Health Advisory
Office of Drinking Water
U.S. Environmental Protection Agency
Washington, DC 20460

This report was prepared la accordance vlch the Memorandum of
Understanding between cht Department of che Army, DapuCy for Environment
Safety and Occupational Health (OASA(X&L)), and tha U.S. Environaantal
Protection Agency (EPA), Office of Drinking Water (ODW), Criteria and
Standards Division, for tha purpose of developing drinking water Health
Advisories (HAs) for selected environmental contaminants, as requested by the
Health Advisories provide specific advice on the levels of contaminants
in drinking water at which adverse health effects would not be anticipated and
which include a margin of safety so as to protect the most sensitive members
of the population at risk. A Health Advisory provides health effects
guidelines, analytical methods and recommends treatment techniques on a
case-by-case basis. These advisories are normally prepared for One-day,
10-day, Longer-term and Lifetime exposure periods where available
toxlcologlcal data permit. These advisories do not condone the presence of
contaminants in drinking water; nor are they legally enforceable standards.
They are not issued as official regulations and they may or may not lead to
the issuance of national standards or Maximum Contaminant Levels (MCLs).
This report is the product of the foregoing process. Available
toxlcologlcal data, as provided by the Army, on the munitions chemical
trinitroglycerol (TNG) have been reviewed and relevant findings are presented
in this report in a manner so ss to allow for an evaluation of the data
without continued reference to the primary documents.
Significant physical and chemical properties of TNG as well as potential
sources of exposure are provided as a meana of identifying the subject
chemical. Additionally, tha pharmacokinetic properties in various species
have been extensively studied and reviewed, and significant points are
presented in some detail. All available toxlcologlcal data, including
short-term, longer-term and lifetime feeding studies in three species and
reproductive, mutagenic, teratogenic and carcinogenic effects, as available,
have been reviewed, and those effeces deemed related to the ingestion of TNG
are summarised. Results of immunologic studies and icthemoglobln assays are
also Included. Summary charts delineating relevant indices of toxicity as
well as tha Incidence of TNG related lesions are Included in Appendices 1-3.
Based on these relevant findings, along with the known pharmacological
effects of TNG when used medicinally as a coronary vasodilator, a Health
Advisory for TNG has been developed vlch specific recommended exposure levels
for One-day, 10-day, Longer-term and Lifetime exposure. These calculations
are made utilizing dose-response data on a mg/kg body weight basis and are
extrapolated with the application of appropriate uncertainlty factors based

upon the nature and extent of available data. Estimates of projected excess
lifetime cancer risks are provided.
While le la recognized chat Ingestion via drinking water may not be the
sole route of exposure to TNG should It become a contaminant of drinking
water, the absence of data on such nonlngestlon routes as cooking, showering
swimming and similar activities along with lack of methods for estimating such
exposure prevent the incorporation of these routes into the development of HA
values. However, the Incorporation of the referenced safety factors into
these values is felt to allow for such additional exposure as well as for
exposure in the most sensitive of the population.
This report also Includes a section describing state-of-the-art methods
of analysis for TNG in drinking water and Includes essential treatment
parameters appropriate for treatment of an affected water supply.
This report has been submitted to a critical review by the EPA to Include
a panel of Health Effects Branch (KEB) toxicologlsts. Their comments, as
appropriate, have been incorporated Into this report.
A companion document, "Data Deficiencies/Problem Areas and
Recommendations for Additional Data Base Development for Trlnitroglycerol" la
Included in this report under Appendix 4.
I would like to express my thanks to Dr. John Glennon, Life Systems,
Inc., who afforded valuable coordination and logistical assistance. I also
thank Dr. Janet Normandy, Ha. Lorl Gordon and Dr. William Hartley who provided
the extensive technical skills required for the preparation of this report.
I am grateful to tha following members of the EPA Tox-Reviev Panel, Or.
Kenneth Bailey, Dr. Aabika Bathija, Dr. Bill Marcus and Mr. Bruce Mints, who
took time to review this report and to provide their invaluable input.
Finally, I would like to thank Dr. Edward Ohanlan, Chief, Health Effects
Branch, Dr. Joseph Cotruvo, Director. Criteria and Standards Division, and Dr.
Penelope Fennar-Crisp, Manager, Htalth Advisory Program, for providing me with
the opportunity and encouragement to be a part of this project.
The preparation of this Adviaorv was funded in part by Interagency
Agreement (IAG) 85-PP5869 between the U.S. EPA and the U.S. Army Medical
Research and Development Command U'SA>TRDC). This IAG was conducted with the
technical support of the U.S. Army Biomedical Research and Development
Laboratory (USABRDL).
Kriahan Khanna, Project Officer
Office of Drinking Water

table of contents
A.	Absorption and Excreelon		iv-1
B.	Distribution		iv-5
C.	Metabolism		IV-5
A.	Health Effects in Humans		V-l
B.	Health Effects in Animal Experiments 		V-4
1.	Short-Term Exposure		v-4
a.	Primary Skin and Eye Irrltaelon Studies		V-7
b.	Dermal Sensitivity Study		v-7
c.	Four-Week Studies 		V-7
2.	Longer-Term Exposure		V-8
a.	Thirteen-Week Studies 		v-8
b.	Lifetime Exposure Studies 		v-10
3.	Genotoxlcology		v-13
4.	Carcinogenicity . 		v-i5
5.	Reproductive Effects		v-ib
6.	Teratogenicity		v-l 9
7.	Other Effects		v~21
a.	Methemoglobln Formation 		v-21
b.	Iraaunologic Effects		v-2 2
A. Quantification of Toxicologic a 1 Effects		-1-3

Table of Contents - continued
B.	One-Day Health Advisory 		VI-4
C.	Ten-Day Health Advisory 		VI-5
D.	Longer-Term Health Advisory 		VI-5
C. Life-Time Exposure		vi-6
F. Evaluation of Carclaoganic Potential		VI-6
1	Indices of Toxicity In Rats Fed TNG		Al-1
2	TNG-ralatad Lesions In tha Rat		 . . . .	A2-1
3	Cardiovascular Ralatad Effects of TNG in Rats		A3-1
4	Data Deficiencies/Problem Areas and Recoanendatlons for
Additional Data Baaa Development for Trlnltroglyeerol 		A4-1

II—1 General Chemical and Physical Properties of TNG 	 II-2
IV-1 Percentage of Orally Administered 1,3- C-TNG
Recovered la 24 Hours	 IV-4
IV-2 Percentage of Orally Administered 1.3- C-TNG Recovered From
Various Organs In 24 Hours	 IV-6
IV-3 Tissyj/Plasm* Radioactivity Ratio 24 Hours After Oral
1,3- C-TNG Administration	 IV-7
IV-4	Accumulation of Metabolites of 1,3- C-TNG in Liver Homogenates .	IV-10
V-l	Summary of Toxicity Studies on TNG in Animals 			V-S
V-2 Incidence of Mortality in Rats Fed TNG for 24 Months		V-12
V-3 Survival Data and Significant Pathology la Rats Nscropsiad
After 24 Months of TNG Feeding			 V-l7
V-4 Significant Pathology in Rats Dying During TNG Feeding Study . . V-18
V-5 Body Weight/Feed Intake During F Gestation 	 V-20
1 o
A1-1 Indices of Toxicity In Male Rats Fed TNG for Various Periods
of Tims	 Al-2
Al-2 Indices of Toxicity In FemAle Rats Fed TNG for Various
Periods of Tla«	 Al-4
A2-1 Incidence of Hepatocellular Carcinoma in Rats Fsd TNG for up
to 24 Months	 A2-2
A2-2 Incidence of Hypsrplaatic Foci of the Liver in Rats Fed
TNG for up to 24 Montha			 A2-3
A2-3 Incidence of Neoplastic Nod ales of che Livsr in Rats Fed
TNG for up to 24 Months	 A2-^
A2-4 Incidence of Interstitial Cell Tumors of the Testis in Rsts Fed
TNG for up to 24 Months			A2-5
A3-1 Incidence of Myocardial Degeneration/Fibrosis in Rats Fed
TNG for up to 24 Months	 A3-2

I. introduction
The Health Advisory (HA) Program, sponsored by the Office of Drinking
Water (ODW), provides information on the health effects, analytical
methodology and treatment technology that would be useful in dealing with the
contamination of drinking water. Health Advisories describe nonregulatory
concentrations of drinking water contaminants at which adverse health effects
would not be anticipated to occur over specific exposure durations. Health
Advisories contain a margin of safety to protect sensitive members of the
Health Advisories serve as Informal technical guidance to assist Federal,
State and local officials responsible for protecting public health when
emergency spills or contamination situations occur. They are not to be
construed as legally enforceable Federal standards. The Advisories are
subject to change as new information becomes available.
Health Advisories are developed for One-day. Tea-day, Longer-term
(approximately 7 years, or 10Z of an individual's lifetime) and Lifetime
exposures based on data describing noncarclnogenlc end points of toxicity.
Health Advisories do not quantitatively Incorporate any potential carcinogenic
risk from such exposure. For those substances that are known or probable
human carcinogens, according to the Agency classification scheme (Group A or
B), Lifetime HAs are not recommended. The chemical concentration values for
Group A or B carcinogens are correlated with carcinogenic risk estimates by
employing a cancer potency (unit risk) value together with assumptions for
lifetime exposure and the consumption of drinking water. The cancer unit risk
Is usually derived from the linear multistage model with 95X upper confidence
limits. This provides a low-dose estimate of cancer risk to humans that is
considered unlikely to pose a carcinogenic risk in excess of the stated
values. Excess cancer risk estimates may also be calculated using the
One-hit, Weibull, Loglt and Problt models. There is no current understanding
of the biological mechanisms Involved in cancer to suggest that any one of
these models Is able £0 predict risk more accurately than another. Because
each model Is based upon differing assumptions, the estimates that are derived
can differ by several orders of magnitude.

Glyceryl trinitrate, a nlcrace ester, is also called nitroglycerin and
trinitroglycerol. These names Imply that the compound contains "nltro" groups
(-NO ), generally associated with aromatic nitrated compounds, and are, therefore,
not in accord with standardized chemical nomenclature. However, since much of the
literature has used the acronym TNG {trinitroglycerol), this report will also
refer to the compound as TNG.
Trinitroglycerol, an explosive aliphatic nitrate ester which has been used
since 1864 as a commercial blasting explosive, is a pale yellow, viscous liquid
prepared by nitrating glycerol with either a mixture of concentrated nitric acid
and concentrated sulfuric acid or "a mixture of concentrated plus fuming nitric
acids" (Sullivan et al., 1979). The content of this mixture Is not clearly
General chemical and physical properties of TNG are presented in Table II-I.
Trinitroglycerol (molecular weight 227.09) does not polymerize and exists as a
stable solid at temperatures below 13.5* C in the form of dipyraaldal crystals.
Liquid TNG begins to become volatile and to decompose at 50-60* C. AC 145* C, the
rate of decomposition is so rapid that the liquid appears to boll. At 256° C, TNG
explodes spontaneously. When stored at temperatures below SO* C, TNG is stable
for many years (Mark et al.* 1965 as cited in Sullivan et al., 1979).
The appreciable solubility of TNG in water (1800 mg/L at 20* C; 2460 mg/L at
60° C) Indicates that environmentally significant concentrations may be dissolved
in waste rinse water and should be expected to remain in the water column.
II -1

Table II-1 General Chemical and Physical Properties of TN'G^
Molecular- Weight
Empirical Formula
Physical State
Specific Gravity
Melting Point
Boiling Point
Heat of Combustion
Autolgnitlon Temperature
Vapor Pressure
Solubility Characteristics:
Carbon Dlaulflde
Conversion Factor
1.2,3- Propanetriol trinitrate, glyceryl
trinitrate (IUPAC name), trinitro-
glycerol, nitroglycerin, trinitrin, TNG.
cn2 - oko2
CH - 0N0„
ch2 - ono2
Pale yellow viscous liquid
1.592 9 20*C
13.5"C / 2.8'C (labils form)
Apparently bolls by decomposing rapidly
at temperatures above 145®C. Explodes
at 256*C«
1580 cal/g
2.6 X 10"4
0.18Z w/v
2SZ v/v
5.6Z v/v
0.83Z v/v
Hg 9 20*C
1 ppo ¦ 9.29 mg/m
References: Merck and Co., 1976. Havley, 1977; Weast, 1971; Rosenblatt et
al.( 1973; Mark «c al., 1965 as cited in Sullivan et al.. 1979

Humans are exposed co TNG In clinical usage as a coronary vasodilator as
veil as in occupational settings and there nay also be exposure in limited
geographical areas to TNG as an environmental contaminant (Rosenblatt et al.,
1973). Within industrial settings, workers may be exposed by inhalation and
through dermal contact (Shiotsuka, 1979). Wastewater discharges from the
manufacture of commercial dynamite preparations or military explosives could
contaminate groundwater and public drinking water supplies.

Information on the metabolism of TNG was collected dur^gg single dose and
prolonged feeding experiments in various species using 1,3- C-TNG. In nose
species absorption was essentially complete in 24 hours with mice absorbing
somewhat less than the other species. From 4.3Z to 6.8Z of the administered dose
was retained in the liver of all species following administration of a single oral
dose of ISO - 200 mg/kg (approximately the LD for rats). Dogs and monkeys also
retained an average of 9.5X to 13.OX of the administered dose in skeletal muscle.
In rats, moat of the absorbed radioactivity was excreted in Che urine and expired
air, whereas mice excreted fairly equivalent amounts in feces, urine and expired
air. The major excretory route in rabbits* dogs and monkeys was the urine.
The urinary metabolites in most species consisted largely of free
mononitroglycerins (MNGs), glycerol and other polar metabolites including glucur-
onides, while TNG end free dlnltroglycsrins (DNGs) were excreted only in small
amounts. Mice excreted only smell amounts of free MNG end DNG- and MNG-
glucuronldes, indicating relatively complete biotransformation in this animal
species. The large amounts of unidentified polar compounds and glycerol found In
all species indicated a rapid denitration of TNG toward glycerol. The partially
nitrated compounds were glucuronidated and excreted la the urine. The glycerol
was excreted in the urine, oxidized to CO^, or converted into a veriety of other
metabolites. Results of biliary excretion studies suggest relatively complete
absorption after oral administration.
In vitro metabolism of TNG by the liver of various species and by the liver
plus other tissue homogenates produced primarily DNGs, suggesting that
the liver in vitro was not able to denitrate TNG as completely as the in vivo
system. Livers from mice and humans had a similarly lov ability to metabolize
TNG in vitro. Placental tissue from mice, rats or humans and mouse fetus, liver
or carcass during lata gestation had a poor ability, relative to liver, to
metabolize TNG in vitro.
A. Absorption and Excretion
Because the data on absorption and excretion of labelled TNG were generally
created together by the authors, the discussion of these topics will be presented
together in this report.
Single dosa studies of the absorption and excretion of TNG were conducted by
Lee et al. (1975) using Charles River CD female rats (175 to 250 g). Each rat was
fashed overnight before being given a tingle oral dose of 18(^mg/kg (approximately
the LD ; Hodgson and Lee, 1975) of TNG spiked with the 1,3- C-labeled compound,
suspended in peanut oil, and administered via intragastric intubation at a volume
of 1 ml/100 gm body weight. After dosing, each rst was placed in a "Roth-Delmar"
metabolism cage where expired CO^, feces, end urine were collected. After 4 or 2^
hours, four rats/time period were anaesthetized and aortic blood was collected.

Liver, kidneys, brain, lungs, thigh muscle, and gastrointestinal tract plus
contents were removed, weighed, homogenized and filtered as needed, and
representative samples were assayed for radioactivity, using a Packard Tricarb
liquid scintillation spectrophotometer.
The authors reported Chae about 80Z to 90Z of the oral dose was absorbed in
24 hours. (It was not specified how these figures were determined but It appears
to be by subtraction of recovered fecal radioactivity from total administered
radioactivity recovered.) The radioactivity remaining In the gastrointestinal
(Gl) tract averaged 59.5Z and 3.0Z at the end of 4 and 24 hours, respectively.
The amounts of radioactivity in the feces during the same time periods were 3.12
and 6.3Z. At the end of 24 hours most of the absorbed radioactivity vas excreted
in the urine and expired air, averaging 39.8Z and 25.5Z, respectively, of the
administered dose.
Using the same procedures as described under Lee et al. (1975) for single
dose rat studies, the absorption, excretion, distribution and metabolism of TNG
were also studied in rats during prolonged feeding studies by Ellis et al.
(1978a). Charles River CD rats vere administered unlabelled TNG In their feed at
either 100 or 10,000 ppm for 3, 12, or 24 months before being administered a
single dose of labelled TNG. (The authors calculated actual Intake levels of the
unlabelled TNG to be 3.04 or 363 mg/kg/day for males and 3.99 or 434 mg/kg/day for
females, low and high doses, respectively.) Prior to administration of the single
oral dose (approximately the	Sod&80n Lee, 1975) of labelled TNG, the
rats were fasted overnight for aooy£ 16 hours. Males and females received a dose
of 180 mg/kg each spiked with 1,3- C-TNG, suspended In peanut oil and
administered via Intragastric Intubation in a volume of.10 ml/kg of body weight.
Immediately after dosing, each rat was placed in a "Roth-Delmar" metabolic cage
for the separate collection of urine, feces and CO in the expired air. At the
end of 24 hours, the animals were sacrificed, blooa vas collected from the
abdominal aorta, and various tissues were .removed, velghed and analyzed for
radioactivity via liquid scintillation spectrometry.
These authors reported that reaults were similar to1 those reported earlier In
rats not fed TNG before the metabolism study as conducted by Lee et al. (1975) and
described previously. There were ao consistent differences between sexes or dose
groups, or aaong the different exposure periods. Radioactivity from the oral dose
was excreted 1a the urine, air, and feces.
Lee et al. (1976) used the saae procedures as described above t^study the
similarities and differences in the absorption and excretion of 1,3- C-TNG
administered as a single oral dose of approximately 180-200 mg/kg (Lee, 1987) to
female albino Swiss mice (20 to 25 g), male and female New Zealand rabbits (2.30
to 2.37 kg), male and female beagle dogs (9.0 to 9.2 kg), and male and female
Rhesus monkeys (3.8 to 3.9 kg). The labelled compound was administered by
intragastric intubation and suspended In peanut oil at a volume of 0.2 ml/10 go
body weight for mice and 1 ml/kg for rabbit9. dogs and monkeys.

For comparison purposes, che mean amounts of radioactivity recovered in 2'-*
hours for the various species have been extracted from the reports and are
presented la Table IV-1. (The data for rats are from the study by Lee et al.,
1975, as previously reported and are repeated here for comparison purposes.)
The authors reported that the absorption in nice (501 to 70Z) vas somewhat
lover than in all other species (75Z to 95Z). In rats the majority of the
absorbed radioactivity vas excreted In the urine and expired air while in alee
radioactivity vas excreted In fairly equivalent amounts in feces, urine and
expired air.
In rabbits, dogs and monkeys, most of ths absorbed radioactivity vas excreted
in the urine with only saall amounts in ths expired air.
In othsr experiments conducted by Lee et al. (1976), biliary excretion of
TNG, its dlnltrolsomers (DNGs) and mononltrolsomsrs (MNGs) vsre compared in female
Charles River CD rata (280 to 320 g). Animals vers fasted overnight before use
and then ths cdmmon bile duct vas cannulated through a midline abdominal incision.
Ooaes of TNG, ONGs or MNCs (180 mg/kg for TNG orapproximately the LD._; Hodgson
^gd Lea, 1975), splksd vlth about 10 uCl of 1,3- C-TNG or the respective
C-lab«led compounds, vere dissolved in peanut oil and administered orally by
Intragastric intubation (volumes not specified).
Bile vas collected at variable intervals up to 24 hours and meaaured by
velghlng. Small volume blood samples (200 ul) vera obtained periodically from the
rats' tails. After 24 hours, the animals vere sacrificed and blood was collected
from the abdominal aorta and the entire length of gastrointestinal tract Including
contents vas removed and combined vlth the feces vhlch vere collected vlthout
urinary contamination. Radioactivity levela in the bile, blood, plasma and GI
tract vere then measured using a Packard Trlcarb 3375 liquid scintillation
In female rats administered 1,3- C-TNG, radioactivity appeared in the bile
within 15 minutes aftsr dosing, reached a peak in 3 hours, and thereafter
decreased. The toeal biliary excretion averaged 13.5Z of the dose in 24 hours.
The blood concentration of radioactivity continued to Increase through the Sth to
6th hour and decreased thereafter.
Gross «C al. (1960) as cited in NIOSH (1978) reported an the dermal
absorption of two formulations of TNG la uistar rats. A gel (93Z TNG) and a soft
paste (22Z TNG) vere applied to one cm of depilated skin on the back of the rat
and the area vas covered vlth an occlusive bandage. Application rates vere 100,
200, 400 or 800 mg for the gel and 100, 200 or 400 mg for the soft paste to 24
rats/concentration/formulation. Dressings remained in place for 4 and 8 days for
the gel and paste, respectively. Amounts of TNG remaining In che residues were
determined by velght for the gel and by a colorimetric method for ths paste. For
the gel, 56.8 and 86.8 mg of TNG was absorbed for the lov and high dose,

la	a /
Table IV-1 Percentage of Orally Administered 1,3- C-TNG Recovered in 24 Hours


(180 - 200 mg/kg)c/

Total Recovery
GI Tract plus
Expired Air
Othar Tissues
..Reference: Lee et al., 1975, 1976; Standard deviation* omitcad.
^(n) • Numbers of anlaAla analyzed.
C Reference: Laa, 1987.

respectively, with an absorption rate of 0.85 mg^cm /hr, Irrespective of the dose.
The absorption rate for the paste was 0.63 mg/ca /hr. The amount of TNG absorbed
from the paste was not indicated.
Wester et al. (1983) as cited in Smith (1986) estimated that following dermal
exposure to 22 TNG ointment, the plasma half-life in monkeys was 6.3 hours.
B. Distribution
Lee et al. (1975) In rae studies and Let et al. (1976) In seudles with mice,
rabbits, dogs and monkeys reported on the distribution of recovered radioactivity
in various tissues. The mean values for recovered radioactivity as a percentage
of administered dose are summarized in Table IV-2.
Lee ee al. (1975) reporeed that, in rats, a significant amount of
radioactivity vas found In the liver, averaging 4.6Z of the dose at> hours; the
amount of radioactivity remained near this level at 24 hours. Small but
significant amounts of radioactivity were also found in various other tissues.
Although the muscle radioactivity was rather high at 4 hours (9.3Z of the
administered radioactivity), the concentration declined to 2.8Z in 24 hours in
direct proportion to the decline In plasma concentration.
Similar results were reported by Ellis et al. (1978a) for rats fe^unlabelled
TNG for 3, 12 or 24 months before receiving a single oral dose of 1,3- C-TNG.
The only major concentration within the body was in the liver, the organ of
metabolism and excretion (in bile).
Based on tissue/plasma radioactivity ratios (radioactivity In 1 ml or gm of
wet tissue per radioactivity In 1 al of plasma), Lee et al. (1975, 1976) concluded
that radioactivity was highly concentrated In the livers of all species vlth that
in mice and rabbits being the most highly concentrated. The kidneys of all
species also contained significant amounts of radioactivity. The authors felt
that the concentration ratio of radioactivity in the kidney probably reflected the
excretion of radioactivity In the urine. The tissue/plasma radioactivity ratios
in various spades are sunarlzed In Table IV-3.
In dogs and aonkays, Lee et al. ( 1976) reported findings similar to those in
rats given a single oral dose of the Labelled TNG. While the percentage of
radioactivity recovered in the muscle was 9.5 and 13.0, respectively, after 24
hours (Table IV-2), the tissue to plasma radioactivity ratios were only 0.8 (Table
C. Metabolism
Analysis of radioactivity by thin layer chromatography (TLC) in the 24-hour
urine from 1,3- C-TNG treated rats as conducted by Lee et al. (1975) indicated
that only a trace of the unchanged TNG remained. Averages of 0.7Z and 0.4X of

Table IV-2 Percentage of Orally Administered I,3- C-TNG Recovered From
Various Organs in 24 Hours
Mlce(4) Rabbits(2)	Dogs(2)
(180 - 200 mg/kg)c/
Liver	4.3
Kidneys	0.3
Spleen	- -
Lungs	0.1
Brain	0.1
Skeletal Muscle
(ac 40Z body
weight)	2.8
Whole Blood
(at 7Z body
velght)	0.7
^Reference: Lee et el., 1975, 1976; Standard deviations omitted,
(n) " number of animals.
Reference: Lee* 1987.

Table IV-3 Tissue/Plasma Radioactivity Rat'io,24 Hours After Oral
1,3- c-TNG Administration3
Rats(4)b/	Mice(4) Rabbits(2)	Dogs(2) Monkeys(3)
(Dose)	(180 - 200 mg/kg)c/
Liver	7.8d/	30.1	21.0	7.5	7.4
Kidney	2.8	4.1	4.3	2.7	2.2
Spleen	-	2.6	3.3	1.8	1.3
Lungs	>1.7	2.8	2.8	1.4	1.3
Brain	1.0	1.2	l.S	0.7	0.9
Muscle	0.6	1.8	0.6	0.8	0.8
. .Reference: Lee ec al., 1975, 1976.
c.(n) - number of animals.
..Reference: Lee, 1987.
Radioaccivlty in 1 gm of wee tissue per radioactiviey in 1 ml of plasma;
standard deviation omitted.

che administered radioactivity were free I,2-DNG and 1,3-DNG, respectively. The
free MNGs accounced for 10.6Z of the administered radioactivity. Analysis of the
MNGs by TIC indicated the ratio of 1-MNG:2-MNG to be 0.9 t 0.2. The majority of
the radioactivity was unidentified polar compounds, these unidentified polar
compounds vera eluted, treated with beta-glucuronldase, and again chrooatographed.
No TNG was detected. However, 10.0Z of the administered radioactivity was
1.2-DNG,	3.SZ vas 1,3-DNG, and 1.5Z was MNGs, all as glucuronldes. Glycerol
accounted for 6.9Z of the administered dose, and 6.2Z remained unidentified polar
These results In rats were compared by Lee et al. (1976) to those in mice,
rabbits, dogs and monkeys afte^a single oral dosa of 180-200 mg/kg (approximately
the U>iq f°r rata) of tha 1*3- C-TNG in volumes of 1 ml/100 gm body weight
for rata. 0.2 ml/10 ga for mice and 1 ml/kg for rabbles, doga and monkeys).
Trlnitroglycerin and free DNGa were excreted only in small amounts in the
urlna of mice, rata and rabbitst (1.2Z or less of tha administered dose) and in
amounts slightly greater in the urine of dogs and monkays (4.3Z and 2.8Z,
respectively). Larga amounts of unidentified polar compounds and glycerol were
found In all species. Mica excreted only small amounts of fraa MNG, MNG-
glucuronldes and DNG-glucuronldea (4.6Z), indicating ralativaly complete
biotransformation 'of TNG. Moat of the urinary matabolltes in rats and rabbits
wera free MNG and DNG-glucuronldes. Dogs and monkays exeretad mostly free MNGs
and MNG-glucuronidaa (23Z).
Tha lover rata of absorption and complete metabolism of TNG In mice was
considered by tha authors to reflect the relatlva toxicity of TNG In this species.
(As Indicated under Short-term Exposure, tha acute oral toxicity of TNG was
slightly lover In mica than in rats.)
Tha matabolism of TNG vas slao studied In rats during prolonged feeding
studies as conducted and dascrlbed by Ellis at al. (1978a). Metabolic reactions
included danltrlflcatlon toward glycerin, glucuronidatlon of varloua
nitroglycerins, and oxidation and other reactions of tha glycerin. The fate of
tha dissociated nitro groups is unknown. Tha urinary matabolltes Included little,
if any, TNG, butrelativaly large amounts of tha di- and mononitroglycerlns and
their glucuromldas. Othar major metabolites included glycerin and some
unidentified polar compounds, presumably metabolites of glycerin.
In experiments by Lea at al. (1976) on
(2.42), other major metabolites in the bile were MNG-glucuronides (2.OS) . Free
MNGs accounted for 1.0Z of Che dose.
In separate experiments by these authors, as summarized In Lee et al. (1976),
tha concentration of 1,3-DNG in the first-hour bile was higher than that in the
five-hour bile sample, whereas the concentration of MNG was higher in-the
five-hour bile than in the first-hour bile. This indicated progressive
denitration of the DNG to MNG.
In vitro studies of the metabolism of TNG vara conducted by Lee et al. (1976).
using honoganates of livers from various species including humans, as well as
placentas and tissues of several species during development. In the in vitro
system as modified from Needleman and Krantz (1965)* 1.3- C-TNG was metabolized
by tha rat liver'primarily to 1,3-DNG and 1,2-DNG after a 15-minute incubation
period. Whan tha reactions were continued for up to 2 hours* tha amount of polar
components Increased aa a function of time. The latter compounda vera thought to
represent MNGa becauae incubation of 1-MNG In the same system resulted in 94Z
present aa tha parent compound after a 90-mlnute incubation. These observations
suggssted that tha rat liver in vitro was not abla to completely denitrace IMG to
glycerol and CO^ as did the in vivo system.
These authors^glso investigated the roles of organa other than the liver in
the conversion of C-TNG to CO^ vhlch occurred in the Intact animal but vas
blocked when the abdominal organs vere excised. Since the removal of the stomach,
lnt^gtines* pancreas* spleen and kidneys apperently interfered with the metabolism
of C-TNG. they postulated that these organa acted in conjunction with the liver
to metabolize TNG. This hypothesis vas tested in the in vitro system by incuba-
ting homogenates of tha^iver alone and^th the addition of various other organs
in the presence of 1,3- C-TNG and 1,3- C-l-MNG. The addition of tha various
tissues removed during excision did not modify tha ability of tha liver to
metabolize either subacrata. Therefore, the results of this experiment still
leave in question eha rola of the liver and various other organa in the complete
denitration of TNG to glycerol and ultimately to
Lee et al. (1976) also conducted comparative studies of TNG metaboll^ by
incubat^gg homogenates of llvsre from various species with 1 omol of 1*3- C-TNG
or 1,3- C-l-MHG for 10 minutes. Results of this study are shovn in Table IV-4.
Livers from rata* dogs, rabbles, arid monkeys vere not able to metabolize
1-MNG to glycerol; tha primary metabolites of TNG vere 1*3-DNG and 1,2-DNG. No
sex diff-erencea In TNG metabolism were observed for any of the animal species
tested. Livers from rats and mice produced higher levela of 1,3-DNG than 1,2-DNG.
In contrast* higher lsvels of r,2-DNG were produced by livers from rabbits, dogs,
monkeys and humans. Livers from mice and humans had a lover ability to metabolize
TNG in vitro, as evidenced by a lover recovery of the labelled metabolites when
compared to other spaciea. In addition, :he human livers had a lover ability to
form 1,3-DNG than any of tha other species The formation of 1,2-DNG by livers
from rats and humans vas reported to be approximately equivalent.

14	a/
Table IV-4 Accumulation of Metabolitg^-of 1,3- C-TNG
la Liver Homogenates
Rata(7) Mice(6) Rabblts<6) Dogs(7)	Monkeys(5)	Huaans(4)
naol/ng Procain
1,3-DNG 40.6d/ 16-7 18^3 2l7i	16^3	12.5
1.2-DNG 28.0 7.0 51.3 46.8	52.7	32.5
Raeloa/ 1.4 2.4 0.36 0.45	0.31	0.38
. (incubated vlth 1 oaol of 1.3- C-TNG.
^Referenda: Lea at al., 1976.
. .(n) ¦ number of determinations
ytnale and female values combined; standard deviations omitted.

Lee ec al. (1976) also studied the mecabolistn of TNG during developmental
stages of various species. Placencal tissue from mice, racs, and humans had less
ability chan the liver to produce 1,3-DNG and 1,2-DNG. Low levels of TNG
metabolism were found in mouse fetuses on day 12 of gestation and In fetal livers
and carcasses on day 18 of gestation. Metabolism of TNG in rat livers increased
between 1 and 7 days after birth and did not change between 7 and 21 days after
birth. Livers from rats at 3 weeks of age produced lower levels of DNGs than did
the adult liver. These results indicated chat at 3 weeks of age, rats do not
produce the adult pattern of TNG metabolites.

v. health effects
A. Health Effects in Humans
Available data on che effaces of TNG la humans is largely based upon its
usage as a vasodilator la che clinical setting. As described in Needleman and
Johnson, 1980, administration of therapeutic doses of TNG, 0.1S to 0.6 mg, affects '
the cardiovascular system and results in vasodilation and a general relaxation of
all smooth musculatura. Sublingual doses of 0.65 mg administered up to 20
times/day and survival after quantities up to 400 mg have been reported by Munch
and Friedland as cited in Oacre and Rosenblatt, 1974. Untoward responses are
secondary Co the therapeutic effect and Include headache, often severe, usually
over the flrac few days of treatment, and "presumably due to dilatation of
meningeal arterial vessels". Arteriolar dilatation In the face and neck also say
result In flushing. Dizziness, weakness and pallor may occur and is usually
transient and associated with postural hypotension. Rash may also occur. It has
.been shown, however, that frequent add repeated exposure can result in the
development of'tolerance to these effects. Several studies on the effects of
industrial exposure, either in the pharmaceutical or explosives manufacturing
process* have provided additional information.
In a controlled experiment reported by Schwartz (1946). 15 edult males were
given a patch test involving the application of a dime-sized aree of dynamite (20X
TNG) to the back over the left scapula, tflthln three hours of application, 14 of
the 15 subjects reported severe, throbbing headaches associated with a feeling of
warmth over the face and some nausea. The headache was not relieved by aspirin
and, in several cases, became more severe, resembling th« headache caused by che
Injection of histamine and its dilatory effect on the intracerebral vessels.
Routine blood counts and electrocardiograms were normal. Further studies
indicated that the headaches could be relieved by Intramuscular Injections of 0.5
gm of caffeine sodium benzoate followed by oral administration of 5-10 mg of
amphetamine sulfate and could be prevented by administration of 15 mg of
amphetamine sulfate for 2-3 days prior to exposure.
Bresler (1966) reporeed that, in severe cases, the Intense headache may be
accompanied by psychic disturbances such as dizziness, mental confusion,
hallucinations* pugnaeiousness and maniacal manifestations. Respiration,
initially hypsrpnelc. becomes dyspnelc and. slow, the skin becomes cold and
cyanotic and ths pulse slows and becomes dicrotic and intermittent. Paralysis and
clonic convulsions may follow. Deech due to respiratory failure can occur in 4-7
hours. These symptoms have been found to be exacerbated by the ingestion of
alcohol wtvlch acts to further relax the blood vessels. Symptoms may also be
enhanced by warm weather, presumably by increasing the levels of TNG in che
atmosphere. Prolonged skin contacc can result in eruptions of variable severity,
particularly ulcerations of the finger tips and below the nail beds. While che
toxic dose varies due to Individual sensitivity, initial immunity Is rarely

L'pon continuous workplace exposure, symptoms usually disappear rapidly,
indicating the development of tolerance to these effects, in 1966, Hanlon and
Frederick reported that workers exposed during the manufacture of TNG tablets
developed headaches and irritation whejj airborne concentrations were in the range
of 0.03 to 0.II ppm (0.28 to 1.02 mg/m ). Exposure was intermittent (2 to 3 times
per week) and tolerance did not develop. Reduction of airborne concentrations to
<0.01 ^pm (0.09 mg/tn ) no longer resulted in headaches.
While tolerance to the early symptoms of TNG exposure, including the
headaches, develops rapidly* other symptoms have been shown to reappear after a
few days avay from the workplace. This form of "organic nitrate dependence" can
be described as the production of physical withdrawal symptoms by removal of the
causitive agent. In the ease of TNG* "Monday morning angina" has been reported as
the most serious effsct of chronic TNG exposure. Sudden death or myocardial
infarction has occurred in individuals with no demonstrable vascular disease after
a few days break in chronic exposure (approximately 24-72 hours). Angina pains
often preceed such an attack.
In 1968, Lund et al. described nine cases of toxic effects due to exposure to
TNG at an explosives factory in Sweden. All but one worker were male* ages
ranging from 35-55 with an average age of 46. Exposure periods ranged from 2-27
years with an average of 12 years. Eight cases were non-fatal but involved the
development of headaches upon Initial exposure and chest pains 30-65 hours after
the last contact with TNG. No pathological conditions wsre found nor were there
any abnormalities in the electrocardiograms (ECGs). No further episodes occurred
in five workers removed from the TNG environment and followed for five years. One
worker, a 35 year-old male, died of myocardial Infarction but showed no signs of
atherosclerosis of the coronary arteries. Five months prior to his death, he
reported attacks of chest pain* occurring on Sunday evenings or during the night
but disappearing shortly after returning to work. The ECG and blood pressure were
normal as was the physical examination.
In 1971, the United States Army Environmental Hygiene Agency (USAEHA) as
cited in Shiotsuka* 1979, conducted an evaluation of the TNG operations and the
workers exposed to TNG at the Badger Army Ammunition Plant in Baraboo, Wisconsin.
Of the 266 potentially exposed workers, 33 (122) were suspected of suffering from
adverse effects due to exposure to TNG. Symptoms Included evidence of hesrt
disease subsequent to expoeure (30Z). chest pains (18Z), death (6Z, specific cause
not reported) and headaches or rash with no other symptoms (21%). A seperate
survey at this same facility indicated chat 90X of the individuals evaluated
reported headaches. Additionally, many workers reported the.occurrence of severe
headaches after returning to work following an absence of a week or more.
Atmospheric samples taken at various locations in the plant indicated
concentrations ranging |rom 0 to 12.53 ag/m with an overall average concentration
of approximately 2 mg/m from a total of 117 samplings.
In 1972, Lange et al., in an independtnt study of what appears to be the same

Army munitions plane in Baraboo, Wisconsin, reported on nine cases of
nonacheromatoua ischemic heart disease among approximately 200 TNG-exposed
workers. Chest pains were reported in eight females and one male, ages 38-54 and
employed for 11-48 months. All experienced angina while away from the workplace
or on Monday mornings, approximately 46-70 hours post-exposure. Two fatalities
occurred, a 40-year old woman with 11 months exposure and a 34-year old man
exposed for four years. No autopsies were performed. Among the non-fatal cases,
aLl developed nonexertlonal chest pains on weekends or Monday mornings with relief
occurring upon return to work. Sublingual TNG on weekends also relieved the chest
pains. Three Individuals showed evidence of impaired left ventricular performance
with exercise intolerance. In general, coronary vessels were normal and the
individuals became asymptomatic upon removal from the TNG environment. In one
cast, coronary angiography conducted during an episode of chest pain showed severe
spasms of the coronary artery which were relieved by sublingual TNG.
A similar but less intensive study at the Radford AAP at RadfoTd, Virginia,
indicated that nearly all sampled areas had TNG concentrations in excess of 2
mg/my the Threshold Limit Value (TLV). Exposure to concentrations greater than 2
mg/m are known to result in headache (ACGIH, 1971 es cited in Shlotsuka, 1979).
Tolerance to this effect is reported to occur after approximately one week'of
exposure. The severity of withdrawal from effects belov the TLV has not been
evaluated. Despite over a century of use aa a vasodilator, no significant acute
toxicity hazards to TNG have been reported nor are any reports of chronic toxicity
or carcinogenicity knovn.
Klock (1975) also reported on the occurrence of substernal chest pain
accompanied by spasm in the midportlon of the right coronary artery that occurred
during coronary arteriography and angiography in a 38 year-old male exposed over a
period of 10 years to dynamite containing variable amounts of TNG and ethylene
glycol dinitrate. The subject was being examined for recurrent attacks of sudden,
severe, crushing, substernal cheat pain that occurred while at rest and while away
from the workplace. Pain was relieved by sublingual TNG or return to work. A
complete hematological and blood chemical workup was normal as were the resting
and exercise ECGs. The coronary arteries were otherwise normal.
In 1978) Feldman et al. reported on two cases in which segments of the right
coronary arteries* appearing angiographically normal* narrowed by 50-90Z after
admlnlstration of 0.4 mg of sublingual TNG. The area of narrowing was not near
the catheter tip* did not respond to additional TNG therapy, persisted for at
least 20 minutes, did not cause cheat pain and could not be reproduced by
subsequent administration of TNG or provoked by ergonovlae maleate.
In an occupational health study of dynamite-producing factories in Sweden,
Hogstedt and Andersson (1979) reported a slight but significant Increase in
mortality rate due to diseases of the circulatory system among workers exposed eo
dynamite in the workplace. Nine deaths due to cardio-cerebrovascular disease
occurred among 88 exposed male workers during a 12 year observation period during

which che expected death rate was estimated to be 4.5. Six of these deaths were
due to ischemic heart disease with the remaing due to cerebrovascular diseases.
One of the deaths was described as the typical "Monday death" occurring in a 26
year-old femala exposed for two years. No signs of coronary sclerosis were noted
in che autopsy.
In mora recent years, physical examination of workers exposed to nitrates in
the dynamite Industry has been followed more routinely. Common findings have
included headache, nausea, palpitations and flushing of face and neck. Other
findings, as cited in NIOSR (1978), Include reduced blood pressures, slight
leukopenia and normochromic anemia but no methemoglobinemia or Heinz bodies
(Maccherlni and Camarri, 19S9); peripheral circulatory disorders and paresthesia
(Rubino at al., 1956), Raynaud's phenomena and transiene ischemia of tha limbs
(Sartalini et al., 1967), alcohol intolerance (Prerovska and Teisinger, 1965) and
peripheral neuropathy (Jacob and Maroun, 1969). Skin rash and edema have been
reported bue not frequently and it was concluded in one study (Planques et al.,
1959) that TNG actad a^ both an irritant and a sensitizer. In those fatal eases
where autopeies' were performed, moderate to severe scleroaia of the coronary
arteries was reported along with thickening of the vessels with hyaline and
connective tissue. All other organa were normal (Carmichael and Lieben, 1963).
8. Health Effects in Animal Experiments
As determined by experiments in dogs, rats, and mice, when TNG was
administered at high dosage levels (these vary with tha duration of TNG
administration), adverse affects were most often observed as a decreaae in weight
gain which appeared to be directly related to a decreaaad food consumption.
Various other tiasuaa were affected including the liver and blood. Table V-l
summarizes these toxicity studies.
1. Short-Term Exposure
Acuta oral LD^ levels for TNG in alee and rata indicated that TNG was
somewhat more toxic in rata than in sice.
In a study by Lea et al. (197)). rets end mice fasted for at least 16 hours
were admlnlstart* 3.41% TNG (as 9.7X TNG la lactose suspended in peanut oil) by
intragastric intubation. Tha volume delivered was not stated. A suspension of
lactose in peanut oil waa used ai the vehicle control. The LD _ waa calculated by
a computer program baaed on the tðod of maximum likelihood or Finney (1971).
Range of doses adminlatered was not specified. Survivors were observed daily for
14 days for delayed mortality or toxic signs.
The acuta oral LD,. levels t S.E. (95X confidence limits) in male and female
rats ware 822 t 54 (70d°953) and 884 t 61 (763-1,055) mg/kg. respectively; in male
and female mice, LDeo lavels were 1,188 : "6 (1,008-1,352) and 1,055 t 63
(895-1,178) mg/kg, reapactlvely. Within cr.e Sour after dosing, "all" the animals

Table V-1 Summary of Toxicity Studies on TNG In Animals
Reference	Species	Dose	Route	Duration
mg/kg/day®^	weeks"'
Lee et al. (1975)	rat, Mouse	—	oral
rabbit	7.29X	dermal, ocular	72 hours
guinea pig	3.412	dermal
Lee et al. (1976)	dog	25, 50, 100, 200 oral	5 days
0.01, 0.1, 1.0
raised to
0.05, 0.5, 5.0

rat (male)
0.8, 6.0, 59.0
0.9, 6.4, 59.3
raised to
2.6, 24.5, 229.5
3.1, 26.5, 233.8

mouse (male)
1.3, 11.5, 106.7
1.3,	10.9, 94.9
raised to
6.4,	60.2, 607.6
6.9, 58.7. 561.2

rat (sale)	1,406	oral	13
(female)	1*416
Ellis et al. (1978a)	dog	1, 5, 25	oral	52
rat (male)	3.04, 31.5, 363	oral	104
(female)	3.99, 38.1, 434
mouse (male)	11.1, 115, 1020	oral	104
(female)	9.7, 96, 1060

Table V-1 - continued

Oketanl et al. (1981a)
5, 1.0, 4.0
days 6-16
Oketanl et al. (1981b)
rat (ule)
10, 20
63 days
prior to mating
14 days prior to
mating through
gestation day 7
Oketanl et al. (1981c)
10, 20
daya 7-17
Oketanl et al. 1198Id)
10, 20
gestation day 1/
through day 21
of lactation
A /
..Unless otherwise stated
Continued for additional weeks

became cyanotic and ataxic. The ears> nose, eyes, paws, and tail appeared very
pale, and respiration was depressed. Death usually occurred within 5 to 6 hours
after dosing. The animals that survived the dose usually recovered in 24 hours.
No gross pathology was noted in the animals that died. Measurements of
methemoglobin were not reported in either species.
The Registry of Toxic Effects of Chemical Substances (RTECS, 1986) lists oral
LD^q values in rats ac S2S tag/kg and in mice ac 500 mg/kg. Neither the strain nor
sex of ehe animals was given nor the vehicle of oral administration. Variations
in these factors may account for the differences in LD^ values.
a.	Primary Skin and Eye Irritation Studies
Using the modified Dralze method for skin and eye irritation, a 7.29Z TNG
paste (25Z peanut oil and 75Z of the 9.72Z TNG in lactose) was applied directly to
the eye or shaved skin (lneact and abralded) of the rabble and evaluated at 24 and
72 hours (Lee ee al., 1975). This concentration of TNG vaa a very mild skin
lrritane but not an eye Irritant.
b.	Dermal Sensitivity Study
Using the "maximization test" described by Magnusson and Kllgman (1969) as
cited in Lee et al. (1975). a preparation of 3.41Z TNG (as 9.7Z TNG in lactose
suspended in peanut oil) was applied to the skin (clipped free of hair) of guinea
pigs. A 40Z positive response vaa elicited. Indicating that TNG is a moderate
sensitizing agent.
c.	Four-Week Studies
In four-veek oral toxicity studies conducted by Lee ee al. (1976) in dogs,
rats, and mlce( the only adverse effecta obeerved were deereaaed feed consumption
and decreased rata of weight gain in high-dose rats.
In the study with dogs* a total of 32 young healthy beagle dogs (6.6 to 11.8
kg) were divided into four groupe. each conelstlng of 4 males and 4 females.
Three groups of dogs vere given 0.01, 0.1 or 1 mg/kg/day of TNG in capsules. The
fourth group served aa conerols and was given empty capsules dally throughoue che
experiment. Tha tests Included hematology, clinical blood chemistry tests, serum
electrolyte measurements,, urinalysis, che presence of occult blood, fasting blood
glucose and bromoeulfophthaleln (BSP) retention. Ae the end of 4 weeks, adverse
effects were not observed in any dogs. Also at the end of 4 weeks of continuous
treatment, one male and one female dog from each dosage group were sacrificed for
necropsy; no adverse affects or lesions were found in any of thess dogs during
gross or microscopic examination of tissues. A No- or Lowest-Observed-Adverse-
Effeet-Level (NOAEL or L0AEL) was not determined for this study due to an
insufficient number of animals to evaluate the effects or lack thereof.

In che four-week, study with racs (Lee ec al., 1976), 64 male and 64 female
young healthy CD Charles River rats were divided into four groups, each consisting
of 16'oales and 16 females. Three groups of rats were fed 0.001, 0.01 or 0.1Z INC
by weight la eh* feed (0.8, 6.0 and 59.0 mg/kg/day for males and 0.9, 6.4, and
59.3. mg/kg/day for females when corrected for evaporation losses and food consump-
tion patterns). The fourth group, serving as the controls, was given tha powdered
standard rodent chov without TNG. Testa conducted vara tha saaa as those dona
with dogs except that BS? retention vas not measured. At tha end of 4 weeks of
treatment, 4 males and 4 females from each group were sacrificed for necropey.
Feed, conaumptlon levels and weight gain of male and female rata at the
highest doaaga laval ware significantly dapraaaad after 4 weeks of treatment.
This effect vaa shown to be reversible when TNG feeding vaa dlacontinued.
Additionally, tha degree of weight depression vaa comparable to the reduced feed
Intake. (Saa Appendix 1, Tablea Al-1 and Al-2, males and feaales, respectively,
4-veek study at 0.1Z) Tha absolute and/or relative valghts of tha liver and/or
kidneys of soma rata fed TNG for 4 weeks were allghtly, but significantly,
different from thoae of the control rats. Vhila soma changes occurred at all test
levels, the effects ware not consistent aa to doae or sex. The authora concluded
that these changea vara not doaa-related and vara not considered to be clinically
significant. Groas and microscopic examination of tiasua at nacropay revealed no
treatment-related lesions at any level. Tha NQAEL based on veight dapreaslon
would bs approximately 6.2 mg/kg/day.
Lee et al. (1976) alao conducted a four-week study in mice. Sixty-four male
and 64 female young healthy albino Swiss mice ware divided Into four groups, each
conalsting of 16 malaa and 16 femalaa. Three groups of mice were fed TNG In the
feed at 0.001, 0.01, or 0.1Z (actual intake levels vere calculated based on
evaporation loaaas and food consumption patterns to be 1.3,,11.5, or 106.7
mg/kg/day, respectively, for aalas and 1.3, 10.9, or 94.9 ng/kg/day, respectively,
for females). Tha 4eh group served aa controla and vas not admlniatered TNG in
the feed. At tha and of 3 veaka, adverse effects vara not obaarvad In any mice.
Starting tha fourth vaakt TNG concentratlona in feed for all groupa vere lncreesed
to 0.005, 0.05 or 0.5% (actual intake levels vera calculated by the authors to be
6.4, 60.2 or 607.6 mg/kg/day, reepectively, for males end 6.9, 58.7 or 561.2
ag/kg/day, respectively, for femalee). Using tha same toxicologlcal parameters as
described abovo for rats (except chat blood chemiatry tests and organ veight
analysis vers set conducted in mice) no creacment-related adverse effects or gross
or microscopic lesions vera observed afcer 4 weeke of TNG treatment. It la
inappropriate to calculate a N0A£L or L0AEL for thia study due to the change of
doses after Week 3 of the atudy.
2. Longer-Term Expoaure
a. Thirteen-Week Studies
In thirteen-veek studies conducted in doga, racs and mice, the only

observable significant effects were elevated SCOT Levels in high-dose rats. These
studies conducted by Lee et al. (1976) were essentially a continuation of the
four-week studies described above.
In the study with dogs* TNG dosages administered during the first 4 weeks
(0."0l, 0.1, and I mg/kg/day in capsules) did noe elicie observable adverse effects
in any of the test animals. Therefore, starting on the fifth veek, dosage levels
were increased to 0.05, 0.5 and 5.0 tag/kg/day for the remainder of the
thirteen-week study and the number of dogs per sex per test group vas decreased
from 4 to 3 at the tvo lover dosages and to 2 dogs per sex at the highest dosage.
Ho treatment-related adverse effects were observed in routine hematological or
clinical chemistry tests. Gross and microscopic examination of tissues from one
dog/sex/level at necropsy shoved no TNG-related changes in tissue after treatment
for 13 weeks. It is inappropriate to determine a N0AEL or L0AEL for this study
due to an insufficient number of animals at its termination.
Similarly, TNG dosages administered to rats during the first five veeks
(0.001, 0.01, 0.1X la the feed) did .not elicit observable adverse effects other
then the intake related velght depression and doses vara, therefore, increased to
0.005, 0.05, and 0.5Z in the feed (2.6, 24.5, and 229.5 mg/kg/day for males and
3.1, 26.5, and 233.8 mg/kg/day for females based on evaporation losses and food
consumption patterns) for the remainder of the thirtaen-veek study (Lee et al.,
1976). No treatment-related adverse effects vera observed in hematological or
clinical chemistry tests vleh the exception of elevated SCOT levels in tvo rats
dosed ac the high level. These elevations vere reversible after a four-week
recovery period when TNG vas not administered. Tissue lesions observed at
necropsy (4 rats/sex/level) vere not considered by the authors to be TNG-related.
^These lesions Included lymphoid hyperplasia and/or pneumonia in the lung,
myocarditis, necrosis of hepatic cella, and hemosiderosis or extrsmedullary
hematopoleals in the spleen.
The absolute and/or relative weights of heart* liver, kidney, thyroid,
adrenal and/or testes of the rats fed TNG for 13 veeks vere slightly but
significantly different from those of the respective control rats. However, with
the exception of the absolute and relative weight of tha thyroid which vas
significantly reduced in males at all test levels In a non-dose-related manner,
these changes seemed to occur at rendoo dosage levels and vere both higher and
lover than coacrols. Therefore, che authors concluded that these changes were not
dose-related and vare not considered to be clinically significant. Body weights
of male and female rats fed TNG at the high dose level (approximately 232
mg/kg/day) were significantly decreased as compared to control rats throughout che
study. A N0AJEL based on body velght depression would be approximately 25.5
In a follow-up study, Lee et al. (19:6) attempted to establish a dose-
response relationship for the adverse effects (depressed feed consumption and
weight gain) observed during the first f	phase of their study viLh rats.

Trlnitroglycerol was fed ac 2.5Z of the diet (consumption levels corrected for
evaporation losses and Intake patterns were reported as 1,406 and 1,416 mg/kg/day
for males and females, respectively) for 13 weeks to 3 male and 3 female young CD
Charles River rats. While these feedlag levels were higher than the acute oral
LD in rats as previously determined by Lea et al. (1975), le can be assumed that
TNG is less toxic when consumed in Che dlae than whan lc is administered as a
single bolus. Both mala and femala rats fed TNG consumed approximately half the
amount of feed as the control rats and lost valght quickly. They continued to
l~ose weight, vara slightly lass activa, and had rough hair coat through the eighth
waak. At this point, both mala and fsaale rats weighed approximately one half as
much as control animals and vara consuming approximately half or less the amount
o£ feed. Thereafter, eheaa rats started to gain weight, food consumption
increased to becvaen 72X and 105Z of control intakes and thalr conditions improved
(Tables Al-l and Al-2, Appendix 1).
Rats fed TNG at this high lavel had soma changes la hematological parameters,
organ weights and lesions In the testes, liver and splean. Erythrocyte count,
hematocrit, hemoglobin concentration, and serum level of alkaline phosphatase
Increased and faating blood glucose decreased in both tha mala and femala rats.
In tha malea, TNG feeding Increased the relatlva weights of spleen, kidney and
brain baaed on tha body weight; in females, TNG feeding increased the relative
veights of splaan and kidnay based on the body valght. Testicular velghta vere
not reported; hovaver, TNG causad teaticular atrophy and degenaratlon, and
aspermatogenesis in tha thraa test malea. Both tha males and the females fed TNG
at approximately 1,400 mg/kg/day had hemosiderosis in the spleen and/or the liver.
In the thirteen-veek studies (continued from the fqur-veek studies by Lee et
al., 1976 and described previously) mice were fad dosaga levels of 6.4, 60.2, or
607.6 mg/kg/day for males and 6.9, 58.7, or 561.2 mg/kg/day for females; there
vere no treatment-related or statistically significant adverse effects observed by
the authors. Absoluts and relative spleen walghts of faaala mice fed TNG at the
middle and high dosa levels vera larger than thoaa of control females but were not
statistically significant dua to large individual variatlona among these animals.
Spleen weights from tha faaala mica fed the lov dosa lavel vara not available.
Possibly related*to tha lncreaaed spleen weight aay have been tha extramedullar
hematopoiesl* which occurred In the liver and splaan of all male and female mice
fed the high and aiddla dosa levels of TNG and tha splsan and/or liver of one male
and one feaal* aouse fed tha lov doee. The authors reported this effact was not
dose-related, was mild and is cosonly observed in rodents. However, the
hemosiderosis of tha splsan vas not noted In any control mice. Therefore, this
study identifies a LOAEL of 6.7 mg/kg/day based upon effects noted in che spleen
of TNG creaced mice.
b. Lifetime Exposure Studies
Long-term exposure to TNG in che diet for up to 12 or 24 monchs has been
reporced co produce adverse effeccs in a variety of species Including dogs, rats

and mice. Among the effects observed were decreases In weight gain and feed
consumption in rats and mice, especially in the high dosage group during the first
week of treatment. Additionally, pathologic alterations were apparent in certain
target organ* including the blood for all three species and the liver and testes
in treated rats only.
Ellis et al. (1978a) studied the effects of oral administration of TNG for up
to 24 months in dogs, rats and mice.
In dogs (6/sex/dose), TNG was administered at 1, 5, or 23 mg/kg/day by
capsule for 12 months. High doss dogs received tvo capsules of ths test substance
daily. Various parameters including routine hematology, clinical chemistry and
hlstopathology were evaluated. Treatment with TNG produced a slight degree of
dose-related, transient methemoglobinemia, but no histopathologic lesions vere
reported (3 dogs/sex/level). In almost all cases, the amount of methemoglobin
was less than 3X indicating it could be due to an artifact in the measurement
methodology. However, after nine months of treatment with TNG, there vas a
significant dose-related trend in methemoglobin levels $ statistically significant
only in males (P < 0.0S). Since the amounts were small, the authors based their
dose relationship on Incidence of occurrence rather than amount present. ~ At the
two highest dose levels, both sexes shoved evidence of methemoglobinemia, while at
the lowest dose level ( 1 mg/kg/day), methemoglobinemia was evident in half of the
males. However, one high dose male and one high dose female had the highest
levels of methemoglobin In the entire study, 5.31 and 3.7Z, respectively.
Methemoglobinemia was no longer apparent by 12 months of feeding and none was
found after one month recovery studies. Because observed effects were minimal
after 12 months of feeding, treatment of the remaining 3 dogs/sex/level was
discontinued for a 4-week recovery study at which time the remaining animals were
necropsled. No necropsy data vere provided from the recovery study. Considering
the posslbllty of artifacts and the lack of a consistent dose-resonse throughout
the study and during the 4-week recovery study, it Is not appropriate to derive a
NOAEL or LOAEL for this study.
Male or female rata (38/sex/treatment level) were administered TNG for up to
24 months at doses of 3.04, 31.5 or 363 mg/kg/day and 3.99, 38.1 or 434 mg/kg/day,
respectively (Bills ee al., 1978a). Various parameters including routine
hematology, clinical chemistry and hlstopathology were evaluated. The low dose
level (3.04 or 3.99 mg/kg/day to males and females, respectively) fed for up to 2^
months produced no adverse effects. At the middle dose level (31.5 or 38.1
mg/kg/day administered to males and females, respectively) some hepatotoxlc
effects (hyperplastic foci) were observed as early as 12 months Into the study.
Rats administered the high dose (363 or 434 mg/kg/day to males and females,
respectively) shoved decreases in weight gain and smaller decreases in feed
consumption. The doses used did not significantly Increase mortality as indicaced
In Table V-2. In the high dose animals, che incidence of mortality was actually
lover than that in the control and the other dosage levels.

Table V-2 Incidence of Mortality In Rats Fed IN'G for 24 Months



Veeks Control
Total Deaths
. .Reference: Elite ec el.» 1978e.
Number of animals/time period; includes only ehoee animals dying and necropsied
.during the course of Che experiment.
..Presumed eo be found deed end discarded.
Starting no. m 38/sex/level. Assume substitutions for early deaths included.

High dose rats also showed methemoglobinemia, hepatomegally, pigmentation of
the epithelium of the kidneys and spleen, and Interstitial cell tumors in the
testes. HiaCopathological lesions in the liver were also reported. These
included primarily fibrous bile duct hyperplasia and non-fibrous (cystic and/or
adenomatoid) hyperplasia as well as hyperplastic foci, neoplastic nodules and
hepatocellular carcinoma. Some of the above effects ae the high dose levels were
evident as early as 12 months Into the study. (See incidence Tables A2-1, A2-2
and A2-3, Appendix 2, for a breakdown as to elm* of occurrence of the
hepatocellular lesions In males and females ae control, middle and high dose
levels.) Evidence of oncogenicity is detailed in Che carcinogenicity section
which follows. A KOAEL of approximately 3.5 mg/kg/day is based upon the
occurrence of hyperplastic foci in the liver of both sexes.
Ellia ae al. (1978a) raporeed that mala and female mica (58/sex/level)
administered the low doaa (11.1 and 9.7 mg/kg/day, respectively) or ehe middle
dose (115 and 96 mg/kg/day, respectively) for up eo 24 months shoved no eoxic
effects aa evidenced by behavior, appearance, body weight, food intake, hematology
and histopaehology. However, ehe high dose, about 1020 and 1060 mg/kg/day TNG
intake for the males and females, respectively, was toxic aa evident by decreased
feed consumption and weight gain, behavioral effects, and methemoglobinemia. The
methemoglobinemia had sequelae including Heinz bodies, anemia and pigment deposits
described aa hemosidarln-lika pigment usually goldsn-brovn and most commonly seen
in the liver. The most common spontaneous lesiona seen in these mice at control
and all test levels Included amyloid deposits, often so widespread as to be
considered "generalized". Other lesions were described as degeneraeive and
attributed eo aging. These included chronic myocarditis and nephritis and fatty
changes, focal necrosis and degeneration of the liver. Various tumors were
reported bue wars not considered by the authors eo be associated with TNG Intake.
These tumors Included bronchoalveolar adenoma, cystadenoma and follicular cell
tumor of eha ovaries and chromophobe adenomas of eha pleuieary. A high mortality
race among mala mica ae the control and all ereacmene levels rssuleed in a
mortality rata of approximately SOZ by 18 moneha at all levels, with no surviving
high-dose males by 24 monehs. The lack of sufficient animals for evaluaelon of
effecea in the males limles ehe usefulness of this study and makes derivation of a
NOAEL or LOAEL inappropriate.
3. Genotoxlcology
Mutagenic activity has not been detected in various studies employing certain
indicator systems* including bacterial and yeaat assays and mammalian assays under
conditions of the tese.
Ellis ec al. (1978a) reported that bone marrov and kidney cell cultures from
rats fed 1Z TNG (approximately 600 mg/kg/day) for 24 months showed no
statistically significant changes in the number of tetraplolds, the frequency of
chromatid breaks and gaps or translocations when compared to untreated controls.
However, biological differences were evident in the number of chromatid breaks ar.d
V — L 3

gaps tn the kidney cultures of treated racs (2.9 i 0.7/50 cells) as compared to
controls (1.6 t 0.5/50 cells)- The authors concluded that since these aberrations
are of questionable relevance with regard to heritable events of Importance to
humans (NRC, 1975 as cited lo Ellis ee al., 1973a), they are not considered
toxicologically important.
Ellis ec al. (1978*) did not detect mutagenic activity in dominant lethal
tests employing Charles River CD male rats that were fed TNG at 0.01Z (3 mg/kg/
day), 0.12 (32 ng/kg/day) or 1Z (363 ag/kg/day) for 13 weeks and subsequently
mated to virgin females of the sane strain. The authors reported that adverse
effects ware not observed on male fertility.
Simon et al. (1977) conducted studies using bacteria (S^ typhimurium strains
TA1535, TA1537, TA1538, TA98 and TA100) with and vlthoue S9 metabolic activation
(added rodent liver homogenate) and yeast (S_. cerevialae D3) to evaluate the
mutagenic activity of TNG at aqueous solubility levels (0.0S to 1.0 al of a
saturated solution) before and after application of chlorine disinfection
techniques. Results were negative both before and after chlorlnatlon.
In contrast, Ellis ee al. (1978b) reported that TOG (1,000 tig/plate)
exhibited weak mutagenic activity in the Salmonella/alerosome plate test against
two of five tester strains (TA-L535 and TA-1537 with and without S9 metabolic
activation* respectively).
Lee et al. (1976) studied the cytogenetic effects of TNG on dogs and rats.
Dogs administered TNG ac 1 mg/kg/day for 4 weeks and 5 ag/kg/day for 9 additional
weeks did not show any apparent numerical or morphological aberrations of
chromosomes in peripheral lymphocytes or kidney cultures. However, the authors
reported that only on* or two dogs were terminated and studied at the end of 4 and
13 weeks. Similarly* In rats an average of 59.0 to 59.3 ag/kg/day of TNG in feed
for 5 weeks and 229.S to 233.8 mg/kg/day for an additional 8 weeks did not cause
any numerical or morphological aberrations of chromosomes in the peripheral
lymphocytes or kidney cultures.
Results of an In vitro single gene autatlon study conducted by Lee et al.
(1976) using wild type Chinese Haa«ear Ovary (CHO-K1) indicated that cells exposed
to 50.0 and 144.ft ug/ml of TNG (concentrations that killed 65Z and 99Z,
respectively* of" the cell population and selected from a single cell survival
curve according to the method of Puck and Itao, 1967), showed no mutants. However,
a positive control* ethyl methaneaulfonace, at a concentration (124 ug/ml) that
killed 85Z of the cell population, induced autants at the frequency of 28.0 x 10
per- mean lethal concentration per cell. This test does not incorporate a
metabolic activating system and, therefore, yields no information on the
mutagenicity of TNG metabolites.

4. Carcinogenicity
Long term studies conducted in dogs and mice gave no evidence of
carcinogenicity; rats. however, exposed co TNG for up co 24 months developed
hepatocellular carcinomas. Additionally, rats fed the middle dose level shoved an
increased incidence of the first stage of the progressive development of hepatic
lesions. Male rats fed TNG at the high dose level showed an Increased incidence
of interstial cell tumors of the testes.
Ellis at al. (1978a) evaluated the incidence of tumor formation in various
organs/systems in several species Including dogs, mice and rata exposed to TNG.
Convincing evidence for oncogenicity vac found only in studies with high-dose
treated rats.
Mala and femala Charles River CD rats (38/sax/dosa) ware exposed to TNG in
tha feed for 24 months at doses of 3.04, 31.5 or 363 mg/kg/day and 3.99, 38.1 or
434 mg/kg/day for males and females, respectively. Following treatment, the
animals vere examined for occurrence of tumors in various organa/systema. At the
low dosa, male (3.04 mg/kg/day) and female (3.99 mg/kg/day) rata vere observed to
have no increased tumor Incidence by comparison to controls. The middle doses,
31.5 and 38.1 mg/kg/day, administered to males and females, raspectlvely, caused
some hepatic lesions (described by the authors as areaa or foci of hepatocellular
alteration, tha flrse stage in tha progressiva development of hepatocellular
carcinoma). Additionally, hepatocellular carcinomas vere reported in 3/33 males
(91) and 2/36 females (6Z) receiving the middle dose level for up to 24 months.
It should be noted that these incidences vere reported based on treatment for 24
months or 24 months plus 1 month recovery period and Included unscheduled deaths.
A breakdown of the Incidence of hepatocellular carcinoma by time of occurrence in
the control, low, middle and high dose groups is shown In Table A2-1, Appendix 2.
Following treatment at tha high dose, (363 and 434 mg/kg/day to males and females,
respectively) TNG ellcleed hepatocellular carcinomas In males (14/29, 48Z) and In
females (11/33, 33Z) compared to 0/32 (0Z) and 0/37 (0Z) respectively, In
untreated controls. The authors noted that tha high doaa la about 3.1 orders of
magnitude greater thaa tha uaual maximum human therapeutic dose (10 mg/human/day
or 0.14 mg/kg/day). It waa also the authors opinion that the uaual dosage route
in human* i* sublingual and only e fraction of that doaa will be immediately
delivered to tha liver while in race the portal circulation carries all the
absorbed TNG to the liver. Thus, the dose delivered to the liver of the high-dose
rats was at least 10,000-fold greater than the maximum dose delivered to human
In addition to the hepatocellular carcinomas, male rats fed the high level of
TNG for up to 24 months displayed a significant increase in the occurrence of
interstitial cell tumors of the testes ae compared to controls. A breakdown in
the incidence of these tumors by time of occurrence In the control and all
treatment levels Is shovn in Table A2-4, Appendix 2. The growth of these tumors
was reported by the authora to produce pressure on the tubules resulting la
atrophy and aspermatogenesis.
v- 1 5

Survival data and the Incidence of che TNG-related lesions, as well as chose
considered pre-carcinogenic lesions for rats necropsied after 24 months feeding
are listed in Tabic V-3. These daca for animals dying during the course of the
feeding study arc shown in Table V-4.
In addition to the increase in hepatocellular carcinomas and testicular
tumors in the high-dose rats, thesa animals shoved a considerable decrease in the
tvo tumor types most commonly associated with death la this strain of rat. Kales
and females shoved a decrease in pituitary adenomas, and females shoved a decrease
In mammary tumors, primarily fibroadenomas. Uhlle tha mechanism of this decrease
vas not known, the authors felt that le contributed, at least In part, to the
increased life span of tha high-dosed rats.
5. Reproductive Effects
Exposure of parental and offspring generations to TNG In tha diet has been
reported to result in impaired fertility in tha high-dose rats (363 and 434
og/kg/day for males and females, respectively) of tha parental generations (F ,
F^) as vail as a reduction in certain parameters in the high-dose litters (F ,
Plb* *2	v*r* considered indirect toxicity attributable to a
testicular tumors and malnutrition.
Ellis at al. (1978a) conducted three-genaratlon raproductlon studies on
TNG-treated rats. Tha initial groups of rats or parental animals (F ; 10 males
and 20 females) vers reported to have been administered TNG In tha diet for 6
months prior to mating at dose levels of 0.01Z, 0.1Z, or 1.0Z (actual Intake
levels vara calculated by the authors to ba 3.04, 31.5 and 363 mg/kg/day,
respectively, for males and 3.99, 38.1 and 434 mg/kg/day, respectively, for
females). It should ba noted, hovever, that an accompanying table Indicated age
at first mating for F^ generation to be 5 months. The test substance vas then
administered, at tha same dose levels, during resultant pregnancies and through
veanlng of thalr F offsprings (2nd litters). Thereafter, TNG vas administered
to salacted F^ oftspring (10 or 12 pairs) during thalr grovth and subsequent
production of a F ganaratlon (second litters). Selected F^ rats (10 or 12
pairs) vera than treated similarly eo F offspring, through vesnlng of their
second littas (7,v)•
The author* reported that following treatment, the mean body velghts of boch
male and femala hlgh-dosa parental rats (363/434 mg/kg/day, raapectlvely) were
significantly all matings. No specific effects on ths fertility of
the F generation vera seen. Hovever, che F generation of high-dose rats had
severely Impaired' fertility. Infertility vas also reported in the F. generation
and vas attributed to severe aspermatogenesis of the high-dose males Taccompanied
by Increased interstitial tissue In the cestes and slgnlflesntly smaller
testicular size). This infertility attributed to males vas confirmed by a third
mating of the high-dose F females with control males resulting in a high
pregnancy ratio (13/14). fhe implantation index and the implant viability index

Table V-3 Survival Data and Significant P'athoLog^ln Rats
Necropsled After 24 Months of TNG Feeding
og/kg/day) 0
Concrol 0.01Z 0.1Z
3.04 31.5
1Z Control 0.01Z 0.1Z	IZ
0	3.99 38.1
No. b/ 4(3)
ac 24 Months
X Hepato- 0
Z Neoplastic 0
Z Hyper-
Z Testi- 25(1)
cular Tumors
9(4) 8(3)
0 ll(l)w 75(6)
9(3) 7(4) 7(4)#/ 18(3)
14(1) 56(10)
14(1) 11(2)
50(2) 67(6) 38(3) 22(2) 28(2) 86(6) 83(15)
25(1) 22(2) 75(6)
Reference: Ellis et al., 1978a.
38/sex/level at start of experiment.
(n) - number of aniaala remaining for recovery study.
All animals died before completion of recovery study.
One animal died before completion of recovery study,
(n) - actual number of aalaals with reported pathology.
V-l 7

Table V-4 Significant Pathology in Racs Dying During TNG Feeding Study
mg/kg/day) 0
Control 0.01X 0.1X	IZ Control 0.01X O.IX	IX
3.04 31.5 363	0 3.99 38.1 434
No. Dying 17
X Hepato- 0
X Neo-
X Hyper-
X Testi-	0
cular Tumors
0 15(2)c/ 50(3)
0 75(3)
24(4) 50(12) 62(8) 60(6) 24(4) 24(5). 61(11) 50(2)
0 8(1) 30(3)
..Reference: Ellis et al., 1978a.
ey38/ssx/levsl at start of experiment.
(n) ¦ actual number of animals with reported pathology.

(which measures implantation losses) were reportedly not significantly different
from untreated control levels.
The authors further reported chat In the offspring, all parameters (liccer
size, live-born index, birch weight, viability and lactacion Indices and weaning
weight) except male/female ratio, were reduced in Che high-dose F litters. Most
measurements were also reduced to some extent in the high-dose F aand F
litters. The authors reported that these adverse effects of the^igh dose of TNG
appeared to be secondary to the poor nutritional status of the dams. A comparison
of animal weight and feed intake is illustrated in Table V-5. It should be noted,
however, that the authors reported a third mating of the high dose F females
(body weight 76Z of controls on Day 0) with control males resulted i& a
reproductive performance comparable to those of untreated controls.
Oketanl et al. (1981b,d) reported on the result of fertility, perinatal and
postnatal studies in SD-Slc rats administered intraperitoneal doses of 0, 1, 10 or
20 mg/kg/day of TNG at various stages of development. In the fertility study
(Oketani et al., 1981b), male rats received the TNG for 63 days prior to mating
while females were dosed for 14 days prior to mating through day 7 of gestation.
No significant effects occurred in the parental animals except for an occasional
and transient convulsion or sedation at the high dose. Body weight and food and
water intakes vera comparable to controls for both sexes throughout. No changes
were evident in fertility indexi numbers of corpora lutaa, implantations and live
or dead fetuses or weight of the placenta and live fetuses. Examination of
fetuses revealed no significant abnormalities or defects.
When these same authors (Oketani et al., 1981d) treated pregnant female
^SD-Slc rats with intraperitoneal dosee of 0, 1, 10 or 20 mg/kg/day of TNG on days
17 of gestation through day 21 of lactation, no algnlflcant effects were noted
except for the occasional and transient conorulalons or sedation at the high dose.
Body weight and food and vater intakes were similar to controls. Administration
of TNG had no affect on gaatatlon period, numbers of Implantations, number and
weight of newborns, percentage of dead pups, delivery race or viability and
lactation indices. No effects were noted on postnatal development, growth or
fertility of the offspring (F^) nor were there any teratogenic effects on the
fetuses (F^) of thaaa animals.
6. Teratogenicity
In teratologlcal studies in rtct. TNG did not produce statistically
significant increases in numbers of adverse effects in soft tissues. Skeletal
anomalies (P < 0.0S) were reported, but were not considered by the authors to be
valuable in assessing teratogenic potential.
Ellis et al. (1978a) reported studying the teratogenic effects of all test
levels of TNG in rats during a three generation study. The design and conduct of
Che study were not fully defined and	ur.s remain as to the time and duraticr
V- 19

Table V-5 Body Weight/Feed Intake During F Gestation^
1 o
W«ighe of F.	Feed Intake of F. Females
Generation ac First Mating	During F,. Gestation
Male (gm) Female (ga)	Females (go)
Control 608 tl5 305 ±4	627 ±51
0.01Zb/ 649 ±21 328 ±5	621 ±27
0.lZc/ 604 ±22 309 ±4	591 ±20
l.0Xd/ 433 ±18 (71Z)-/ 247 ±4 (81Z)	409 ±39 (65Z)
Reference: Ellis et al, 1978a.
,3.04/3.99 mg/kg/day, males and females, respectively.
..31.5/38.1 mg/kg/day, males and females, respectively.
-363/434 mg/kg/dayf males and females, respectively.
Percentage as compared to controls.

of dosing during chls study and che age and status of che females used.
Specifically, the experimental methods, as outlined in Ellis et al. (L978a),
indicate that che Fq females from the reproduction 3tudy were discarded following
weaning of the second litter. The body of this report, however, indicates chat
these females were mated a third time and were subsequently used for the
teratology study. Neither procedure fits che protocol described la Appendix II of
the report of Ellis ec al. (1978a) in which virgin females are treated during the
period of organogenesis, i.e. days 6-15 In the rat.. Nevertheless, the reported
results of chls seudy (the third mating of the females of che F generation)
indicated developmental anomalies In sofc tissue and the skeleton from the high
dose level (434 mg/kg/day). The authors reported chat examination of the soft
tissues of one-half of the fecuses disclosed only a single malformation,
diaphragmatic hernia, in che high-dose group which appeared to be related to che
TNG exposure. However, che incidence of anomalies was not Judged to be
slgnlflcanc by che cwo-sample rank cest. In concrete, examination of the skeleton
of the ocher half of the fetuses revealed a statistically significant (P < O.OS)
increase in the incidence of absent and 'incomplete ossification of the hyoid bone
in the high-dose group when compared to controls. The stsrnebrae, centrum and
bones of che skull were not similarly affected. However, these skeletal anomalies
are considered indicators of delayed development fall inco a category of
malformations considered common and are not normally Indicative of eeracogenlc
Okecanl ec al. (I98la,c) reported on che resulcs of ceracological studies in
Japanese albino rabbles creaced intravenously with 0, 0.5, 1.0 or 4.0 mg/kg/day of
TNG on days 6 - 18 of gestation and in SD-Slc rats treated intraperltoneally with
0, 1, 10 or 20 mg/kg/day of TNG on days 7-17 of gestation. No significant effects
occurred In dams of either species exeepc for transient convulsions ac the high
dose level in both species along with occassional sedation in dama of the
high-dosed rats. No effects were noted on body weights or food and water Intakes,
numbers of corpora lutaa, Implantations and live and dead fetuses or placental
weights. No differences from controls were saen in development or the numbers of
external, visceral or skeletal abnormalities in the fetuses. Administration of
TNG during gestation was neither cerotogenlc nor embryotoxic to either species ac
che doses tested.
7. Other Effects
Other potential effects of TNG exposure chat have been tested include
methemoglobin formation and immunologic responses as indicated by changes in serum
concentrations of Ig£. Dose-related incidences of methemoglobin formation In dogs
have been demonstrated; however, there was no demonstrable effect on IgE formation
in dogs.
a. Methemoglobin Formation
Since TNG could also cause metheaioglobin«mia in humans, Lee et al. (1976)

conducted a scudy in dogs investigate the do9e-response relationship, if any,
associated with mathemoglobln formation, the disappearance of methemoglobin from
the blood, and the effects of repeated TNG administration.
A total of 16 young adult beagle dogs vers divided Into £our groups, each
consisting of 2 males and 2 females. Each group was given TNG in capsules as a
single daily dose at 25, 50, 100 or 200 mg/kg/day for 5 consecutive days. Mo
control dogs were Included in this study. After each daily dosa, blood samples
from all dogs vara taken at 0.5, 1, 2, 4, 8, 16 and 24 hours for tha determination
of mathemoglobln utilizing the method of Dubovaki (1964) and cocal hemoglobin
concentration measured as cyanomethemoglobln according to the standard method
reported by Selegson (1958).
Administration of TNG at 25 or 50 mg/kg/day for 5 eonaecutiva days did not
causa any apparent adverse signs aa indicated by appaaranca and bahavior; 100 or
200 mg/kg/day cauaad cyanosis beginning 2 to 3 houra after doaing and lasting a
"few" hours. In addition, 200 mg/kg/day temporarily dacraasad their physical
activity. All dosaa quickly caused formation of mathamoglobin, and concentrations
raachad peaks in 1 to 4 hours. Tha halght of tha paaka and tha duration of the
mathamoglobin concentrations vara both dosa-relatad. At 50 mg/kg/day, tha
mathaaoglobln concentration as X of total hemoglobin raachad a high of about 15X.
At 100 mg/kg/day( ths hlghast level vas approximately 42Z. Mathamoglobin
disappeared from tha blood In 8 to 24 hours aftar dosing dapandlng upon tha dose.
After dosing with TNG at 25, 50, or 100 mg/kg/day for 5 consecutive days, the
authors concludad that thara appeared to ba no tolerance in tha formation of
mathemoglobin; nor vas thara any accumulative affect on tha formation of
mathemoglobln. Howavar, aftar 200 mg/kg/day of TNG, tha mathamoglobin
concentration on tha second day were considerably higher than thosa on tha first
day, and tha mathamoglobin at 2 hours on tha third day vas highar than that on the
second day. Thay laearpratad this as probably indicating soma cumulative effect
of TNG on tha formation of mathamoglobin. Peak concentrations at this dosa level
reached approximately 30% of total hemoglobin.
b. Immunologic Effects
Lee et ml. (1976) conducted scud lee to assess tha potential lmunologlc
responses to adainistration of TNG using the immunodiffusion technique of Mancinl
(1964). Dogs administered TNG at 0.01 to 5.0 mg/kg/day in tha diet for up to 13
weeks exhibited no significant changes in serum concentrations of IgE antibodies
or sensitization or allergic reactions vhen compared to unaxposad controls.
Similarly, rats administered TNG at 2.5X of tha diet for 13 weeks showed no
apparent alterations in serum concentrations of IgE.

VI. health advisory development
Available toxicity studies in various animal species, for periods ranging
from a single dose to continuous 24-month feeding studies* along with data on the
therapeutic and adverse effects of TNG in humans have been evaluated. Several
toxicity endpolnts considered relevanc to a possible HA development for TNG In
drinking water have been Identified. These endpolnts Include the direct
cardiovascular effects In humans treated with TNG for symptoms of angina
(Needleman and Johnson, 1980). Adverse effects, secondary to the desired
therapeutic effect but occurring at the same dose levels. Include orthostatic
hypotension resulting in dizziness and fainting, headaches, often severe and
prolonged, flushing of the face and neck and rapid pulse race (U.S. Pharmacopeia
D.I., 1980; Needleman and Johnson, 1980). In animals, the significant findings
include metheaoglobln formation In dogs (Lee ae el., 1976), testicular tumors
accompanied by atrophy and aspermatogenesis and hepatocellular carcinomas in TNG
treated rata (Ellis ee al., 1978a). A decrease In ehe race of weight gain in
rats, evident In the early weeks of all feeding studies ae various dose levels, is
felt eo be secondary to a corresponding decrease In food Intake.
The feeding of TNG to mAle and female rats (3/sex) ae an average dose of
approximately 1410 mg/kg/day (2.5Z in the diet) for up eo 13 weeks (Lee ec al.,
1976) eseabllshed a correlation with the decreased welghe gain/ decreased food
lncaka observed In other feeding studies. The mosc significant findings In this
study, however, were the moderate to Bevere testicular aerophy, mild to moderate
testicular degeneration and severe aspermatogenesis of ehe cesees in the treated
males. Similar testicular findings also occurred in raes fed TNG ae approximately
360 mg/kg/day for up eo 24 months. Ellis ee al. (1978a) raporeed ehese effects to
be secondary eo ehe development of the testicular tumors also found in ehe
24-moneh feeding seudy. No eeseicular tumors were reporeed in ehe 13-week study,
nor were ocher TNG-relaead lesions described alehough ehese animals did develop
some degree of heaosiderosls of the spleen and liver. While hemosiderosis of
these organs is reporeed as often seen In rodenes, no such findings were evident
in the unereaeed conerol raes (5/sex) in this seudy.
No eeseicular effeces were reported in dogs fed TNG ae levels up to 25
mg/kg/day for up eo 12 aonchs, nor were any taseicular eumors reporeed to have
occurred in sale mice fed TNG at levels up to 1040 mg/kg/day for up to 24 months
(Ellis ee al., 1978a). The incidence of atrophy of the tesees of treated mice in
this study did noe significantly differ from thac found in conerol mice. A high
mortaliey race was observed, however, in all male mice, conerol and treated. The
lack of sufficient animals for evaluation of eeseicular effeces afcer 24 months
feeding wleh TNG somewhae limits the usefulness of this study. It appears,
therefore, thae raes are ehe most sensitive species for ehe evaluaelon of
testicular effeces due eo treatment with TNG.
Rats fed TNG ae levels up to an average of approximately 400 mg/kg/day for up
to 24 months developed hepatocellular carcinomas as a result of the TNG feeding.
vi -1

Tumors were reported by Ellis et al. (1978a) ac both the middle dose (35
mg/kg/day) and high dose levels (400 mg/kg/day); racs ac the lower (3.5 mg/kg/day)
and middle dosa levels also shoved changes described as hyperplastic foci or areas
of altered hepatocytes, changes considered preliminary to che development of
eumors. These altered hepatocytes progressed to neoplastic nodules at the
aid-dose level before developing into hepatocellular carcinomas. All three stages
were evident as early as twelve months into the study. Cholangiofibroais,
proliferation of the bile ducts and fibrous tissue, was also reported as early as
12 months in the study and occurred at the high dose level. No related findings
were reported In dogs or mice at any treatment level.
Methemoglobln formation was observed in studies In boeh dogs and rats with
dogs appearing mora sensitive to this effect. In an actempt to correlate
methemoglobln formation to TNG intake, dogs were dosed by capsule vlth 23, 50, 100
or 200 mg/kg/day of TNG for up to 5 days (Lee et al., 1976). All dogs readily
produced methemoglobln but only ths two higher doses produced significantly
elevated levels, 421 and SOS, respectively, along with signs of cyanosis (at both
levels) and decreased physical activity at ths highest dosage lavel. While no
control animals were concurrently analyzed in this study, it can be assumed that
statistically significant levels of methemoglobln would not be produced In the
untreated animal. Dogs fed TNG for up to 12 months at 1, 5 and 25 mg/kg/day
shoved no toxic signs axcept for a slight Increase in methemoglobln formation
amounting to an average of less that 2Z (Ellis ec al., 1978a). While
methemoglobinemia occurred after 9 months of feeding and appeared to be
dose-related, the low level of methemoglobln (less than 2Z) and the specific time
of occurrence (neither before nor after the 9 month test period) gives little
value to these findings. Furthermore, dosage levels la this study did not appear
to be high enough to elicit any significant or observable toxic signs in this
species and further treatment was terminated after 12 months of feeding.
Rats and mice fed TNG for up to 24 months at doses up to 400 and 1040
mg/kg/day, respectively, also shoved evidence of methemoglobln formation. Ia
rats, Ellis et al. (1978a) reported levels of approximately 20Z in males and
females fed an averaga of 400 mg/kg/day. This lavel of methemoglobln formation
was evident afear 3 months of feeding end continued through 6 months on the test
diet. Thereafter, methemoglobln levels varied considerably, from approximately IZ
to 341, vlth lass than 1Z methemoglobln detected after 24 months of feeding ac che
high dose laval. These same authors reported that methemoglobinemia vlth
anemia, Helns bodies and pigment deposits also occurred at the high dose level
(1040 mg/kg/day) in mica. Hovever, average methemoglobln values reported over che
24-month study nevar exceeded the level found in untreated controls (3.6Z). A
marked species difference In response co aethemoglobln formation has been reporced
by Beard and Noa (1981) vlth man generally appearing more sensitive to this effect
as compared to dogs, and rats being che '.«ast sensitive. These studies appear co
substantiate that report regarding the relative sensitivity of dogs and rats.
While it is knovn that TNG has a pharmacological effect on the cardiovascular

(CV) system in humans, none of the referenced animal studies specifically examined
the CV effects of TNG feeding in animals. The only available histopathological
effect reported la these studies was an apparent dose related decrease in the
incidence of ayocardial degeneration in the mature (older than 12 months) rat fed
TNG at levels up to an average of approximately 400 mg/kg/day for up to 24 months
(Table A3-1, Appendix 3). This effect vas particularly noticeable in the female.
The occurrence of this lesion vas reported to begin around 18 months in the study
with only one reported occurrence before .12 months. The significance of this
effect, if any, is not known.
In order to determine which of the species studied might most closely
parallel the effects of TNG feeding in the human, the pharmacokinetics of TNG in
the various species vas examined. The only available comparison vas an in vitro
study using livers from various species, including human livers. Data from this
study indicated	like the dog, rabbit and monkey, human livers, in vitro,
metabolized 1,3- C-TNG primarily to 1,2-DNG as compared to 1,3-DNG while racs
and mice produced higher levels of 1,3-DNG. Additionally, livers from mice and
humans had a lover ability to metabolize TNG in vitro, as evidenced by a lover
recovery of labelled metabolites when compared to other species. This Information
provides only a qualitative comparison and Is not sufficient to determine vhlch
species most closely resembled man in its metabolism of TNG.
A. Quantification of Toxlcological Effects
Health Advisories (HAs) are based upon the Identification of adverse health
effects associated with the most sensitive and meaningful noncarcinogenlc endpoint
of toxicity. The induction of this effect is related to a particular exposure
level over a specified period of time and is most often, determined from the
results of experimental animal studies. The general formula is as follows:
(UF(s)) ( L/day) " 	mg/L
NOAEL • No-Obeerved-Adverse-Eff«ce Level
LOAEL ¦ Lowest-Observed-Adver»«-£ffects Level
(exposure dose in mg/kg/day)
BW * assumed body weight of protected individual
(10 kg for child or 70 kg for adult)
UF(s) • uncertainty factors(s), based zn quality and nature of data
V I - 3

L/day " assumed water consumption
(1 L/day for child or 2 L/day for adult)
B. One-day Health Advisory
No daca in animals were located In the available literature that were
considered suitable for ths derivation of a One-day HA. Available data on the
adverse effects of therapeutic doses of TNG administered sublingually and orally
in humans In the treatment of acute anginal attacks was felt to be adequate for
this calculation. These effects are primarily orthostatic hypotension and
The therapeutic dose range for sublingual TNG la 0.15 to 0.60 mg Initially,
with up to 10 mg/day being the prescribing limit and 0.3 mg often enough to
produce tha desired pharmacologic effect (Needleman and Johnson, 1980). It has
been reported that TNG Is readily absorbed after sublingual administration with
pharmacological effects evident In 1 to 3 minutes and remaining effective for 20
to 30 minutes* occasionally for up to one hour (PDR, 1985).
Relaxation of vascular smooth muscle Is tha principal pharmacologic action.
It may also reduce systolic, diastolic and mean arterial blood pressure, slightly
increase tha heart rata and, at doses not effeceiva on syseaaic arterial pressure,
may produce arteriolar dilatation In the face and neck, resulting in a flush.
This same dose may also cause headache, often severe, presumably due to dilatation
of the meningeal arterial vessels. A marked hypotensive effect may also follow
sublingual administration resulting in transient episodes of dizziness, weakness
and other manifestations of cerebral ischemia (Needleman and Johnson, 1980).
In vlev of the foregoing Information, development of a human no effect level
based on the therapeutic dose is appropriate and would provide a protective level
for sensitive members of tha population. Trlnitroglycarln may also be
administered orally; however, this dosage form is unreliable for therapeutic
purposes since there is variable absorption from tha allmentrary tract (U.S.
Pharmacopeia D.I., 1980). In this form, 10-fold higher doses are necessary to
produce tha mm pharmacological effects. This indicates that absorption of TNG
by the gaserolacaatinal tract is probably significantly less than via the
sublingual rout*.
Assuming tha average weight of the subject was 70 kg, the usual dosage would
range from approximately 0.002 to 0.009 mg/kg, with an arithmetic mean of
approximately 0.005 mg/kg. Using an uncertainty factor of 10 would reduce the
dose to well below the level causing therapeutic effects and provide additional
protection for sensitive Individuals. The HA value for a 10 kg child consuming l
liter of water is as follows:
°-°°5	* 10 - 0.005 og/L
10 x 1 L

0.005 mg/kg " mean therapeutic dose
10 kg * assumed body weight of a child
10 - uncertainty factor, chosen In accordance vlch
NAS/0DW guidelines for human data. Reduces che mean
therapeutic dose to a no effect level.
1 I ¦ assumed dally water consumption of a child
A four-week feeding study in rats, conducted by Lee et al. (1976), sight seem
the most appropriate for determining a 10-day HA. In this study, rata were fed
TNG in the diet at levels ranging from 0.0012 to 0.12 (TNG intake vas calculated
to be an average of 0.85, 6.2 or 59.2 mg/kg/day at the three treatment levels).
However, the only observable effect In this study was a slight decrease in the
rate of'weight gain at the high dose which appeared to be directly correlated with
a parallel decrease in food consumption. Similar studies conducted by these same
authors la dogs and nice also produced no observable toxic effects, and feeding
levels were raised 5-fold in all studies for the remaining weeks of a 13-week
A five-day dosing study conducted by Lee et al. (1976) in dogs la not
appropriate for determination of a 10-day HA as the only parameter of toxicity
examined was methemoglobin formation which is not the most sensitive observable
effect. Therefore, none of the foregoing studies are deemed appropriate for
development of a 10-day HA. It is, therefore, recommended that the One-day HA
value of 0.005 mg/L be used aa an estimate of the 10-day HA since this would be
most protective of the more sensitive members of the population.
D.	Longer-term Health Advisory
Thirteen-week feeding studies conducted by Lee et al. (1976) in dogs, rats
and mica ae levels up to 5, 230 and 590 mg/kg/day, respectively, produced no
observable health effects useful in determining HA values. The extramedullar
hematopoieai* of the liver and spleen of all male and female mice fed TNG at
approximately 60 and 580 mg/kg/day and la one male and one female at approximateLy
6.6 mg/kg/day was mild and of questionable significance. A similar feeding study
by these same authors in which rata were fed TNG at 1410 mg/kg/day for 13-week.s
produced severe testicular damage but did not provide sufficient data for
determining a no effect level. The thirteen-week study in dogs produced no
adverse effects la animals exposed to che highest (5 mg/kg/day) dose.
Data on the health effects of longer-term exposure in humans to TNG are

available- Industrial workers exposed co TNG initially experience severe
headache, dizziness and postural weakness during the first several days of
exposure. Tolerance then develops, but symptoms may reappear after a few days
away from the source of exposure. The most serious effect of chronic exposure is
a form of organic nitrate dependence, and individuals with no demonstrated organic
vascular disease have died suddenly or developed myocardial Infarction after a few
days break In chronic exposure. Additionally, there are now well-documented cases
of severe myocardial Ischemia, relieved by TNG, during withdrawal from chronic
exposure to an organic nitrate (Needleman and Johnson, 1980). Taking into
consideration this added risk due to the development of tolerance to the adverse
effects of TNG, it seem* appropriate to use the One-day HA value of 0.00S og/L for
longer-term exposure.
E.	Life-time Exposure
Lifetime experiments conducted by Ellis et al. (1978a) in rats exposed to
levels of TNG ranging from 3.0 to 430 mg/kg/day shoved evidence of development of
TNG-relaeed lesions In males and females at th« two higher dose levels. These
lesions Included hepatocellular carcinomas at the middle (33 mg/kg/day) and high
dose level (400 mg/kg/day) and Interstitial cell tumors of the testes in males fed
TNG ae 360 mg/kg/day. These data Indicated a NOAEL of 3.S mg/kg/day over a
lifetime. Non-carcinogenic effects at the middle dose level Included hyperplastic
foci In the liver of both males and females. These effeces indicate a NOAEL of
3.5 mg/kg/day for noncarclnogenlc effects. Lifetime experiments conducted by
Ellis et al., 1978a in male and female mice exposed to levels of TNG ranging from
9.7 to 1060 mg/kg/day resulted In decreased velghc gain and methemoglobinemia at
the high dose level only. However, based on the development of tolerance in
humans chronically exposed to TNG, as described under the Longer-term HA, the use
of the One-day HA value of 0.003 mg/L Is recommended.
F.	Evaluation of Carcinogenic Potential
Even though TNG haa been extensively used as a drug, there is no reported
evidence of carcinogenicley in humans. It haa been tested in lifetime studies in
dogs, rats and mica. Only ehe studies in rats were positive and details are
provided belov. Alehough a cancer risk assessment is provided based upon the
study by Elll* et al. (1978a) TNG Is not considered a probable human carcinogen
based upon eh* weight of evidence of che existing data. The EPA does not usually
provide quantleaelve risk assessment for chemicals with limited evidence of
carcinogenicity in animals. However, in che lnteresc of a complete evaluation of
the rat study, a quantitative risk assessment is provided.
In the lifetime study conducted by Ellis et al. (1978a), male and female CD
rats exhibited a significant dose-related increase in hapatocelullar carcinoma
when exposed to doses of TNG averaging 3 5 or 400 mg/kg/day in the diet for up co
24 months* Additionally, male CD rats exposed at approximately 363 mg/kg/day in
the dleC for the same time period developed a significantly Increased incidence of
interstitial cell tumors of the testes.

Xhe incidence of hepatocellular carcinoma used co calculate the carcinogenic
risk assessment are 0/32, 0/3^, 3/33 and 14/29 for aale racs at control, low,
middle and high doses, respectively. The multistage model was used for
high-to-low dose extrapolation (Crump and Watson, 1979; Hove and Crump, 1982).
GLOBAL83 vas used to fit the data in the experimental dose range and to obtain
upper 95Z confidence limits on the combined incidence of hepatocellular
carcinomas. The multistage model conforms to a biological model of tumor
initiation and promotion (Crump et al., 1977) and provided an adequate fit to the
dose-response data for TNG.
Animal doses were converted to equivalent human exposures using a surface
area correction assuming a 0^35 kg rat* an^ a 70 kg human. The human slope
estimate (q^*) is 1.66 x 10~ (mg/kg/day) for the multistage model. The slope,
q^*t is taken as an upper bound of the potency of the chemical to Induce cancer at
lov doses below the experimental dose range. The q * may be used to estimate the
upper 95Z confidence llmle on extra risk by q^* x Dose, In units of mg/kg/day.
The 9SZ lower confidence limit on the dose producing an extra risk* R, may be
estimated by R/q^*. Assuming that a 70 kg human adult consumes 2 liters of water
a day over a 70 year lifespan the estimated cancer risk is as follows:
Level of	TNG in Drinking Water
Human Risk	mg/L	ug/L
10"4	0.21	210
10~5	0.021	21
10"6	0.0021	2.1
For comparison purposes, drinking water concentrations associated with an
excess cancer risk of 10~ were 2.0 ug/L, 121 ug/L, 0.4 ug/L and 0.1 ug/L for the
one-hit, probit, logie and tfeibull models, respectively.

VII. analysis
Several methods have been published for nitroglycerin analysis. Two of chase
methods appear co be more useful for che determination of low concentrations
(below 10 ppm) and these procedures are described.
The following methods for determining TNG in explosive and pharmaceutical
preparations have been listed by Rosenblatt et al. (1973) as cited in Sullivan et
al. (1979). In general, these methods were not reported Co be useful for low
concentrations of TNG in aqueous solutions.
1.	The Du Pone nitrometer method which measures nitric oxide gas liberated
from the aster by mercury.
2.	Hydrolysis of the ester followed by reduction of the NO - and analysis
of N02-.	3
3.	Reduction by tltanous chloride and back titration of the excess reagent
with feerie alum.
4.	Reduction with ferrous chloride and titration of Che ferric iron with
tltanous chloride.
5.	Infrared detection of nitrate groups.
6.	Ferrous sulfate-sulfuric acid colorlaetrlc method.
Two instrumental methods exist which are capable of quantitatlng TNG in
environmental samples ae concentrations below 10 ppm. Walsh (1976) and Veltzel et
al. (1976) as cltsd in Sullivan et al. (1979) described high pressure liquid
chromatographic separation followed by UV detection of the nitrate ester groups.
Gas-liquid chromatography (GLC) using electron captura has also been successfully
employed (Stilwell et al., 1976 and Hackley et al.* 1974 as cited in Sullivan et
al., 1979). Both techniques involve solvent extraction of the TNG from water or
sediment with benxene, ethyl acetate, or methylene chloride• followed by gel
cleanup and concentration of the excrace, if required. Because TNG is extremely
thermolabile, gantla drying at iaBtanc temperatures Is required. Ultra-violet
detection Is not specific for TNG but can be used to identify peaks of TNG or ics
less completely nitrated analogs. One advantage of ultra-violet detection is chat
it is done at ambient temperatures. Weltzel et al. (1976) as cited in Sullivan ec
al. (1979) reported detection limlcs of 0.002 mg/L in water and 1 mg/kg In
sediments from the New River near Radford, Virginia, site of a manufacturing
plant. Walsh (1976) as cited in Sullivan ec al. (1979) was able to detect as
little as 0.02 mg/L (100 ppm gave a full-scale response in a 25 ul sample).
The high column and detector temperature# required for GLC cause breakdown
and loss of sensitivity to very low levels of TNG; however, sensitive detection Ls

feasible if column temperatures are maintained at LOO® C to 1<»Q® C. Rosseel and
Bogaerc C L973) as cited in Sullivan ec al. ( 19 79) were able to detect
concentration* of 0.5 ug/L la plasma using electron capture techniques. Stilvell
et al. (1976) as cited in Sullivan ec al. (1979) reported 100Z recovery of TNG
standards at 1 mg/L but only 28Z recovery at 0.L ag/L. These authors reported
chae detection with GLC electron capture was possible Co levels as low as 0.01
mg/L la water and 0.05 og/kg in sediment but chac reproducibility was poor below
0.6 mg/L and 1.5 mg/kg for water and sediment, respectively. Hackley ec al.
(1974) as cited in Sullivan ec al. (1979) effectively used flame Ionization
decectlon as well as electron capture co quantitate TNG and related compounds at
concentrations of 10-1000 ag/L.

VIII. treatment
The EPA (1976) as clced la Sullivan «c al. (1979) has reviewed TNG waste
treatment processes and stated that vaatevater with up to 1,200 mg/L TNG is
amenable to treatment by the activated sludge process after addition of lime to
reduce the explosion hazard.
Bogatko (1978) as cited In Sullivan et al. (1979) studied the blodegradation
of nltraes esters In activated sludge systems. Bogatko's results suggest that
microbial activity does not completely degrade TNG since only a 40 to 45Z
reduction in 30 mg/L concentrations of mixtures of TNG and ethylene glycol
dlnltrate could be attained.
The Incomplete biotransformation of TNG by bacteria was also studied by Walsh
(1976) as cited in Sullivsn et al. (1979) who showed that TNG at concentrations
above 75 mg/L was toxic to microflora of activated sludge. Ae a concentration of
70 mg/L bacterial cultures were able to degrade approximately 50Z of the TNG.
Characterization of TNG wastewaters and studies of the transformation
products formed during microbial or chemical wastewater treatment suggest that all
of the mono- and dinltrated Isomers of glycerol as well as TNG can be present. In
addltiont slda reactions apparently occur which result in transformation products
other than inorganic nitrites and nitrates, CO and glycerol (Capellos et al.,
1978; Fraser, 1968; Hackley et al., 1974; and welsh, 1976 as cited in Sullivan et
al., 1979).
Fraser (1968) as cited In Sullivan et al. (1979) studied the hydrolysis of
„JNG, 1,3-DNG, and 1' and 2-MNG In ethanol solutions of NaOH and found evidence of
ooth alpha and beta elimination followed second order reactions. The relative
ease of hydrolysis was: TNG > 1,3-DNG > 1-MNG > 2-MNG In order of decreasing
reaction rate. Basic hydrolysis using Ca(OB) has been proposed as a treatment
method for TNG wastavatar; therefore, the kinetics of basic hydrolysis of TNG and
the isomeric DNG's in aqueous solution have been studied by the U.S. Army Medical
Research and Development Command. This research has been conducted since July
1976 (Capellos et al., 1978 as cited In Sullivan et al., 1979), and has resulted
in the isolation of soma of the chemical breakdown products. In general,
hydrolysis reactions followed second order kinetics in aqueous Ca(OH) solutions.
The relatlva ease of hydrolysis In water vae different than in alcohol. Based on
the second ordar decay constant, the 1,3-DNG Isomer was the most reactive material
followed by 1,2-DNG, than TNG (Capellos et al., 1978 as cited In Sullivan et al.,
Waste treatment at a certain TNG manufacturing plant has consisted of cacch
basins to trap sedimented and precipitated TNG. Waste streams at this plant are
neutralized before collection (Patterson et al., 1976 as cited In Sullivan et al..
1979). Proposed treatment of these waters viil include Ca^H)^ hydrolysis as
discussed above, followed by an activated sludge process employing a rotating

biological contactor. Treatment of.wastes high In TNG Is still in the
experimental stage (EPA, L976 as cited in Sullivan et al., 1979) and reported
effectiveness of biological waste treatment to completely degrade TNG remains
somewhat questionable.

Based on the available animal data and the effects attributed to therapeutic
doses of TNG vhen administered In the treatment of acute angina or when humans are
exposed in the industrial setting, the HA for One-day, 10-day, Longer-term and
Lifetime exposure has been determined to be 0.005 mg/L. Based on a study In rats
In vhlch TNG exposure resulted in hepatocellular carcinomas* quantitative risk
estimates are also provided.
A companion report, "Data Deficiencies/Problem Areas and Recommendations for
Additional Data Base Development for Trinltroglycerol" (Appendix 4), summarizes
the scope of existing data reviewed for this HA. This report delineates those
areas where additional data and/or a clarification of existing data would be
advantageous for a sound HA.
IX- 1

X. references
ACGIH, 1971. American Conference of Governmental Induscrial Hygienists, Committee
on Threshold Limit Values: Documetation of the threshold limit values for
substances in workroom air. 3rd edition, 2nd printing, 1974, pp. 110-111. As
cited in Shiotsuka (1979).
Bartalini, E., et al., 1967. Epidemiological and clinical features of
occupational nitroglycol poisoning in Italy. Med. Lav. 58:618-623 as cited in
NIOSH, 1978.
Beard, R.R. and No«, J.T., 1982. Patty's Industrial Hygiene and Toxicology, 3rd
ed., Vol. 2A» John Wiley & Sons, New York, p. 2416.
Bogatko, (1978) cited In Sullivan et al., 1979. Primary reference was not given.
Brassier, R.R., 1966. Nitroglycerin reactions among pharmaceutical workers. Ind.
Med. 18:519-523.
Capellos, C. et al., 1978. Kinetic studies and produce characterization during
the basic hydrolysis of glyceryl nitrate esters, Technical Repore. AR
LCD-TR-79022, AD E 400341, Energetic Materials Division* Large Caliber Weapons
System Laboratory, U.S. Army Armament Research and Development Command, Dover,
NJ. As cited in Sullivan et al. (1979).
Carmichael, P. and Lieben, J., 1963. Sudden death In explosives workers. Arch.
Environ. Health 7:424-439.
Crump, K.S., et al., 1977. Confidence Intervals and tests of hypothesis inferred
from animal carcinogenicity data. Biometrics, 33:437-451.
Crump, K.S. and Watson, W.V., 1979. GLOBAL79. A Fortran program to extrapolate
dlchotomous animal carcinogenicity to lov dose. Prepared for The Office of
Carcinogen Standards, OSHA, U.S. Department of Labor, Contract 41USC252C3.
Dacre, J.C. and Rosenblatt, D.H. 1974. Mammalian toxicology and toxicity co
aquatic orgaalna of four Important types of waterborae pollutants - An extensive
literature evaluation. Technical Report 7403. U.S. Army Medical Bioengineerlng
Research and Development(Laboratory. Aberdeen Proving Ground, MD.
Dubowski, K.M., 1964. Measurement of hemoglobin derivatives. Hemoglobin Ics
Precursors and Metabolites, J.B. Lippincott Company, Philadelphia, PA.
Ellis, H.V., III et al., 1978a. Mammalian toxicity of munitions compounds
Phase III: Effects of life-time exposure Part II: Trlnitroglycerln. Progress
Report No. 8. Midwest Research Insticuce. Kansas City, MO, Contract No.
DAMD-17-74-C-4073. AD A078746.

Ellis, H.V.t III et al., 1978b. Mammalian toxicity of munitions compounds Phase
I: Acuce oral toxicity, primary skin and eye irritation, dermal sensitization,
disposition and metabolism, and Ames tests of additional compounds. Report No. 6.
Midwest Research Institute, Kansas City, MO, Contract No. DAMD-17-74-C-4073, AD
Environmental Protection Agency (EPA), 1976. Development document for interim and
final effluent limitation guidelines and proposed new source performance standards
for the explosives manufacturing point source category. EFA 440/1-76-060-J. As
cited In Sullivan et al. (1979).
Feldman, R.L. et al., 1978. Unusual vasomotor coronary arterial responses after
nitroglycerin. Aa. J. Cardiol. 42:517-519.
Finney, D.J., 1971. Problt Analysis, Cambridge University Press.
Fraser, R.T.M., 1968. Stability of nitrate esters. Chea. & Ind. 33:1117-1118.
As cited in Sullivan et al. (1979).
Gross, E. et al., 1960. [Absorption of glyceryl trinitrate (nitroglycerin)
through the skin.] Arch. Toxicol. 18:331-334 (Ger) as cited in NIOSH, 1978.
Hackley, B.E., Jr. et al., 1974. Environmental quality standards research on
wastewaters of Army Ammunition Plants. Edgewood Arsenal Technical Memorandum
EC-TM-74004, Aberdeen Proving Ground, MD. As cited in Sullivan et al. (1979).
Hanlon, J.J. and Fredrick, W.G., 1966. Great lead controversy. Arch. Environ.
Health (lett) 12:676.
Hawley, G.G., 1977. The Condensed Chemical Dictionary. 9th ed. Van
Hodgson, J.R. and Laa( C.C., 1975. Trinitroglycerol metabolism: Denltratlon and
glucuronlde formation in the rat. Toxicol. Appl. Pharmacol. 34; 449-455.
Hogstedt, C. and Andersson, K., 1979. A cohort study on mortality among dynamite
workers. J. Occup. Mad. 21:553-556.
Howe, R.B. and Crump, K.S., 1982. GI.CBAL82: A computer program to extrapolate
quantal animal toxicity data to low doses. Prepared for The Office of Carcinogen
Standards, OSHA, U.S. Department of Labor, Contract 41USC252C3.
Jacob, J.C. and Maroun, F.B., 1969. Peripheral neuropathy in a person sensitive
to dynamite. Can. Med. Assoc. J. 101:102-104. as cited in NIOSH, 1978.
Klock, J.C., 1975. Nonoccluslve coronary disease after chronic exposure to
nitrates: Evidence for physiologic nlcra:e dependence. Am. Heart J. 89:510-513.
¦<. i

Lange, R.L. et al., 1972. Nonatheronatous ischemic heart disease following
withdrawal from chronic industrial nitroglycerin exposure. Circulation 46:
Lee,- C.C., 1987. Personal communication.
Lee, C.C. et al., 1975. Masnalian toxicity of munition compounds: Phase I.
Acute oral toxicity, primary skin and eye irritation, dermal sensitization, and
disposition and metabolism. Report No. 1. Midwest Research Institute, Kansas
City, MO, Contract Ho. DAMD-17-74-C-4073, AD B011150L.
Lee, C.C. et al., 1976. Mammalian toxicity of munition compounds Phase II:
Effects of multiple doses Part I: Trinitroglycerin. Report No. 2. Midwest
Research Institute, Kansas City, MO, Contract No. DAMD-17-74-C-4073, AD A047067.
Lund, R.P. et al., 1968. Withdrawal symptoms in workers exposed to
nitroglycerine. Brit. J. Industr. Med. 25:136-138.
Maccherini, I. and Camarri, E., 1959. [Nltroglycol poisoning.] Med. Lav.
50:193-201 (Ita) as cited in NIOSH, 1978.
Magnusson, B. and Klingman, A.M., 1969. The identification of contact allergens
by animal assay. The guinea pig maximization test. J. Invest. Derm. 52:268-276.
As cited in Lee et al. (1975).
Mancini, G. et al., 1964. Immunochemical quantitation of antigens by single
radial immunodiffusion. Isnunochemlstry, 2:235.
Mark, H.F. et al., 1965. Klrk-Othmer Encyclopedia of Chemical Technology. 2nd
ed. Vol. 4. Intersclence Inc. As cited in Sullivan et al. (1979).
Merck and Co., Inc., 1976. The Merck Index. An encyclopedia of chemicals and
drugs. 9th ed. Merck and Co., Inc., Rahway, NJ.
Munch, J.C. and Frledland, B. 1963. Glycerol trinitrate. I. Acute toxicity.
Ind. Med. Surg* 34: 143-145* As cited in Dacre and Rosenblatt (1974).
Needleman P.* and Johnson, Jr., E.M., 1980. The Pharmacological Basis of
Therapeutics* 6th ed. Chapt. 33, MacMillan Publishing Co. Inc., New York.
Needleman, P. and Krantz, J.C., 1965. The biotransformation of nitroglycerin.
Biochem. Pharmacol., 14:1225.
NIOSH, 1978. National Institute of Occupational Safety and Health. Criteria for
a recommended standard..occupational exposure to nitroglycerin and ethylene glycoL
dlnltrate. DHEW (NIOSH) Publication No. '3-167.

N'RC, 1975 - National Research Council. Committee for the Working Conference on
Principles of Protocols for Evaluating Chemicals in the Environment. Principles
for Evaluating Chemicals in the Environment. Washington, D.C. National Academy
of Sciences.
Patterson, J.U. et al., 1976. State-of-the-art: Military explosives and
propellants production industry. Vol. II. EPA 600/2-76-213C. As cited in
Sullivan et al. (1979).
PDR, 1985. Physicians Desk Reference. 39th ed. Medical Economics Co., Inc.,
Oradell, N.J. p. 1546.
Puck, T.T. and Kao, F.T., 1967. Genetics of somatic mammalian cells. V.
Treatment with 5-bromodeoxyuridine and visible light for Isolation of
nutritionally deficient mutants. Proc. Nat. Acad. Set. 58:1227.
Oketani, Y. et al., 1981a. [Toxlcological studies on nitroglycerin (NK- 843) (6)
Teratologlcal study in rabbits.] Pharmacometries. 22(5):633—638 (Jap).
Oketani, Y. et al., 1981b. [Toxlcological studies on nitroglycerin (NK-843) (7)
Fertility study in rats.] Pharmacometries 22(5):639=648 (Jap).
Oketani, Y. et al., 1981c. [Toxlcological studies on nitroglycerin (NK-843) (8)
Teratologlcal study In rats.] Pharmacometries. 22(6):737-751 (Jap).
Oketani, Y. et al., 198ld. [Toxlcological studies on nitroglycerin (NK-643) (9)
Perinatal and postnatal study in rats.] Pharmacometries. 22(6):753-763 (Jap).
Planques, J. et al.. 1959. [Cutaneous and nervous allergic reactions induced by
dynamite.] Arch. Mai. Prof. Med. Trav. Secur. Soc. 20:187-188 (Fre) as cited in
NIOSH, 1978.
Prerovska, I. and Teislnger, J., 1965. Clinical picture of chronic intoxication
with dinitrodiglycol. Prac. Lek. 17:41-43 as cited in NIOSH, 1978.
Rosenblatt, D,B. et al.. 1973. Munitions production products of potential concern
as vaterboraa pollutants-Phase I. U.S. Army Medical Environmental Engineering
Research Unit Report No. 73-07, AD 912752 Edgevood Arsenal, MD.
Roaseel, M.T. and Bogaert, M.G., 1973. GLC determination of nitroglycerin and
isosorbide dinltrate in human plasma. J. Pharm. Scl. 62:5. As cited in Sullivan
ee al. (1979).
RTECS, 1980. Registry of Toxic Effects of Chemical Substances. National
Institute of Occupational Safety and Health (NIOSH).

Rubino, G.F. ec al., 1956. [Effects of chronic exposure co mtroglycol in
dynamite manufacture.] Minerva Med. 47:1124-1128 (lea) as cited in NIOSH, 1978.
Schwartz, A.M., 1946. The cause, relief and prevention of headaches arising from
contact with dynamite. N. Eng. J. Med. 235:541-544.
Selegson, D., 1958. Standard Methods of Clinical Chemistry, Academic Press, Inc.,
N.Y. Vol. 2:52.
Shiotsuka, R.N. 1979. Occupational health hazards of nitroglycerin with special
emphasis on tolerance and withdrawal effects — a literature review. Technical
Report 7903. U.S. Army Bloengineerlng Research and Development Laboratory, Fort
Detrlck, MD.
Simmon, V.F. et al., 1977. Mutagenicity of some munition wastewater chemicals and
chlorine test kit reagents. Final Report. SRI International, Menlo Park, CA,
Contract No. DAMD-17-76-C-6013, AD A057680.
Smith, J.G., 1986. Water quality criteria of nitroglycerin. Oak Ridge National
Laboratoy. Final Report. AD 0RNL-6180. Oak Ridge, TN.
Stllwell, J.M. et al., 1976. Aquatic life field studies at Badger Army Ammunition
Plant. Final Report. Phase II. U.S. Army Medical Research and Development
Command Contract No. DAMD 17-74-C-4123, Battelle Columbus Laboratories,
Columbus, OH. As cited In Sullivan et al. (1979).
Sullivan, J.H. et al., 1979. A summary and evaluation-of aquatic environmental
data in relation to establishing water quality criteria for munitions-unique
compounds. Part 2: Nitroglycerin. Final Report. U.S. Army Medical Research and
Development Command, Contract No. DAMD 17-77-C-7027, Water and Air Research, Inc.,
Gainesville, FL.
U.S. Pharmacopeia Dispensing Information (D.I.), 1980. U.S. Pharmacopeial
Convention Inc., Mack Publishing, Co., Easton, PA, pp. 334-336.
Walsh, J.T., 1976. Chemical characterization of nitroglycerin blodegradatlon
products. U.S. Army Natlck Research aad Development Laboratory. Unpublished
manuscript. Natlck, MA. Aa cited in Sullivan et al. (1979).
Weast, R.C., 1971. Handbook of Chemistry and Physics. 51st ed. The Chemical
Rubber Co., Cleveland, OH.
Wester, R.C. et al., 1983. Pharmacokinetics and bioavailability of intravenous
and topical nitroglycerin in the Rhesus monkey: Estimate of percutaneous
first-pass metabolism. J. Pharm Scl. 72:745-748.

w«ici«l, S.L. et al., l976. Aquatic field surveys at Iowa, Radford, and Jollet
Aray Ammunition Plant,. final Report. Volu«. II. Radford Ar.y Munition PUn,
Arny f ® Research and Development Command Coneraet No. DAMD 17-75-c 5046
Environed Control Technology Corp.. Ann Arbor. MI. As ctc.d in Sullivan IT
AX • C 197y/ •


Table Al-I Indices of Toxicity in Hale RaCs Fed TNC for Various Periods of Time8^

°-,X h!
31.5 ag/kg/day
0.1Z -
59 Bg/kg/day
230 ag/kg/day
1406 ag/kg/day
Index of Toxicity
12 BO.
24 mo.
4 weeks
13 weeks
12 B0
i. 24 ao.
4 wk. 13 wk.
Body Weight as Z
of Control
51* 55X
Food Intake as Z
of Control
44Z 72Z
c /
262 7
Liver Lesions
CareI noma
Neoplastic Nodules
Hyperplastic Foci
3 /4
0/ 3
3/4 <2/4)d/

Table A1 — I - continued

31.5 ag/kg/day
o.ix -
59 mg/kg/day
230 ag/kg/day
363 ng/kg/day
1406 ag/kg/day
Index of Toxicity
12 m. 24 »o.
4 weeks
13 weeks
12 m. 24 so.
4 wk. 13 wk.
Atrophy and/or
Interstitial Cell
0/4 1/29
0/4 12/25
"(lit 11 irm c : tlllb ct a 1 . , If) J Ba •
,'M< Imak* corrected for evaporation loss and Intake patterns.
1 jiai lnr.lcally significant as compared to controls.
d^Anlaald fed for 13 weeks with TNG and allowed to recover for 4 weeks.

Table AI-2 Indices of Toxicity in Female Rats Fed TNG for Various Periods of Tlme3^

0.11 .
0.1Z -
59 mg/kg/day
230 mg/kg/day
2. 5X
1406 mg/kg/day
Index of Toxicity
12 mo.
24 mo.
4 weeks
13 weeks
12 DO.
24 mo.
4 wk . 13 wk.
Body Weight as Z
of Control
65Z 80Z
Food Intake as Z
of Control
50Z 105Z
Liver Lesions
Neoplubtlc Nodules
Hype i (< 1 db t lc Koc 1
Ctio 1 anglof ibrosla
4/4 <2/4)d/

** Reference: Ellis et al., 1978a.
'tn<; Intake corrected for evaporation loss and Intake patterns.
Statistically significant as compared to controls.
^Anlujdly fed lor I J weeka with TNG and allowed to recover for 4 weeks.


Table A2-L Incidence of Hepatocellular Carcinoma in Racs fed T>G
for l'p To 24 Months*
0. iz
Sacrif iced^'
. 3/33
Z with
Reference: Ellis at si., 1978*.
Number with hepatocellular carclocma/nuaber neeropsied.
Fed TNG for 52 weeks. control faed for 4 weeks.
F«d TNG for 104 weeks, control feed for 4 weeks.
38/sex/level ac start of experlaent

table A2-2 Incidence of riyperplastic Foci of vhe Liver in Rats Fed 7\G
fdr Up To 24 Months
0. 1Z
Z with
, 67Z
..Reference: Ellis et el., 1978*.
'Number with hyperplastic foci/number necropsied.
Fed TNG for 52 vteka, control feed for 4 weeks.
fed TNG for 104 weeks, control feed for 4 weeks.
38/sex/level ae scare of experiment

Table A2
-3 Incidence of
for L'p To
Nodules of , the L
24 Months*
iver In Rats Fed



0. iz
Z with
b .Reference: Ellis et al., 1978a.
'Number with neoplastic nodules/nuaber n«crops,ied.
jyFed TNG for 52 weeks, control feed far 4 weeks.
(fed TNG for 104 weeks, control feed for 4 weeks.
6 38/sex/level at the start of the experlaeat.
A 2--

Table Al-- Incidence of Interstitial Cell Tumors or the Testis in Rats Fed ~
TNG tor Up To 24 Months*'
(3.04 mg/kg/day)
(31.5 og/kg/day)
(363 mg/kg/da
Total .
0 /
Z with
Cell Tumors

9 J
. .Reference: Ellis et al., 1978*.
,Number with latarscitlal call tumors of cestis/number necropsied.
j,Fad TNG for 32 waaks, control diet for 4 weeks.
.Fed TNG for 104 weeks, control diet for 4 weeks.
38/sex/level at start of experiment.


Table A3-L Incidence of u<-ocardial Degeneracion/Fibrosis in Racs Fed
TNG for Up To 2U Monchs
Males	Females
Control O.OIZ	O.IZ	1Z	Control 0.01Z O.IZ	IZ
CO)	(3.05) (31.5)	(363)	(0)	(3.99) (38.1) (434)
15/32C/ 20/34	16/33	6/29	12/37	10/40 5/36	1/33
47X	591	45Z	21J	32X	25Z	14Z	3Z
..Reference: Ellis et el., 1978a.
.Dose in ag/kg/day.
Number vicb myocardial degeneration or fibrosia/number necropaied.

Data Deficiencies/Problem Areas and Keconaendatlons For
Additional Data Base Development for Trlnlcroglycerol

A 4-2

the Office of Drinking Water (ODW), Environmental Protection Agency (EPA), in
conjunction with the Department of the Army, has reviewed the available data on
trinitroglycerol (TNG) for the purpose of developing a Health Advisory (HA) useful
in dealing with contamination of drinking water, to include "state-of-the-art"
information on health effects, analytical methodology and treatment technology.
This information is contained in detail in. the report entitled "Health Advisory on
Tha objective of this document is to provide an evaluation of data
deficiencies and/or problem areas encountered in the review process for TNG and to
make recommendations, as appropriate, for additional data base development. This
document ia presented as an independent analysis of the current status of TNG
toxicology, as relates to its possible presence in drinking water, and Includes a
summary of the background information used in the development of the HA. For
greater detail on the toxicology of TNG, the Health Advisory on Trinitroglycerol
should be consulted.
Trinitroglycerol Is a pale yellow, viscous, volatile liquid used in
commercial and military explosives and, clinically, as a vasodilator (Rosenblatt
ec al., 1973). It is produced for military use at selected Army Ammunition Planes
(AAPs) and, because of its appreciable solubility in water (1800 mg/L at 20" C),
can be expected to be found dissolved in waste rinse water. It is incompletely
biotransformed by bacteria and, at concentrations above 75 mg/L, is toxic to
microflora of activated sludge (Walsh, 1976 as cited In Sullivan et al., 1979).
Proposed treatment consists of Ca(0H)2 hydrolysis of the neutralized waste streams
followed by an activated sludge process employing a rotating biological contactor
(Patterson et al.t 1976 as cited in Sullivan et al., 1979).
The pharmacokinetic properties of TNG have been extensively studied in
several species including dogs, rate and mice. Lee et al. (1975, 1976) lndicaced
that TNG is readily absorbed in all species with absorption essentially complete
in 24 hours. Tha absorbed radioactivity Is highly concentrated in the liver of
all species vleh the kidneys also retaining significant amounts.
In vivo studies indicate TNG is rapidly denitrated with urinary metabolices
consisting largely of free mononitroglycerins (MNGs), glycerol and other polar
metabolites including dlnitroglycerln (DNG) and MNG glucuronldes. Mice are
reported to metabolize TNG more completely than any other species tested (Lee ec
al., 1975, 1976). The liver of several species, Including humans, was unable to
denitrate TNG in vitro as completely gs the in vivo system. Excretion of absor.-ec
radioactivity after dosing with 1,3- C-TNG was primarily in the urine with mice

excrecing equivalent amounts In tne reces and expired air and rats excreting
lesser amount* in expired air. Results of biliary excretion studies suggest
relatively complete absorption after oral administration (Lee et al., 1975, 1976).
Acute toxicity studies conducted by Lee et al. (1975) indicated LD values
averaging 850 mg/kg la racs and 1120 mg/kg in mice. Toxic signs included
cyanosis, ataxia and respiratory depression. Death, when ie occurred, vas usually
in 5 to 6 hours after dosing.
Four-week feeding studies conducted by Lee et al. (1976) in dogs, rats and
mice at levels up to 1, 59 and 100 mg/kg/day, respectively, produced no toxic
signs except for an lncake related depression in weight gala In the high dosed (59
ag/kg/day) rats. Doses in these animals were raised 5-fold to averages of 5, 230
and S80 mg/kg/day, respectively, and continued through a total of 13 weeks. No
additional toxic signs related to TNG intake were reported at any treatment level
except for an elevated SGOT level In tvo of the high dosed rats and extramedullar
hematopolesis of the liver and spleen and a variable Increase in spleen weight in
the TNG treated mice. The significance of these findings, If aay, waa not
A follow-up study with rats fed an average of 1410 mg/kg/day of TNG for 13
weeks (Lee et al., 1976) established the relationship between the decreased
intake/decreased weight gain pattern seen in previous studies. The most
significant toxic sign* in this study, however, were testicular atrophy and
degeneration and aspermatogenesis in. the test males. Both sexes also showed
evidence of hemosiderosis of the spleen and/or liver.
Feeding of TNG to dogs for 12 months at levels up to 25 mg/kg/day and for up
to 24 months In rats and mice (Ellis et al., 1978a) at levels up to 400 and 1040
mg/kg/day, respectively, produced variable degrees of methemoglobinemia in all
three species. In rats* these levels were significantly Increased (p<0.05) in the
high dosed group (400 ag/kg/day) evident after 3 months of treatment and remaining
so through the 18th month of treetaent. While actual levels were not
significantly Increased in dogs or alee, sequelae related to methemoglobinemia,
including Helns bodies* anemia and pigment deposits* vera evident In the high
dosed alee (1040 ag/kg/day). Both racs and alee also showed a decrease in weight
gain and food Intake during the early weeks of the study. A high mortality rate
among male mice of the control and all cast levels resulted In no surviving
high-dose males by 24 months. The bosc significant findings In these studies were
among rats which. In addition to enlarged livers and pigmentation of the
epithelium of the kidneys and spleen, developed hepatocellular carcinomas and
various degrees of pre-carclnogenic lesions when treated at both the middle and
high dose levels (averaging 35 and 400 mg/ltg/day, respectively). Male rats also
developed an'Increased Incidence of interstitial cell tumors of the testes after
feeding with TNG at approximately 360 ag/kg/day for 24 months.
Trinitroglycerol vas reported to be generally non-mutagenic in various
A4- U

indicator systems including mammalian, bacterial and yeast assays (Ellis et ai.,
1978a; Simmon et al., 1977; and Lee et al., 1976). Only weak mutagenic activity
in the Salmonella/microsome plate assay against two tester strains was reported by
Ellis et al; (1978b).
Impaired fertility in high dosed rats (363 and 434 mg/kg/day, males and
females, respectively) of the F and F parental generations and reduction in
several parameters of the high dosed litters	?[.» ? ^ w#re considered as
indirect toxic effects attributed to the testicular effects in the high dosed
males and to the generally poor nutritional status of the parental animals at the
time of mating (Ellis et al., 1978a). No effect on fertility or perl- and
postnatal development were noted in rats receiving intraperitoneal doses of TNG up
to 20 mg/kg/day (Oketani et al., 1981b,d).
No significant teratogenic effects were reported following intravenous (up co
4 mg/kg/day) or intraperitoneal (up to 20 mg/kg/day) doses of TNG in rabbits and
rats, respectively (Oketani et al., 1981a,c). Ellis et al. (1978a) also reported
no tertogtnic effects in rats dosed up to 434 og/kg/day; however, the experimental
design of this study vas not fully defined.
Exposure to TNG had no effect on the immunologic response of dogs and rats
but dosing with TNG for 4 days at levels up to 200 ag/kg/day in dogs did produce a
significant increase in methsmoglobin formation along with cyanosis and decreased
physical activity at the two highest dosage levels (Lee et al., 1976).
Several methods of chemical analysis, apparently adequate for detection of
TNG at low concentrations in water, have been cited in Sullivan et al. (1979).
Treatment methods for TNG wastewater (Patterson et al., 1976 as cited in Sullivan
et al., 1979) following sedimentation and precipitation procedures have been
described and, while still in the experimental stage at the time of publication,
appear to be adequate for treating the levels of TNG likely to be encountered.
Based on the significant findings of the foregoing studies along with the
available data on the clinical use of TNG in humans, HA values for one-day,
10-days, longer-term and lifetime exposures were established et 0.00S mg/L.
Available data on the pharmacokinetics, health effects, analysis and
treatment of TNG have been reviewed.
The pharmacokinetic properties of TNG have been studied in various species
and results indicate that absorption and metabolism occur rapidly and are similar
in all species studied. Limited in vitro studies with human livers Indicate thac
man also metabolizes TNG in a similar manner and further studies are unlikely to
yield additional data pertinent to the development of HA values.

The available studies on the toxicity of TNG include LD, 's in rats and mice
and short-cera (13-week) and longer-term (24-month) studies in dogs, racs and
mice, including assessments for carcinogenic potential. While the levels selected
for ch« short-term studies were generally too lov to establish effect/no-effecc
doses, the longer-term studies clearly established toxic levels in at least one
species. Additional short-tern studies have, hovever, established toxic effects
in rats at a higher dose level and the levels at which methemoglobinemia becomes
significant In dogs. While no high-dose treated male mice vere available after 24
months of feeding due to a high mortality rata aaong all mala alee, controls
Included, the apparently low toxicity of TNG In this species indicated that little
additional Information would have been made available from this study. Additional
short-term studies at high dose levels selected to establish an effect/no-effect
dose range would be useful; however, the available human data are adequate for
establishing short-term HA values.
Three-generation reproduction studies in rats, mutagenicity assays in several
systems and iomunologlcal studies In dogs and rats have also been reported. All
studies appear adequate for use In HA development.
Teratogenic studies In rats or rabbits provided no conclusive evidence of
teratogenicity. However, questions remain as to the tlma and duration of dosing
and tha age and status of cha females used in tha study conducted by Ellis et al.
(1978a). Specifically, tha experimental methods, as outlined, indicated that the
Fq females from tha reproduction study were discarded following weaning of the
second litter. Tha body of this report, however, indicated that these females
were mated a third time and were subsequently used for the teratogenic study.
Neither procedure fits the protocol described in Appendix II of the report of
Ellis et al. (1978a). This referenced protocol indicates that virgin females are
treated during the period of oncogenesis, I.e. days 6-15 in the rat. It Is,
therefore, considered advisable to either repeat the tertogenle study according to
an accepted standard protocol or to clearly ascertain that the procedures used in
the prassnt study would yield data valid for use In the assessment of
Several methods of analysis for TUG la wastewater have been reported. The
limit of detectabllity, 0.002 mg/L, using the method of Weitzel et al. (1976) as
described la Sullivan et al. (1979), appears adequate to detect levels of TNG chat
may be considered-hazardous to health. Likewise, several methods of treatment of
water contaminated with TNG, including the method of Patterson et al. (1976), have
been described by Sullivan et al. (1979) and appear adequate.
Based on the above discussion, the following concluslons/recomendations can
be made:
1. The available studies on TNG tcxUlt> ire generally considered adequate

for development of a HA useful in dealing vieh the potential
contamination of drinking water.
A repeat of the teratogenicity study according to standard FDA protocols
is advisable to assess the teratogenic potential of TNG.
Aside from the aforementioned data deficiency, no further studies on
TNG, as relates to its possible presence in drinking vater, are deemed
necessary at this time.

Ellis', H.V., III ec al., 1978a. Mammalian toxicity of munitions compounds Phase
III: Eff«cts of life-time exposure Part II: Trinitroglycerin. Progress Report
No. 8. Midwest Research Institute, Kansas City, MO, Contract No.
DAMD-17-74-C-4073, AD A078746.
Ellis, H.V., III et al., 1978b. Mammalian toxicity of munitions compounds Phase
I: Acute oral toxicity, primary skin and eye irritation, dermal sensitization,
disposition and metabolism, and Ames teats of additional compounds. Progress
Report Mo. 6. Midwest Research Institute¦ Kansas City, MO, Contract No.
DAMD-17-74-C-4073, AD A069333.
Lee, C.C. et al., 1975. Manalian toxicity of munition compounds: Phase I.
Acute oral toxicity, primary skin and eye irritation, dermal sensitization, and
disposition and metabolism. Report No. 1. Midwest Research Institute, Kansas
City, MO, Contract No. DAMD-17-74-C-4073, AD B011150L.
Lee, C.C. et al., 1976. Mammalian toxicity of munition compounds Phase II:
Effects of multiple doses Part I: Trinitroglycerin. Report No. 2. Midwest
Research Institute, Kansas City, MO, Contract No. DAMD-17-74-C-4073, AD A047067.
Oketani, Y. et al., 1981a. [Toxicologlcal studies on nitroglycerin (NK-843) (6)
Teratologlcal study in rabbits.] Pharmacometrics. 22(5):633-638 (Jap).
Oketani, 7. et al., 1981b. [Toxicologlcal studies on nitroglycerin (NK-843) (7)
Fertility study in rats.] Pharmacometrics 22(5):639-648 (Jap).
Oketani, Y. et al., 1981c. [Toxicologlcal studies on nitroglycerin (NK-843) (8)
Teratologlcal study in rats.] Pharmacometrics. 22(6):737—751 (Jap).
Oketani, Y. et al.. 1981d. [Toxicologlcal studies on nitroglycerin (NK-843) (9)
Perinatal and postnatal study in rats.] Pharmacometrics. 22(6)1753—763 (Jap).
Patterson, et al., 1976. State-of-the-art: Military explosives and
propellants production Industry. Vol. II. EPA 600/2-76—213C as cited in Sullivan
et al., 1979.
Rosenblatt, D.B. et al., 1973. Munitions production products of potential concern
as waterborn* pollutants-Phase I. U.S. Army Medical Environmental Engineering
Research Unit Report No. 73-07, AD 912752 Edgewood Arsenal, MD.
Simmon, V.F. et al., 1977. Mutagenicity of soma munition wastewater chemicals and
chlorine test kit reagents. Final Report. SRI International, Menlo Park, CA,
Contract No. DAMD-17-76-C-4073, AD A057680.

Sullivan, J.H. ec al., 1979. A summary and evaluation of aquatic environmental
data in relation co establishing water quality criteria Cor munition-unique
compounds. Parr 2: Nitroglycerin. Final Report. U.S. Army Medical Research and
Development Command, Contract No. DAMD 17-77-C-7027, Water and Air Research, Inc.,
Gainesville, FL.
Walsh, J.T., 1976. Chemical characterization of nitroglycerin biodegradatlon
produces. U.S. Army Naeick Research and Development Laboratory. Unpublished
manuscript. Naticlc, MA, as cited in Sullivan et al., 1979.
Weitzel; R.L. et al.. 1976. Aquatic field surveys at Iowa. Radford, and Joliet
Army Ammunition Plants. Final Report. Volume II. Radford Army Ammunition Plant.
U.S. Army Medical Research and Development Command, Contract No. DAMD
17-75-C-5Q46, Environmental Control Technology Corp., Ann Arbor, MI, as cited in
Sullivan et al., 1979.

«d«aa. A spaca of daA:i".s sf -ir.xr.cvr :a-sa :c:^rrtd <- — ;« i:;eT -,-4	.
feadlsg SC or cotton lUtafJ at :oZ in :!*.« iiit. >a -.soar 3: .as -
:a c*a NC craatad aiea and ®*y. :n«r#£or«. ba raiarad zs :ha cr#a::<:	:
:*« compound via a 3i;htnlu> ai v«c, ' ' Zi'.'.s «c al., 1-30)
Nterocalluloaa was raporcad to ba non-outaganlc la various indicator vj*»=i
<£11is ac *1., 1976, 1973, 1930). thraa-ganaraeion raproductlsn seudias m ratj
Indicated chat HC did nee advarsaly affact caproduccicn but that tha non-nutr
bulk of tha I OX dlat aay cam# an advarsa affaec during pariods of high
nutritional daauad, such as prtgnancy or lactation, a* avldancad by a dacrtasa -.n
lactation lndaa and waighc ac vaanlng (Ellis at al., 1930).
No earstoganic scudlas wort raporcad. Expoaura to NC did noe affact era
ineunological raaponaa of dogs or rata (Cilia at al., 1976).
Kachoda of analysis (Barklay and losaablatt, 1978) and traacaane (Xosanblatc
at al., 1973) adaquata for dataecioa and raaoval of HC at lavala which aigtit b«
daaoad haaardoua to haalch hava baas praaaaead la datail la "Haalch Advisory on
Tha laek of coxlcologlcal indlcacora, along vlch tha apparane
noa-abaorpclou/noa-dlgaatloa of ths HC flbars, lad to cha conelualon that HA
valuas for NC la drinking watar wars uaaacassary.
Avallabla data on tha pharaacoklaaclc*. haalth «ffacts, analysis and SO
vastavatax traataaat hava baaa ravlavad.
Vhlla tha avallabla data oa tha aatabolisa of HC ara 1laicad In scops, tha
chaalcai aad physical natura of HC gaaarally supports tha fladlag of tha ona
avallabla study that HC paaaaa through tha 61 tract apparaatly unchangad and
uaabaorbad. Additional acudlas would* tharafora, ba uaaacassary.
Tfca avallafcla atudiaa «n tha toxicity of HC lacluda l"0«o'c 1a	nica
aad short-cam (11 vsab) aad loa«ar~car* (24 aoatb) studies la dogs, tats and sic*
that lacludaA aasassaaaca for posslbla carcinogenicity. Thraa-ganaracion
raproductldfestadia* la rata, Mutagenicity assays 1ft baetarlal aad cycogaaatie
systaa* safelaueloflcal atudiaa la dags sad rat* haw* also baaa raportad. All
scudlas ayya**~ada^aax« tar uaa la HA davalopaaac.
rurthar larastlgacloa of tha causa of tha spata of daatha la alca occurring
at alu aoatha la tha loagar-tara faadlaf study saa* warranted but, la via* of the
otharwlaa low toxicity af * alan* with tha oefcer awatLafela data, would not ba
necessary for 84 davalopwaa*. Tarat manic itsrilas vara aot raporcad; hovavar. the
apparaat laablllty of K ta ba absorbed mmlA praclada It* actlag as a potential
teratogen. Tharafora. additional atudiaa aaaa uaaacassary ac this tiaa.

. r r 5) . icvavar, s::i :ed l"- :r	--".:ars „;
edta* asd ^.ypiraaia 0f :r.a	txcrtai::es. a??ar«-.:l/ :3 :-
^'C. *or« »aaaleiva d«raal	s:udi«s, available. vouli :• a
::t ba d»««ad aaetssary is	:«ga?ds isn o! :r.# drir.
Savaral saehods of. analysis of NC in vasttvacar hava b«
a sachod adapead eo tha Tachnicon Autoanalyxar. Tha lisle o
0.4 ag/l, appears adaquata (or dataraining drinking vaear co
to palaeablliey concaraa.
Haehods for the traataaac of vastavaear by chaaical and
da|radaeioB hava baaa davalopad. Tha inaoiubllity of HC la <
to ba rasevad by coagulaeloa and/or flleraeion aaka it uniik.
•xcaaalva traacaane saasuraa would ba raqulrad.
Baaad on tha abova dlscuaaloa, tha following coaclualoa:
ba ud«:
1.	Tha avallabia aeudlaa oa NC toxicity ara adaquata
HA uaaful la da*llag with eoaeaaia*eloa of drlaklai
2.	No algalfleant data gap* or problas araaa ralatlva
vacar ulit.
3.	No furthar aeudlaa oa NC. aa ralataa eo lea poaalbj
drlakiag vatar, ara aaeaaaary ae ehla tlma.

3arltlay. J.J. and RosanbUtc. D.H., '.9:3. Aut'5=ae«d nltracai:-l:s« a-a. 3.5.
:\S. Amy Madlcal 81o«nglr.a«rini Rasaareh and Davalapsant '.aboratary Tec--.;4.
Rapore 7807. aD AQ67081, Face Otcrlck, MD.
Ellis. H.V., hi «e al.. 1976. ^uullaa toxicity of ounlclon* compounds ?kas«
II: Effaces of aultipla dosas Pare IV; Nitroeallulosa. Raport No. 5. "idvast
Rasaareh laatttuta. Kansas City, MO. Contrsee Me. DAMD-17-74-C-40 7 3, AD aC620!5.
Ellis. H.V., in «c al., 1978. Hinallaa toxicity of aunlelons compounds ?kasa I
Acuta oral toxicity, priaary akin and ay« Irritation, dtraal sansltizattan,'
disposition and aatabolisa, and Asaa eases of additional eoapounds. Raport so. 6
Mldvast Rasaareh Inseieuea. Kansas Cley, MO, Conersce Mo. DAJffl-17-74-C-40 7], aD
Ellis, H.V., in ae al., 1980. Maaaallan toxicity of aunlelons compounds ?