the data
integrated into an overall conclusion and characterization of hazard as presented in
Sections 4.5, 4.6, 5, and 6?
COMMENT: All of the reviewers agreed with the characterization of n-hexane data as
inadequate for the assessment of human carcinogenic potential. Two reviewers provided
comments for clarification of the cancer data.
RESPONSE: Suggestions for clarification were added to the document. The reference has been
changed to the 2005 U.S. EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).
PUBLIC COMMENTS
Two public comment submissions were received during the external peer review and
public comment period. These comments were distributed to the external peer review panel
prior to the public panel meeting and discussion of the assessment. Both submissions had
several editorial suggestions and requested clarifications to specific portions of the text. These
changes were incorporated in the document as appropriate and are not discussed further.
1) Inhalation reference concentration (RfC) for n-hexane
COMMENT: A single commenter indicated that the selection of the human study (Sanagi et al.,
1980) as the principal study for the derivation of the RfC for n-hexane was appropriate. The
commenter noted that use of the Sanagi et al. (1980) study would provide a more conservative
RfC for n-hexane due to the coexposure to acetone and possible potentiation of n-hexane-
induced neurotoxic effects. In addition, this commenter suggested that the n-hexane exposure
A-9
-------
reported in the human study by Sanagi et al. (1980) is a minimal LOAEL or a NOAEL for the
following reasons: (1) the authors' conclusions regarding the significance of the reported
findings; (2) differences in exposure conditions between exposed and control groups (i.e., ball
milling operations workers); (3) significance of electrophysical findings; and (4) inconsistencies
in the magnitude of the findings.
RESPONSE: All of the human occupational n-hexane exposure studies also contain coexposure
to other chemicals, many of which have been shown to potentiate n-hexane-induced
neurotoxicity and metabolism. Specifically, Sanagi et al. (1980) reported coexposure to acetone,
a chemical that has been shown to exacerbate metabolism of n-hexane to its toxic metabolite,
2,5-hexanedione, and n-hexane-induced neurotoxicity and reproductive toxicity. In addition,
neurological deficits such as vibration sensation loss and headache have been reported to result
from exposure to high noise levels such as those present in the factory where workers observed
by Sanagi et al. (1980) worked. The contribution of both acetone coexposure and noise to
n-hexane-induced neurotoxicity is unknown.
The external peer review panel felt that the effects observed by Sanagi et al. (1980) did
not represent a NOAEL or a minimal LOAEL but rather they agreed that the n-hexane exposure
level represented a LOAEL. In addition, EPA has determined that the Huang et al. (1989)
animal study is the most appropriate study upon which to determine the point of departure for the
derivation of the RfC for n-hexane. EPA has considered the external peer review panel's
comments collectively and decided that the exclusion of the human study by Sanagi et al. (1980)
as the principal study is scientifically justified based on the coexposure of the study subjects to
acetone and n-hexane (see response above).
COMMENT: A single comment was received requesting additional clarification for the basis of
the selection of the BMDL for the 12-week exposure in male rats as opposed to the 8- or 16-
week exposures in the Huang et al. (1989) study.
RESPONSE: The goodness of fit and AIC values for the 8- ,12-, and 16-week exposures were
relatively comparable for the Huang et al. (1989) male rat MCV data set. BMCL values from
these data sets were within twofold of each other, suggesting that all the modeling results are
equally plausible. The lowest point of departure, based on decreased MCV at 12-weeks
exposure (BMCL = 215 mg/kg-day) was selected for derivation of the RfD corresponding to the
A-10
-------
most sensitive neurological effect. In addition, a discussion of the modeling of the 8, 12, and 16-
week n-hexane exposure data has been added to Appendix B.
2) Database uncertainty factor for the derivation of the RfC for n-hexane
COMMENT: Both public comment submissions suggested that the database uncertainty factor
of 3 was inappropriate given the combined databases for n-hexane and commercial hexane and
the evidence for the mode of action of n-hexane-induced neurotoxicity. These commenters also
felt that the concern for increased susceptibility to n-hexane-induced neurotoxicity in the
developing fetus is contradictory to evidence suggesting that younger animals may be less
sensitive to n-hexane-induced neurotoxicity.
RESPONSE: Clarification and justification for the database uncertainty factor have been added
to the document. Specifically, information has been added to Section 5.2.3 detailing the only
study reporting susceptibility differences between weanlings and adults (Howd et al., 1983) and
additional multigeneration reproductive and developmental toxicity studies with a mixture
containing n-hexane (Daughtrey et al., 1994a; BRRC, 1991, 1989a, b; IRDC, 1986). Due to the
lack of studies investigating the susceptibility of weanling and young animals compared with
adult animals and multi-generation reproductive and developmental toxicity studies evaluating
the effects of pure n-hexane exposure on the fetus, a UFDB of 3 was applied (see response above).
A-ll
-------
APPENDIX B: BENCHMARK DOSE (BMD) ANALYSIS
Benchmark dose modeling was performed on the Mast (1987) data sets for reduced
ossification of sternebrae and reduced fetal body weight, the Ono et al. (1982) data sets for MCV
and mixed MCVs, and the Huang et al. (1989) data sets for reduced MCV. The criteria for
selecting models used here are consistent with recommendations in the EP'A''s Benchmark Dose
Technical Guidance Document (U.S. EPA, 2000c) except as noted where model selection criteria
may disagree in ranking the models.
MODELING OF DICHOTOMOUS DATA SETS
In rats, Mast (1987) reported a statistically significant increase in the incidence of litters
with fetuses with reduced ossification of sternebrae 1-4 (dichotomous data). Dichotomous data
from this study were analyzed by BMDS and examined for toxicological relevance. All nested
models for dichotomous variables available in EPA's BMDS (version 1.3.2) were fit to the data
presented in the BMDS output in this appendix. The nested logistic (NL), National Center for
Toxicological Research (NCTR), and Rai and vanRyzin (RvR) models allow for the possibility
that the variance among the proportions of pups affected in individual litters is greater than
would be expected if the pups were responding completely independently of each other (U.S.
EPA, 2000c). A 10% increase in the incidence of reduced ossification of sternebrae 1-4 was
selected as the benchmark response because it was a response rate that fell within the range of
experimental dose levels used in the Mast (1987) study. All of the models provided similar fits
to the data, based on the summary results reported in the BMDS output and the detailed
examination of graphs and goodness-of-fit statistics. Model fits were improved slightly by the
incorporation of litter size (as a litter-specific covariate) and by intra-litter correlations, as
determined by comparisons of AIC values.
Modeling of reduced ossification of sternebrae 1—4 litter incidence data (Mast, 1987)
A summary of the incidence data for reduced ossification of sternebrae 1-4 in individual
fetuses for both the control and exposure groups is shown in the output for the BMD modeling
results of the data presented in this appendix.
B-l
-------
The best-fitting models used the total number of nonresorbed pups in the litter as the
litter-specific covariate that had the smallest AIC compared with comparable models with other
litter-specific covariates. The NL model with the default parameter specification n greater than
or equal to one was fit to the data and compared with the NCTR and RvR models with similar
specifications. The default option forced the value of the exponent to be set equal to one. While
the RvR models achieved a small AIC, they were usually not able to estimate a BMCL value and
so are not discussed further. The NL model in the BMDS nested module produced the highest
goodness of flip-values. Values were larger than 0.10 over a range of parameter values and
model specifications. While the NCTR models had slightly smaller AIC statistics than
comparable NL models, the differences were less than 2 AIC units, and none of the NCTR
models had ^-values greater than 0.10. We therefore selected the NL models as providing the
best description of the data. The default option forced the value of the exponent to be set equal
to one. The NL model had a better goodness of fit and smaller AIC. This model provided an
AIC of 3.41 with a/>-value of 0.183. The corresponding BMC and BMCL values were 1571 and
943 ppm, respectively.
MODELING OF CONTINUOUS DATA SETS
Continuous data models (linear, polynomial, power, or Hill where possible), either with a
constant variance or with variance as a power function of the mean value (using an additional
model parameter), were fit to the following continuous data sets from the Huang et al. (1989),
Ono et al. (1982), and Mast (1987) studies which were selected for benchmark dose modeling:
decreased MCV, decreased MCV and mixed nerve conduction velocity, and fetal body weight
deficits, are the respective endpoints. The BMDS (version 1.3.2) was used to calculate potential
points of departure for deriving the RfC by estimating the effective dose at a specified level of
response (BMC) and its 95% lower bound (BMCL). EPA's Benchmark Dose Technical
Guidance Document (U.S. EPA, 2000c) recommends, in the absence of some idea of the level of
response to consider adverse, selecting as the BMR level for continuous data a change in the
mean equal to one SD from the control mean.
The variance (square of the standard deviation) of the within-dose-group continuous
endpoints is modeled in general by a two-parameter power function independently of the dose-
response model for the within-dose-group mean response. This is done by fitting a model to the
sample variances within each dose group, based only on the within-dose-group sample mean and
not on the modeled dose-response function for the mean responses. Two situations lend
themselves to a priori specification of the exponent in the variance function model, denoted rho.
The first situation is to specify rho = 0, a situation often encountered in continuous models in
B-2
-------
which the random errors appear to have a constant within-dose-group variability at all doses, as
seems to be the case for the Mast (1987) fetal weight data described below. Another common
situation is when the standard deviation of the within-dose-group data increases in direct
proportion to the mean so that the ratio of the SD to the mean within each dose group (the
within-dose-group coefficient of variation) is nearly constant, thus rho = 2. This case more
accurately describes the Ono et al. (1982) MCV data. The variance model or specification of rho
affects the influence (statistical weight) of observed mean values in each dose group and was
considered in the analysis.
Modeling of Reduced MCV and Reduced Mixed Nerve Conduction Velocity
The Ono et al. (1982) study used three dose groups of 0, 200, and 500 ppm. The Hill
model in BMDS has four adjustable parameters in the dose-response function and cannot be
reliably estimated using data with only three dose groups without specifying the value of at least
one parameter in either the dose-response model or the variance function model. Even when the
Hill model can be fitted to the data, over-parameterizing the model may not make it possible to
evaluate the goodness of fit of the model to the data or to estimate the BMCL. The Hill model
was not evaluated in the Ono et al. (1982) study as there was little basis for specifying a large
number of parameters in that model. The overall means and standard deviations for reduced
MCV following inhalation exposure to n-hexane from the Ono et al. (1982) study are presented
below in Table B-l.
Table B-l. Mean nerve conduction velocity in rats exposed to n-hexane
n-hexane exposure group
0
200
500
Number of
rats
8
8
8
Mean nerve
conduction velocity
49.6
42.9
36.2
Standard
deviation of
MCV
7.6
3.5
1.1
Source: Onoetal., 1982.
The models that were evaluated included the polynomial model of degree 1 (linear),
the polynomial model of degree 2 (quadratic), and the power function model. Neither the
linear nor the quadratic model provided adequate goodness of fit [all P(H4) < 0.0001] for any
model specification. None of the power function models provided an adequate fit to the Ono
et al. (1982) data unless one or two parameters of the model were specified. The best
estimates of the shape parameter or exponent in the power function model were in the range of
-------
0.7 to 0.8 and only provided BMCL estimates when the value of the exponent or power (rho)
in the variance function model was specified as being in the range of approximately 2 to 3.
Table B-2 below shows some of the BMC modeling results.
The sensitivity of the BMCL to the exponent was explored by holding y constant at
some middle values within the confidence interval for that parameter and setting rho as a
constant over the range of 2 to 4. An adequate fit of the BMDS power function model can be
obtained over a wide range of parameter values (Table B-l). However the BMCL estimates
among acceptable models vary over a threefold range even for very small deviations in the
parameter estimates, depending more on assumptions rather than what can be estimated from
the data. The BMCL estimates shown in Table B-2 range from 81 to 85 ppm when rho > 2 but
could be as small as 26 to 29 ppm due to the uncertainty about the power function exponent in
dose-response model (here denoted y),
(Mean response in dose group j) = intercept + slope (concentration in group j) y
The dose-response function for these data can therefore be modeled, but the estimation
of the variance function leads to a high degree of uncertainty in the BMCL estimate. For
example, P(H3) = 0.00147 for the bottom 4 model runs in Table B-2 and P(H3) < 0.0001 for
all other models evaluated with the variance function exponent not specified. No model
available in the BMDS described the sharp decrease in variance with increasing n-hexane
concentration (decreasing mean MCV) unless rho > 2.5, which is not commonly encountered.
Table B-2. Parameters and modeling results for the n-hexane exposure MCV data
from Ono et al. (1982)
Parameters of power model
Dose response power
exponent
Estimate
0.716
0.708
0.704
0.702
0.75
Standard
error
0.268
0.252
0.246
0.242
fixed
Variance exponent
Estimate
2
2.5
2.75
2.9
2
Standard
error
fixed
fixed
fixed
fixed
fixed
Results
DOF
for
AIC
4
4
4
4
3
AIC
98.95
97.54
96.85
96.44
96.97
DOF
for
P(H4)
1
1
1
1
2
P(H4)
0.0141
0.0318
0.0477
0.061
0.0488
BMC
108
107
106
106
116
BMCL
26
28
29
29
81
11/17/05
B-4
DRAFT-DO NOT CITE OR QUOTE
-------
0.75
0.75
0.75
fixed
fixed
fixed
2.5
3
3.5
fixed
fixed
fixed
3
o
J
3
95.57
94.21
92.89
2
2
2
0.0928
0.194
0.3745
117
118
120
82
84
85
DOF = Degree of freedom.
By holding one or two parameters constant, the apparent variability in the parameter
estimates is constrained, and the BMCL is larger than in the unconstrained case. Specifying the
two parameters decreases the uncertainty associated with the BMCL so that the BMCL values
are larger than they might be if the parameter estimates were not specified or constrained (Table
B-2). In summary, the data limit the confidence that can be placed in any BMCL estimated
from Ono et al. (1982) data by BMDS.
The overall means and standard deviations for reduced mean total and distal mixed nerve
conduction velocity following inhalation exposure to n-hexane from the Ono et al. (1982) study
are presented below in Tables B-3 and B-4.
Table B-3. Mean total mixed nerve conduction velocity in rats exposed to n-hexane
Exposure group
0
200
500
Number of rats
8
8
8
Mean MCV
59.3
53.3
50.9
SDofMCV
4.1
3.0
1.3
Source: Ono etal. (1982)
Table B-4. Mean total distal nerve conduction velocity in rats exposed to
n-hexane
Exposure group
0
200
500
Number of rats
8
8
8
Mean MCV
52.3
47.6
45.3
SDofMCV
3.3
1.7
1.6
Source: Ono etal., 1982.
Continuous data models (linear, polynomial, power, or Hill where possible), either with
a constant variance or with variance as a power function of the mean value (using an additional
model parameter), were fit to total and distal mixed nerve conduction velocity endpoints using
U.S. EPA BMDS (version 1.3.2) and the same strategy as in the preceding analyses. These
endpoints had the same issues as in the MCV analyses. The standard deviation and variance
were much larger in the control animals than in those exposed to n-hexane and sharply
decreased with increasing dose. The variance decreased much more rapidly with increasing
B-5
-------
dose than did the mean dose-response function which required an unusually large value of the
exponent rho in the variance function model.
The power function model provided the best model as indicated by the AIC criterion,
but an adequate fit to either the dose-response model or the variance model was not found
among the models that were evaluated unless the values of both the dose-response function
power and the variance function exponent rho were specified as if they were known rather than
being estimated from the data. The BMDL estimate usually was either not computable or was
computable but nearly equal to zero (because the power was much less than one and the fitted
power function had a very steep slope near the control exposure) unless both parameters were
specified. When the constant variance option (rho = 0) was selected, either of the BMDS
quadratic (polynomial) or power function models perfectly fit the mean mixed nerve conduction
velocity data and fit nearly as well when the specification rho = 2 was made. The differences in
fit occurred because values of rho not equal to zero in the variance function model implied
different weights for fitting the mean responses at the three dose groups, but visual inspection
of the results and small chi-squared residuals showed that the deviations were small and of little
concern for BMDL estimation. The power function model was selected over the quadratic
model because the power function model produced a decreasing estimated dose-response at all
dose levels within the observed range of the data. Some quadratic models showed a small
increase in mixed nerve conduction velocity at high n-hexane concentrations and were not
selected as this result seemed biologically implausible.
All data sets had a large value of rho and a small value of the exponent n, usually
around 0.4. The a priori plausible estimates of the variance function exponent are specified as 0
(additive error) or rho = 2, which most closely matched the data. A positive BMDL estimated
concentration could be obtained unless the value of the power in the power function dose-
response model was specified along with specifying rho. The results for quadratic models fit
the data about as well as the power models without requiring specification of any of the
parameters of the dose-response function. The advantage of the quadratic models is that they
do not have an infinitely steep slope near the control concentration of 0 ppm n-hexane whereas
the power function models with exponent <1 do have an infinitely steep slope near the control
concentration of 0 ppm n-hexane. However, the power model has greater plausibility than the
quadratic model because estimates of MCV decrease with increasing concentration.
The models evaluated included the polynomial model of degree 1 (linear), the
polynomial model of degree 2 (quadratic), and the power function model. Neither the linear nor
the quadratic model provided adequate goodness of fit [all P(H4) < 0.0001] for any model
specification. None of the power function models provided an adequate fit to the data unless
B-6
-------
one or two parameters of the model were specified. The best estimates of the shape parameter
or exponent in the power function model were in the range of 0.36 to 0.49, and only provided
BMCL estimates nontrivially greater than zero when the value of the power function exponent
was specified and the power y in the variance function model was specified as being in the
range of roughly 2 to 4. The equation of the power model is the same as presented above.
Only results for specified rho of 0 and 2 are shown. Tables B-5 and B-6, below, show
the results for total and distal mixed nerve conduction velocity data sets. The BMCL estimates
are considered low (i.e., 13 to 33 ppm) due to the uncertainty about the power function
exponent in the dose response model. The sensitivity of the BMCL to the exponent was
explored by holding y constant at middle values within the confidence interval for that
parameter and setting y as a constant over the range 2 to 4.
An adequate fit of the BMDS power function and quadratic models can be obtained over
a wide range of parameter values. However, the BMCL estimates among acceptable models
vary over a threefold range even for modest deviations in the parameter estimates, depending
more on what one is willing to assume is known rather than what can be estimated from the
data. The BMCL estimates shown above depend on the mixed MCV index and on the model.
Highlighted in bold type is a set of consistent models that fit the data well as indicated by low
AIC and small chi-squared residuals at concentrations around the estimated BMCL even when
it is not possible to formally test the goodness of fit using the likelihood ratio test for H4, as
often happens when one or more parameters are specified. The BMCL for the recommended
models are shown in Table B-7.
Table B-5. Estimates of BMCL and goodness of fit statistics for total mixed nerve
conduction velocity from the Ono et al. (1982) study
Exponent in power
model estimate
Estimate
0.491
0.367
0.377
0.4
0.4
0.5
0.5
0.382
Standard
error
0.231
0.232
0.232
Specified
Specified
Specified
Specified
0.205
Exponent in variance
model estimate
Estimate
14.44
0
2
0
2
0
2
o
J
Standard
error
7.61
Specified
Specified
Specified
Specified
Specified
Specified
Specified
DOF
for
AIC
5
4
4
3
3
3
3
4
AIC
77.31
81.97
80.08
79.99
78.09
80.27
78.39
79.25
DOF
for
P(H4)
0
0
0
1
1
1
1
0
P(H4)
NE
NE
NE
0.888
<10'4
0.581
<10'4
NE
BMC
(ppm)
145
25.9
30
33
35
56
60
33
BMCL (ppm)
NE
NE
0+
15
16.8
30
33.2
0+
B-7
-------
0.387
0.4
0.5
0.199
Specified
Specified
4
13
14.58
Specified
6.47
7.67
4
4
4
78.5
75.5
75.31
0
1
1
NE
<10'4
<10'4
36
95
150
0+
NE
NE
Quadratic model
Exponent in variance model estimate
Estimate
3.46
0
1.5
1.875
Standard error
4.84
Specified
Specified
Specified
5
4
4
4
80.89
79.97
78.52
78.19
0
1
1
1
NE
<10'4
<10'4
<10'4
87
80
83
83
57.7
51.3
53.8
54.5
DOF = Degree of freedom.
NE = Could not be estimated.
0+ = any very small positive number <0.05 for BMCL in this table.
Table B-6. Estimates of power model BMCL estimates and goodness of fit
statistics for distal mixed nerve conduction velocity from the Ono et al. (1982)
study
Exponent in power
model estimate
Estimate
0.408
0.435
0.426
0.4
0.4
0.5
0.5
0.421
0.418
0.4
0.5
Standard
error
0.171
0.225
0.206
Specified
Specified
Specified
Specified
0.199
0.192
Specified
Specified
Exponent in variance
model estimate
Estimate
10.5
0
2
0
2
0
2
3
4
10.5
10.59
Standard
error
Specified
Specified
Specified
Specified
Specified
Specified
Specified
Specified
Specified
4.82
5.08
DOF
for
AIC
5
4
4
3
o
J
3
o
6
4
4
4
4
AIC
66.24
69.48
67.7
67.5
65.71
67.56
65.82
66.93
66.25
64.24
64.49
DOF
for
P(H4)
0
0
0
1
1
1
1
0
0
1
1
P(H4)
NE
NE
NE
0.875
<10'4
0.776
<10'4
NE
NE
<10'4
<10'4
BMC
(ppm)
61
34
35
28
29
48
51
35
36
59
91
BMCL
(ppm)
NE
0+
0+
13
14.3
26.4
28.3
0.11
NE
NE
NE
Quadratic model
Exponent in variance model estimate
Estimate
4.996
Standard error
0.235
DOF
for
AIC
5
AIC
112.3
DOF
for
P(H4)
0
P(H4)
NE
BMC
(ppm)
11850
BMCL
(ppm)
102.3
-------
0
1.625
Specified
Specified
4
4
67.48
0
1
NE
<10'4
80
81
51.3
53
DOF = Degree of freedom.
NE = Could not be estimated.
0+ = any very small positive number <0.05 for BMCL in this table.
The lack of fit of the BMDS variance function to the observed variances without use of
an untypically large value of rho leads to some questions about the extent to which these results
can be generalized. The models selected above all have rho = 2 (constant coefficient of
variation at all exposure levels) or as close as possible for the quadratic models. The power
function model with power = 0.4 is probably the most typical case but will yield BMCL
estimates lower than other models in Table B-4.
Table B-7. BMCL estimates for power and quadratic models with various
parameters from the Ono et al. (1982) study
Model and parameters selected for
BMCL estimation
Power = 0.4, rho = 2
Power = 0.5, rho = 2
Quadratic, rho near 2
Total MCV
17
33
55a
Distal MCV
14
28
53a
a Approximate largest value of rho for which a positive BMCL was found by halving intervals between
rho = 0 and 2 among the quadratic models with specified rho.
Modeling of reduced fetal body weight data (Mast, 1987)
All four BMDS continuous endpoint models were initially considered for the Mast
(1987) fetal body weight data. Specifically, continuous data models (linear, polynomial, power,
or Hill where possible), either with a constant variance or with variance as a power function of
the mean value (using an additional model parameter), were fit to the data using U.S. EPA
BMDS (version 1.3.2). The continuous data from the Mast (1987) study for reduced mean fetal
weight averaged across litters within each exposure group are shown in Table B-8.
B-9
-------
Table B-8. Mean of litter mean pup body weight in rats exposed to n-hexane
Exposure group (ppm)
0
200
1000
5000
Number of litters
23
24
27
28
Mean of litter mean
fetal body weight (g)
3.48
3.54
3.27
2.97
Standard deviation
0.37
0.36
0.32
0.38
Source: Mast, 1987.
The current version of BMDS does not provide models for continuous endpoints among
the BMDS nested modules for developmental toxicology data in which individual fetal
responses provided quantal response data as in Mast (1987). However, valid BMDS analyses
may be carried out using each litter as the unit of analysis in the BMDS continuous model
because the litters themselves correspond to different dams and are statistically independent,
unlike the individual fetal weight data that exhibit intralitter correlation.
The linear, quadratic, and power dose response models among the continuous-variable
models provided satisfactory goodness of fit ^-values [P(H4) > 0.10] using a constant variance
model (rho = 0) or a variance model with rho estimated from these data (BMDS version 1.3.2)
and provided an adequate fit to the data (with a goodness of flip-value >0.1). The model with
the lowest AIC was used to establish the BMC (see BMDS output). The continuous Hill model
with an unspecified exponent could not be fitted to the data due to lack of degrees of freedom
(DOFs) since the number of dose groups in the modeled data set were equal to or less than the
number of parameters estimated in the Hill dose response model. The results are shown in
Table B-9. Even the Hill model gave a satisfactory fit to the data by specifying an exponent of
1 or 2, either of which is a plausible value. The lowest AIC occurred when the data were fit by
a quadratic (polynomial of degree 2) model with constant variance, but the linear model with
constant variance had an AIC value close to that of the quadratic and also provided an adequate
fit to the data. The quadratic model with constant variance was therefore selected and provided
a BMCL estimate of 848 ppm n-hexane.
Visual inspection of a plot of the predicted and observed means also indicated a
reasonable fit of these models to the data overall, especially in the range nearest the estimated
BMCL to be used as a point of departure (see BMDS outputs). The model-predicted BMC
value associated with an approximate decrease in mean pup body weight of 10.6 % was 1540
ppm. The corresponding BMCL value was 848 ppm.
The estimated BMCL of 848 ppm for a quadratic model with constant variance (Table
B-10
-------
B-9) is reasonably robust to model specification or selection among models that fit the data
well: quadratic with nonconstant variance, BMCL = 838 ppm; Hill with exponent = 1 and
constant variance, 835 ppm; and Hill with exponent 2 and constant variance, 873 ppm. The
power function models did not fit the mean fetal body weight reduction data as well but
produced modestly larger (23% and 26%, respectively) BMCL values with either constant or
nonconstant variance. The linear model with constant variance has nearly as low an AIC value
as the quadratic, does not fit as well as the quadratic, power, or Hill models, and has an
estimated BMCL about three times larger.
Table B-9. Results from fitting BMDS models to mean fetal weight data from Mast
et al. (1987)
Dose-response model
Linear
Quadratic
Power
Hill exponent = 1
Hill exponent = 2
Variance
model
Power
Constant
Power
Constant
Power
Constant
Power
Constant
Power
Constant
DOF
for
AIC
4
3
5
4
5
4
5
4
5
4
AIC
-100.12
-102.02
-102.4
-104.37
-99.28
-101.22
-100.17
-102.13
-101.17
-103.13
P(H4)
DOF
2
2
1
1
1
1
1
1
1
1
P(H4)
0.1635
0.1603
0.2466
0.2517
0.1164
0.1166
0.2097
0.2130
0.4476
0.4326
BMC
(ppm)
3361
3435
1535
1540
2467
2506
1839
1858
1451
1464
BMCL (ppm)
2507
2655
838
848
1094
1120
824
835
862
873
Modeling of reduced MCVfollowing 8-, 12-, and 16-week n-hexane inhalation exposure data
The overall observed means and standard deviations for reduced MCV following 8-, 12-,
and 16-week exposure to n-hexane from the Huang et al. (1989) study are presented in Table B-
10. The data for decreased MCV were presented graphically in the Huang et al. (1989) study as
mean MCV ± SEM . The study authors were contacted to obtain the raw data for decreased
MCV. Dr. Huang was unable to provide these data due to length of time since the study was
completed and his relocation to several other institutions. Thus, these values were taken directly
from the graph in this figure by physical measurement AND converted to the scale of the
measurements, and the SEM was converted to SD. BMDS version 1.3.2 does not have a time
series analysis option, thus the Huang et al. (1989) data were analyzed separately for each of the
successive four-week observations, rather than as a time series in which each individual
measurement (4, 8, 12, and 16 weeks) may be correlated with those at another week.
B-ll
-------
Table B-10. Mean reduced MCV in rats exposed to n-hexane for 8,12, and
16 weeks
Exposure (ppm)
Number of animals
Observed mean MCV
Observed SD
8 weeks exposure
0
500
1200
3000
8
8
8
8
11.5
8.96
7.83
7.41
0.8
0.8
0.8
1.07
12 weeks exposure
0
500
1200
3000
8
8
8
8
11.8
8
7.45
4.06
0.8
0.667
1.07
0.8
16 weeks exposure
0
500
1200
3000
8
8
8
8
11.7
9.91
6.42
2.64
1.07
0.667
1.07
1.07
Source: Huang etal., 1989.
The Hill model provided the best fit to the data for 8-, 12-, and 16-week exposure to
n-hexane, but the parameters for the 12-week exposure duration were substantially different
from those for the 8-week and 16-week exposure duration data (Table B-l 1). The best fitting
model for 8- and 16-week exposure was the constant-variance Hill dose-response function with
the exponent specified as equal to 2 (Table B-l 1). The Hill model for 12 weeks exposure had
P(H4) = 0.3132 and an AIC = 27.35, the second best fitting modeling results. However, when
the Hill exponent was estimated from the 12-week exposure data and a constant-variance model
was assumed, the AIC was only slightly larger at 27.36 and the value of P(H4) = 0.5655 was
much larger. The goodness of fit statistics P(H4) for the models with the Hill exponent = 2 were
< 0.002 and considered inadequate. The AIC values for these models were about 10 units higher
than for the smaller values of the shape parameter (specified as 1 or estimated as 0.6), so the
value of 2 was rejected. We cannot offer any explanation for this difference, but note that the
best empirical estimate for the shape parameter was about 0.6, even further from the value of 2
used to model the other two exposure durations. The variance function P(H3) = 0.3946 fit the
variances well and the chi-squared residuals were much smaller for exponents fixed at 1 or
estimated at 0.6, than when the shape parameter was set equal to 2.
For the 12-week exposure data, a BMCL of 121.6 ppm n-hexane was estimated from a
constant-variance model in which exponent or shape parameter in the Hill dose-response
function was specified as 1, but estimating the shape parameter from the data as about 0.604
yielded a BMCL of 34.7 ppm, greater than threefold lower. The numerical stability of the
B-12
-------
estimates from the model with the shape parameter estimated from the data is doubtful, as all
four of the Hill model parameters were estimated from only four dose groups and the standard
error of the shape parameter could not be estimated. The BMCL = 121.6 ppm appears to be a
more reliable estimate.
Table B-ll. Results of fitting the Hill function dose-response model to rat
MCV data for 8,12, and 16-week exposures to n-hexane (Huang et al., 1989)
Hill model exponent
Fixed
2
2
1
1
NA
NA
Estimated
NA
NA
NA
NA
0.607
0.604
Variance model
exponent
Fixed
NA
0
NA
0
NA
0
Estimated
0.525
NA
-0.127
NA
-0.134
NA
DOF
for
AIC
5
4
5
4
6
5
AIC
39.97
38.3
29.32
27.35
29.34
27.36
DOF
for
P(H4)
1
2
1
2
0
1
P(H4)
0
0.0013
0.1301
0.3132
NE
0.5655
BMC
(ppm)
448
396
150.5
156.3
60.5
69.2
BMCL
(ppm)
217
310
111.4
121.6
NE
34.7
BMCLHEC
(mg/m3)
382.4
546.3
196.3
214.3
NE
61.2
NA = Not applicable.
NE = Cannot be estimated.
The BMCL (1SD) of 121.6 ppm (430 mg/m3) for decreased MCV in rats exposed to
n-hexane for 12 weeks was chosen as the point of departure based on the sensitivity of this
neurological effect following n-hexane exposure and the confidence in the modeling results from
this study at low doses. For the Huang et al. (1989) data set, the excess risk is equivalent to an
approximately 6.8% change in response. A summary of the BMD modeling results of the Huang
et al. (1989), Ono et al. (1982), and Mast (1987) data is presented in Table B-12.
Table B-12. Benchmark dose modeling results of n-hexane inhalation toxicity
studies for selection of the principal study
Referenc
e
Mast,
1987
Mast,
1987
Endpoint
Reduced
ossification of
sternebrae 1-4
Decreased fetal
body weight
gam
Dose
groups
4
4
Model
Nested
logistic
Quadratic
Fixed
parameters
n>l
n>l
Goodness of
fit/7-value
0.0001
(group:
0.1834)c
0.2517
AIC
1433
-104.37
BMCa
(ppm)
1571
1540
BMCLa
(ppm)
943
848
BMCLHEC
(mg/m3)
2770
1494
B-13
-------
Huang et
al., 1989
Huang et
al., 1989
Huang et
al., 1989
Ono et
al., 1982
Ono et
al., 1982
Ono et
al., 1982
MCV 8 weeks
MCV 12
weeks
MCV 16
weeks
MCV
Mixed MCV
(total)
Mixed MCV
(distal)
4
4
4
3
3
3
Hill
Hill
Hill
Power
Power
Power
rho = 0
n=2
rho = 0
n = 1
rho = 0
n=2
rho = 3.5
n = 0.75
rho = 2
n = 0.5
rho = 2
n = 0.5
0.789
0.313
0.779
0.3745
< 0.0001
< 0.0001
27.22
27.35
34.88
92.84
78.39
65.82
198
156
367
120
60
51
143
122
321
85.4
33
28
252
215
566
150.5
58.1
49.3
a BMR = 10% for quantitative endpoints and BMR = 1 SD for continuous endpoints.
b BMCLHEC presented in mg/m3 and conversion described in Section 5.2.2.1.
0 All models have been adjusted for either total litter size (late resorptions) or number of viable fetuses for
other endpoints. The nested module estimates p-value for goodness of fit by applying a chi-squared test to
data grouped by the strata or levels of these covariates as well as by dose. The chi-squared goodness of fit
test may give different results than the standard methods used elsewhere in this table.
B-14
-------
Output B-l: Nested, logistic model results for reduced ossification of sternebrae 1-4 in rats
data from Mast (1987)
Nested Logistic Model with 0.95 Confidence Level
T3
£>
1
g
03
LL
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
— MootQ/Hl^^ioti^"* ~~
i Nested Log i sue :
- Rl\/ir^ I /"MA/OP R/^i if"i/H -
: DIVIU LUWCI DOUIIU :
r ^^-^^^ __^-^
[ /^^^^^^^^^ I
r ^^ ^^^~~~^~~^ \
f / ^^^ \
L ^^^ J
^ ^
? ^
F. BMDL BMD .1
1000
2000 3000
dose
4000
5000
19:01 07/292004
Parameter constraints: Exponent n > 1.
Covariate: Number of implanted sites in the litter, including those resorbed.
Benchmark Response: BMR = 0.10 (nested quanta! endpoint).
BMDL(0.10, 95% confidence) = 943 ppm n-hexane.
The probability function is:
Prob. = alpha + thetal*Rij + [1 - alpha - thetaPRij]/ [l+exp(-beta-theta2*Rij-rho*log(Dose))],
where Rij is the litter specific covariate.
Restrict Power rho >= 1.
Total number of observations = 102
Total number of records with missing values = 0
Total number of parameters in model = 9
Total number of specified parameters = 0
Maximum number of iterations = 1250
Relative Function Convergence has been set to: le-008
B-15
-------
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 0.174419
beta= -9.52447
thetal = 0
theta2 = 0
rho= 1
phil = 0.29301
phi2= 0.130921
phi3 = 0.279439
phi4= 0.178497
Parameter Estimates
variable
alpha
beta
theta
theta2
rho
phil
phi 2
phi 3
phi 4
estimate
0.416634
-10.3618
-0.0161654
0.0543009
1
0.274038
0.122715
0.272467
0.178054
standard error
0.140121
2.39704
0.0172964
0.151612
bounded
0.13398
0.0625457
0.0889558
0.0654785
Analysis of Deviance
Model Log(likelihood) Deviance Test DF
Full model -582.697
Fitted model -708.705 252.015
Reduced model -832.35 499.306
P-value
94 2.2259458e-016
101 O.0001
AIC:
1433.41
Litter Data
Lit.-Spec. Litter
Dose Cov. Est. Prob. Size
chi-squared
Expected Observed Residual
0.
0.
0.
0.
0.
0.
0.
.0000
.0000
.0000
.0000
.0000
.0000
.0000
9.0000
10
11
12
12
12
13
.0000
.0000
.0000
.0000
.0000
.0000
0.271
0
0
0
0
0
0
.255
.239
.223
.223
.223
.206
9
10
11
12
12
12
13
2.440
2
2
2
2
2
2
.550
.627
.672
.672
.672
.684
9
3
1
0
0
2
1
2.7529
0.1754
-0.5949
-0.9253
-0.9253
-0.2327
-0.5573
B-16
-------
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
13.0000
14.0000
15.0000
15.0000
16.0000
16.0000
16.0000
16.0000
16.0000
17.0000
17.0000
17.0000
18.0000
18.0000
19.0000
19.0000
9.0000
10.0000
12.0000
13.0000
13.0000
13.0000
13.0000
13.0000
13.0000
14.0000
15.0000
15.0000
15.0000
15.0000
15.0000
16.0000
16.0000
17.0000
17.0000
17.0000
17.0000
17.0000
18.0000
18.0000
2.0000
9.0000
9.0000
13.0000
13.0000
13.0000
14.0000
14.0000
15.0000
15.0000
15.0000
15.0000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.206
.190
.174
.174
.158
.158
.158
.158
.158
.142
.142
.142
.126
.126
.109
.109
0.279
0.263
0.232
0.217
0.217
0.217
0.217
0.217
0.217
0.201
0.186
0.186
0.186
0.186
0.186
0.170
0.170
0.155
0.155
0.155
0.155
0.155
0.140
0.140
0.405
0.307
0.307
0.254
0.254
0.254
0.242
0.242
0.229
0.229
0.229
0.229
13
14
15
15
16
16
16
16
16
17
17
17
18
18
19
19
9
10
12
13
13
13
13
13
13
14
15
15
15
15
15
16
16
17
17
17
17
17
18
18
2
9
9
13
13
13
14
14
15
15
15
15
2.684
2.664
2.612
2.612
2.528
2.528
2.528
2.528
2.528
2.411
2.411
2.411
2.262
2.262
2.080
2.080
2.507
2.630
2.784
2.815
2.815
2.815
2.815
2.815
2.815
2.816
2.787
2.787
2.787
2.787
2.787
2.728
2.728
2.640
2.640
2.640
2.640
2.640
2.522
2.522
0.811
2.762
2.762
3.305
3.305
3.305
3.382
3.382
3.438
3.438
3.438
3.438
0
0
0
0
0
2
5
0
o
J
6
2
1
0
2
1
o
5
o
5
2
3
0
3
0
1
3
8
8
2
2
2
0
2
4
3
0
0
5
2
0
6
2
2
3
0
7
9
1
2
9
12
1
13
2
-0.8881
-0.8493
-0.8087
-0.8087
-0.7664
-0.1600
0.7496
-0.7664
0.1432
1.0753
-0.1231
-0.4227
-0.6761
-0.0783
-0.3259
0.2774
0.2603
-0.3119
0.0965
-1.2054
0.0793
-1.2054
-0.7772
0.0793
2.2205
2.1457
-0.3168
-0.3168
-0.3168
-1.1221
-0.3168
0.5018
0.1073
-1.0268
-1.0268
0.9183
-0.2488
-1.0268
1.3444
-0.2017
1.5187
0.0965
-1.1194
1.1389
1.7554
-0.7106
-0.4050
1.6457
2.3972
-0.6825
2.6771
-0.4025
B-17
-------
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
15.0000
15.0000
16.0000
16.0000
16.0000
16.0000
16.0000
16.0000
16.0000
17.0000
17.0000
17.0000
18.0000
18.0000
18.0000
6.0000
8.0000
10.0000
12.0000
14.0000
14.0000
14.0000
14.0000
14.0000
14.0000
15.0000
15.0000
15.0000
15.0000
15.0000
15.0000
15.0000
16.0000
16.0000
16.0000
16.0000
16.0000
16.0000
17.0000
17.0000
18.0000
18.0000
19.0000
0.229
0.229
0.217
0.217
0.217
0.217
0.217
0.217
0.217
0.205
0.205
0.205
0.193
0.193
0.193
0.442
0.427
0.414
0.404
0.395
0.395
0.395
0.395
0.395
0.395
0.391
0.391
0.391
0.391
0.391
0.391
0.391
0.388
0.388
0.388
0.388
0.388
0.388
0.386
0.386
0.384
0.384
0.383
15
15
16
16
16
16
16
16
16
17
17
17
18
18
18
6
8
10
12
14
14
14
14
14
14
15
15
15
15
15
15
15
16
16
16
16
16
16
17
17
18
18
19
3.438
3.438
3.472
3.472
3.472
3.472
3.472
3.472
3.472
3.486
3.486
3.486
3.482
3.482
3.482
2.651
3.417
4.143
4.843
5.529
5.529
5.529
5.529
5.529
5.529
5.871
5.871
5.871
5.871
5.871
5.871
5.871
6.216
6.216
6.216
6.216
6.216
6.216
6.564
6.564
6.918
6.918
7.279
2
1
1
4
11
3
0
0
6
6
0
0
5
6
2
o
J
1
4
11
7
8
6
0
8
o
3
4
0
2
4
2
8
1
9
4
10
11
8
12
4
9
9
3
5
-0.4025
-0.6825
-0.6648
0.1419
2.0243
-0.1270
-0.9337
-0.9337
0.6797
0.6522
-0.9046
-0.9046
0.3818
0.6332
-0.3726
0.2086
-1.1527
-0.0571
2.1063
0.4418
0.7421
0.1415
-1.6603
0.7421
-0.7594
-0.5296
-1.6619
-1.0958
-0.5296
-1.0958
0.6026
-1.3789
0.7454
-0.5931
1.0132
1.2809
0.4777
1.5486
-0.6510
0.6186
0.5028
-0.9460
-0.5245
Combine litters with adjacent levels of the litter-specific covariate
within dose groups until the expected count exceeds 3.0, to help improve
the fit of the XA2 statistic to chi-squared.
Grouped Data
Mean chi-squared
Dose Lit.-Spec. Cov. Expected Observed Residual
9.5000
4.990
12 2.0018
B-18
-------
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
11.5000
12.0000
13.0000
14.5000
15.5000
16.0000
17.0000
18.0000
19.0000
5.299
5.344
5.369
5.277
5.140
10.111
7.233
4.524
4.161
1
2
1
0
0
10
9
2
4
-1.0809
-0.8188
-1.0221
-1.1718
-1.1134
-0.0168
0.3057
-0.5334
-0.0343
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
200.0000
9.5000
12.5000
13.0000
14.5000
15.0000
16.0000
17.0000
18.0000
5.137
5.598
14.074
5.602
11.147
5.456
13.198
5.044
5
3
15
10
6
7
7
8
-0.0496
-0.8028
0.1774
1.2692
-1.0362
0.4307
-1.0783
0.8080
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
1000.0000
5.5000
11.0000
13.0000
14.0000
15.0000
16.0000
17.0000
18.0000
3.572
6.067
6.610
6.765
20.627
24.305
10.459
10.445
5
7
10
11
31
25
6
13
0.5515
0.2289
0.7388
0.8773
1.1856
0.0706
-0.6680
0.3709
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
5000.0000
7.0000
10.0000
12.0000
14.0000
15.0000
16.0000
17.0000
18.0000
19.0000
6.068
4.143
4.843
33.172
41.098
37.293
13.128
13.836
7.279
4
4
11
32
21
54
13
12
5
-0.7711
-0.0571
2.1063
-0.1437
-2.1502
1.8260
-0.0229
-0.3134
-0.5245
Chi-square= 32.30 DF = 26 P-value = 0.1834
To calculate the BMD and BMDL, the litter specific covariate is fixed at the mean litter specific covariate of control
group: 14.826087
Benchmark Dose Computation Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD =
1571.05
BMDL= 943.119
B-19
-------
Output B-2: Continuous Hill model results for mean fetal body weight in rats from Mast
(1987)
Polynomial Model with 0.95 Confidence Level
o
Q.
to
CD
CO
CD
3.6
3.4
3.2
2.8
Polynomial
BMD Lower Bound
BMDL
BMD
0
1000 2000 3000
dose
4000
5000
15:2508/222005
Polynomial Model. Revision: 2.2 Date: 9/12/2002
Input Data File: U:\IRIS\N-hexanE\MAST\MASTRATFW082205.(d)
Gnuplot Plotting File: U:\IRIS\N-hexanE\MAST\MASTRATFW082205.plt
MonAug22 15:25:53 2005
Mast Rat Fetal Wt by Litter Quadratic model rho==0
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*doseA2 + ...
Dependent variable = MEAN
Independent variable = DOSE
rho is set to 0
B-20
-------
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
alpha = 1
rho = 1 Specified
beta_0 = 3
beta_l = -0.001
beta 2 = 0
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
alpha 0.124672 0.0174576 0.090456 0.158889
beta_0 3.53112 0.0584714 3.41652 3.64572
beta_l -0.000281312 0.000116308 -0.000509272 -5.33529e-005
beta 2 3.38002e-008 2.19346e-008 -9.19091e-009 7.67912e-008
Asymptotic Correlation Matrix of Parameter Estimates
alpha beta_0 beta_l beta_2
alpha 1 1.8e-007 -2.6e-007 2.6e-007
beta_0 1.8e-007 1 -0.7 0.64
beta_l -2.6e-007 -0.7 1 -0.99
beta 2 2.6e-007 0.64 -0.99 1
Table of Data and Estimated Values of Interest
Dose N ObsMean ObsStdDev EstMean EstStdDev ChiA2
Res.
0
200
1000
5000
23
24
27
28
3.48
3.54
3.27
2.97
0.37
0.36
0.32
0.38
3.53
3.48
3.28
2.97
0.353
0.353
0.353
0.353
-0.694
0.885
-0.2
0.00656
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
B-21
-------
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al 55.842420 5 -101.684839
A2 56.285063 8 -96.570125
fitted 55.185409 34 -104.370819-102.370819
R 37.662042 2 -71.324083
Test 1: Does response and/or variances differ among dose
levels
(A2vs. R)
Test 2: Are Variances Homogeneous (Al vs A2)
Test3: Does the Model for the Mean Fit (Al vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) Testdf p-value
Testl 37.246 6 <0001
Test 2 0.885286 3 0.829
Test3 1.31402 1 0.2517
The p-value for Test 1 is less than .05. There appears
to be a
difference between response and/or variances among the
dose levels.
It seems appropriate to model the data
The p-value for Test 2 is greater than .05. A
homogeneous variance
model appears to be appropriate here
The p-value for Test 3 is greater than .05. The model
chosen appears
to adequately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
B-22
-------
BMD= 1540.16
BMDL= 848.016
B-23
-------
Output B-3: Continuous, Hill model results for decreased MCV in rats following 8 weeks
exposure to n-hexane from Huang et al. (1989)
Hill Model with 0.95 Confidence Level
12
11
o>
co
o 10
Q.
CO
0>
03
CD
8
7
6
Hill
BMD Lower Bound
BMDL BMD
500
1000
1500
dose
2000
2500
3000
12:2207/082004
Parameter constraints: Rho = 0 (constant variance) and n = 2.
Benchmark Response: BMR = 1 standard deviation of control group.
BMDL(1 std. Dev., 95% confidence) = 143 ppm n-hexane.
The form of the response function is:
Y[dose] = intercept + v*doseAn/(kAn + doseAn)
Dependent variable = MEAN
Independent variable = HEXANE_CONC
rho is set to 0
n is set to 2
Power parameter is not restricted
The variance is to be modeled as Var(i) = alpha * mean(i) A rho
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
B-24
-------
User Inputs Initial Parameter Values
alpha = 0.1
rho =1 Specified
intercept =12
v = -3
n = 1 Specified
k=100
Asymptotic Correlation Matrix of Parameter Estimates
alpha
rho
intercept
V
n
k
alpha
1
0
0
0
0
0
rho
0
1
0
0
0
0
intercept
0
0
1
0
0
0
V
0
0
0
1
0
0
n
0
0
0
0
1
0
k
0
0
0
0
0
1
Parameter Estimates
Variable Estimate
Std. Err.
alpha
rho
intercept
V
n
k
0.67071
0
11.5075
-4.13841
2
398.844
1
1
1
1
1
1
Data and Estimated Values of Interest
dose
0
500
1200
3000
n
8
8
8
8
observed
mean
11.5
8.96
7.83
7.41
observed
standard
deviation
0.8
0.8
0.8
1.07
estimated
mean
11.5
8.98
7.78
7.44
estimated
standard
deviation
0.819
0.819
0.819
0.819
chi square
0.00239
-0.0197
0.0604
-0.0431
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = arpha*(Mu(i))Arho
B-25
-------
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -9.585726 5 29.171453
A2 -9.044284 8 34.088568
A3 -9.371958 6 30.743917
fitted -9.609309 4 27.218617
R -35.201386 2 74.402771
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) Test df p-value
Testl 52.3142 6 <0001
Test 2 1.08288 3 0.7812
Test3 0.655348 2 0.7206
Test 4 0.474701 2 0.7887
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .05. Consider running a
homogeneous model
The p-value for Test 3 is greater than .05. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .05. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD= 198.109
BMDL= 143.112
B-26
-------
Output B-4: Continuous, Hill model results for decreased MCV in rats following 12 weeks
exposure to n-hexane from Huang et al. (1989)
Hill Model with 0.95 Confidence Level
12
10
O
Q.
tr
05
o>
Hill
BMD Lower Bound
BMDL BMD
0
16:2408/092004
500
1000
1500
dose
2000
2500
3000
Parameter constraints: Rho = 0 (constant variance) and n = 2.
Benchmark Response: BMR = 1 standard deviation of control group.
BMDL(1 std. Dev., 95% confidence) = 122 ppm n-hexane.
The form of the response function is:
Y[dose] = intercept + v*doseAn/(kAn + doseAn)
Dependent variable = MEAN
Independent variable = HEXANE_CONC
rho is set to 0
n is set to 1
Power parameter restricted to be greater than 1
The variance is to be modeled as Var(i) = alpha * mean(i) A rho
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
B-27
-------
User Inputs Initial Parameter Values
alpha = 0.0001
rho = 1 Specified
intercept = 0
v = -l
n = 1 Specified
k = 1000
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -n have been estimated at a boundary point, or have been specified by the user, and do
not appear in the correlation matrix)
alpha
rho
intercept
V
k
alpha
1
0
0
0
0
rho
0
1
0
0
0
intercept
0
0
1
0
0
V
0
0
0
1
0
k
0
0
0
0
1
Parameter Estimates
Variable Estimate
alpha
rho
intercept
Std. Err.
v
k
0.673532
0
11.6883
-13.7905
2469.9
Data and Estimated Values of Interest
dose
0
500
1200
3000
n
8
8
8
8
observed
mean
11.8
9.06
7.45
4.06
observed
standard
deviation
0.8
0.667
1.07
0.8
estimated
mean
11.7
9.37
7.18
4.12
estimated
standard
deviation
0.821
0.821
0.821
0.821
chi square
0.127
-0.378
0.334
-0.083
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = arpha*(Mu(i))Arho
B-28
-------
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -8.528285 5 27.056569
A2 -7.585712 8 31.171423
A3 -8.515700 6 29.031399
fitted -9.676481 3 25.352961
R -50.666148 2 105.332296
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) Test df p-value
Testl 86.1609 6 <0001
Test 2 1.88515 3 0.5966
Test3 1.85998 2 0.3946
Test 4 2.32156 2 0.3132
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .05. Consider running a
homogeneous model
The p-value for Test 3 is greater than .05. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .05. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 156.287
BMDL= 121.613
B-29
-------
Output B-5: Continuous, Hill model results for decreased MCV in rats following 16 weeks
exposure to n-hexane from Huang et al. (1989)
Hill Model with 0.95 Confidence Level
12
10
co
I 8
co
0
tr
c r-
as b
0
Hill
BMD Lower Bound
BM.PL.BMP.
0
500
1000
1500
dose
2000
2500
3000
17:5807/062004
Benchmark Response: BMR = 1 standard deviation of control group.
BMDL(1 std. Dev., 95% confidence) = 321 ppm n-hexane.
The form of the response function is:
Y[dose] = intercept + v*doseAn/(kAn + doseAn)
Dependent variable = MEAN
Independent variable = HEXANE_CONC
rho is set to 0
n is set to 2
Power parameter is not restricted not restricted
The variance is to be modeled as Var(i) = alpha * mean(i) A rho
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
B-30
-------
alpha = 0.1
rho = 1 Specified
intercept = 20
v = -10
n = 1 Specified
k = 1000
Asymptotic Correlation Matrix of Parameter Estimates
Parameter Estimates
Variable Estimate
alpha 0.852015
rho 0
intercept 11.6245
v -10.3269
n 2
k 1171.69
Std. Err.
Data and Estimated Values of Interest
dose
0
500
1200
3000
n
8
8
8
8
observed
mean
11.7
9.91
6.42
2.64
observed
standard
deviation
1.07
0.667
1.07
1.07
estimated
mean
11.6
10
6.34
2.66
estimated
standard
deviation
0.923
0.923
0.923
0.923
chi square
0.0797
-0.139
0.0837
-0.0248
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = alpha*(Mu(i))Arho
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -13.304172 5 36.608345
A2 -12.188625 8 40.377250
A3 -13.188105 6 38.376209
fitted -13.437576 4 34.875151
R -57.381404 2 118.762807
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
B-31
-------
(A2vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) Test df p-value
Testl 90.3856 6 <.0001
Test 2 2.2311 3 0.5258
Test3 1.99896 2 0.3681
Test 4 0.498942 2 0.7792
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .05. Consider running a
homogeneous model
The p-value for Test 3 is greater than .05. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .05. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 367.09
BMDL= 321.332
B-32
-------
Output B-6: Power model results for decreased MCV in rats following exposure to n-hexane
from Ono et al. (1982)
Power Model with 0.95 Confidence Level
55
50
CO
o
Q.
CO
0
o: 45
£Z
05
0
40
35
Power
BMDL BMD
0 100
14:4608/242005
200 300
dose
400
500
Power Model. SRevision: 2.1 $ $Date: 2000/10/11 20:57:36 $
Input Data File: U:\IRIS\N-hexanE\ONO\ONO_RAT.(d)
Gnuplot Plotting File: U:\IRIS\N-hexanE\ONO\ONO_RAT.plt
WedAug24 14:46:472005
Ono Rat MCV Power Model rho==3.5 n==0.75
The form of the response function is:
Y[dose] = control + slope * doseApower
B-33
-------
Independent variable = DOSE
rho is set to 3.5
power is set to 0.75
The variance is to be modeled as Var(i) = alpha*mean(i)Arho
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
alpha = 0.001
rho = 1 Specified
control = 59
slope = -0.0001
power = 1 Specified
Asymptotic Correlation Matrix of Parameter Estimates
alpha rho control slope power
alpha 1 -1 -0.37 0.27 0.26
rho -1 1 0.36 -0.27 -0.25
control -0.37 0.36 1 -0.64 -0.59
slope 0.27 -0.27 -0.64 1 1
power 0.26 -0.25 -0.59 1 1
Parameter Estimates
Variable Estimate Std. Err.
alpha 2.71502e-005 0.000344776
rho 3.5 3.28236
B-34
-------
control 50.2999 1.91213
slope -0.136897 0.23451
power 0.75 0.2655
Table of Data and Estimated Values of Interest
Dose N ObsMean ObsStdDev EstMean EstStdDev ChiA2Res.
0 8 49.6 7.6 50.3 4.95 -0.141
200 8 42.9 3.5 43 3.77 -0.0317
500 8 36.2 1.1 35.8 2.73 0.137
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = alpha*(Mu(i))Arho
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al
A2
A3
fitted
R
-48.403560
-37.407394
-42.464475
-43.446606
-59.621629
4
6
5
3
2
104.807120
86.814789
94.928950
92.893212
123.243258
B-35
-------
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) d.f p-value
Testl 44.4285 4 <00001
Test 2 21.9923 2 1.677e-005
Test3 10.1142 1 0.001471
Test 4 1.96426 2 0.3745
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .05. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is less than .05. You may want to consider a
different variance model
The p-value for Test 4 is greater than .05. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD= 119.577
BMDL= 85.3538
B-36
-------
Output B-7: Power model results for decreased total mixed nerve conduction velocity in
rats following exposure to n-hexane from Ono et al. (1982)
Power Model with 0.95 Confidence Level
CD
c
CO
CD
64
62
60
58
56
r-A
54
52
50
Power
BMDL BMP
0
14:4208/242005
100
200 300
dose
400
500
Power Model. SRevision: 2.1 $ $Date: 2000/10/11 20:57:36 $
Input Data File: U:\IRIS\N-hexanE\ONO\ONO_RAT.(d)
Gnuplot Plotting File: U:\IRIS\N-hexanE\ONO\ONO_RAT.plt
WedAug24 14:41:582005
Ono Rat (Total) MNCV Power Model rho==2 n==0.5
The form of the response function is:
Y[dose] = control + slope * doseApower
Dependent variable = MEAN
B-37
-------
rho is set to 2
power is set to 0.5
The variance is to be modeled as Var(i) = arpha*mean(i)Arho
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
alpha = 0.001
rho = 1 Specified
control = 59
slope = -0.0001
power = 1 Specified
Asymptotic Correlation Matrix of Parameter Estimates
alpha rho control slope power
alpha 1 -1 -0.17 0.1 0.085
rho -1 1 0.17 -0.098 -0.083
control -0.17 0.17 1 -0.49 -0.42
slope 0.1 -0.098 -0.49 1 1
power 0.085 -0.083 -0.42 1 1
Parameter Estimates
Variable Estimate Std. Err.
alpha 0.00253828 0.0536687
rho 2 5.24144
control 59.2077 1.10404
B-38
-------
slope -0.38656 0.621848
power 0.5 0.256451
Table of Data and Estimated Values of Interest
Dose N ObsMean ObsStdDev EstMean EstStdDev ChiA2Res.
0 8 59.3 4.1 59.2 2.98 0.0309
200 8 53.3 3 53.7 2.71 -0.163
500 8 50.9 1.3 50.6 2.55 0.132
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = arpha*(Mu(i))Arho
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Warning: Likelihood for fitted model larger than the Likelihood for model A3.
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -36.984508 4 81.969016
A2 -32.573332 6 77.146663
A3 -68.919854 5 147.839707
fitted -36.192522 3 78.385045
R -48.756302 2 101.512604
B-39
-------
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) d.f p-value
Testl
Test 2
Test3
Test 4
32.3659
8.82235
72.693
-65.4547
4
2
1
2
<00001
0.01214
<.00001
<.00001
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .05. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is less than .05. You may want to consider a
different variance model
The p-value for Test 4 is less than .05. You may want to try a different
model
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 59.5474
B-40
-------
BMDL= 33.1628
B-41
-------
Output B-8: Power model results for decreased total distal nerve conduction velocity in rats
following exposure to n-hexane from Ono et al. (1982)
Power Model with 0.95 Confidence Level
CO
c
o
o>
o:
56
54
52
50
48
46
44
Power
BiyiDL BMP
0
14:5408/242005
100
200 300
dose
400
500
Power Model. SRevision: 2.1 $ $Date: 2000/10/11 20:57:36 $
Input Data File: U:\IRIS\N-hexanE\ONO\ONO_RAT.(d)
Gnuplot Plotting File: U:\IRIS\N-hexanE\ONO\ONO_RAT.plt
WedAug24 14:54:162005
Ono Rat Distal MNCV Power Model rho==2 n==0.5
The form of the response function is:
Y[dose] = control + slope * doseApower
B-42
-------
Independent variable = DOSE
rho is set to 2
power is set to 0.5
The variance is to be modeled as Var(i) = alpha*mean(i)Arho
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
alpha = 0.001
rho = 1 Specified
control = 59
slope = -0.0001
power = 1 Specified
Asymptotic Correlation Matrix of Parameter Estimates
alpha rho control slope power
alpha 1 -1 -0.21 0.23 0.22
rho -1 1 0.21 -0.23 -0.22
control -0.21 0.21 1 -0.48 -0.4
slope 0.23 -0.23 -0.48 1 0.99
power 0.22 -0.22 -0.4 0.99 1
Parameter Estimates
Variable Estimate Std. Err.
alpha 0.00190554 0.0368726
rho 2 4.94577
B-43
-------
control 52.2904 0.838647
slope -0.319513 0.468668
power 0.5 0.234407
Table of Data and Estimated Values of Interest
Dose N ObsMean ObsStdDev EstMean EstStdDev ChiA2Res.
0 8 52.3 3.3 52.3 2.28 0.0042
200 8 47.6 1.7 47.8 2.09 -0.0824
500 8 45.3 1.6 45.1 1.97 0.0782
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = arpha*(Mu(i))Arho
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Warning: Likelihood for fitted model larger than the Likelihood for model A3.
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -30.737669 4 69.475338
A2 -27.954058 6 67.908116
A3 -44.665496 5 99.330992
fitted -29.909322 3 65.818644
B-44
-------
R -43.520950 2 91.041900
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) d.f p-value
Testl 31.1338 4 <.00001
Test 2 5.56722 2 0.06181
Test3 33.4229 1 <.00001
Test 4 -29.5123 2 <.00001
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .05. Consider running a
homogeneous model
The p-value for Test 3 is less than .05. You may want to consider a
different variance model
The p-value for Test 4 is less than .05. You may want to try a different
model
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD= 51.0371
B-45
-------
BMDL = 28.8322
B-46
------- |