Carcinogenesis vol.26 no.7 pp.1272--1279, 2005
doi:10.1093/carcin/bgi078
Advance Access publication April 7, 2005
Hemoglobin adducts, urinary metabolites and health effects in 2,4,6-trinitrotoluene
exposed workers
Gabriele Sabbioni1,2,, Yu-Ying Liu3, Huifang Yan3
and Ovnair Sepai4,5
1
Institute of Environmental and Occupational Toxicology, Casella Postale
108, 6780 Airolo, Switzerland, 2Walther-Straub-Institut f€ur Pharmakologie
und Toxikologie, Ludwig-Maximilians-Universit€at M€unchen, 80336
M€
unchen, Germany, 3Chinese Academy of Preventive Medicine, Nan Wei
Road 29, Beijing 100050, China, 4Department of Environmental and
Occupational Medicine, The Medical School, University of Newcastle upon
Tyne, Newcastle upon Tyne, UK and 5Health Protection Agency,
CHAPD-HQ, Chilton OX11 0RQ, UK
To whom correspondence should be addressed
Email: [email protected]
2,4,6-Trinitrotoluene (TNT) is an important occupational
and environmental pollutant. In TNT exposed humans, the
notable toxic manifestations have included aplastic anemia,
toxic hepatitis, cataract, hepatomegaly and liver cancer.
Therefore, we developed methods to biomonitor workers
exposed to TNT. The workers were employed in a typical
ammunition factory in China. The controls were recruited
from the same factory. We determined hemoglobin (Hb)
adducts and urine metabolites of TNT. Hb-adducts of TNT,
4-amino-2,6-dinitrotoluene (4ADNT) and 2-amino-4,6dinitrotoluene (2ADNT), and the urine metabolites of
TNT, 4ADNT and 2ADNT were found in all the workers
and in a few controls. 4ADNT was the main product.
Although the levels of 2ADNT correlated well with
4ADNT, 2ADNT was not found in all the samples. Therefore, 4ADNT was the best marker of exposure for Hbadducts and urine metabolites. The levels of the urine
metabolites and Hb-adducts were related to the health
status of the workers. The Hb-adduct 4ADNT was statistically significantly associated with risk of hepatomegaly,
splenomegaly and cataract. The odds ratio (OR) for cataract, splenomegaly and hepatomegaly were 6.4 [95% confidence interval (CI) ¼ 1.4--29.6], 9.6 (1.1--85.3) and 7.6
(1.3--43.7), respectively. No correlation was found between
urine metabolites and health effects. These results were
tested for confounding factors like age, workyears, smoker
status, smoke years, cigarettes per day and hepatitis
B status using stepwise forward logistic regression analysis.
In the case of splenomegaly, hepatitis B status is a confounder. In the case of cataract, age is a confounder. The
Hb-adduct, 4ADNT, is a good biomarker of exposure and
biomarker of biological effect.
Abbreviations: 2ADNT, 2-amino-4,6-dinitrotoluene; 4ADNT, 4-amino-2,6dinitrotoluene; 4ABP, 4-aminobiphenyl; ECG, electrocardiogram; EI-MS,
electron ion ionization-mass spectrometry; GC--MS, gas chromatography-mass spectrometry; Hb, hemoglobin; NCI, negative chemical ionization;
PFPA, pentafluoropropionic anhydride; SGPT, serum glutamicpyruvic
transaminase; TNT, 2,4,6-trinitrotoluene.
Carcinogenesis vol.26 no.7 # Oxford University Press 2005; all rights reserved.
Introduction
2,4,6-Trinitrotoluene (TNT) is an important environmental
and occupational pollutant (1--6) (http://www.atsdr.cdc.gov/
toxprofiles/phs81.html). In TNT exposed humans the notable
toxic manifestations have included; aplastic anemia, toxic
hepatitis and cataracts (1,3,4). Similar conditions have been
observed in animals. In China, chronic occupational exposure
to TNT caused mainly hepatomegaly and cataract. The incidence of hepatomegaly was 41% and that of cataract was 79%
in the TNT workers with prolonged exposure, whereas anemia
was found rarely (7). A recent retrospective study, on male
workers exposed to TNT over 1 year from eight Chinese
military factories, from 1970 to 1995 demonstrated a higher
relative risk for malignant tumors, especially liver cancer (8).
The morbidity of total malignant tumor in male TNT exposed
workers was markedly higher than that of controls, and the
relative risk was 2.3. Liver cancer morbidity was 31.9% of
the total malignant tumor, and its mortality was 3.97 times of
the controls. A preliminary study of a German population
living near the sites of two World War II munitions plants,
indicated an association between increased rates of some types
of leukemia and living in a town near TNT waste from these
plants (9). The study showed increased relative risk of acute
myelogenous leukemia for adult males and females living near
the former explosives plants when compared with adults in a
neighboring county. The relative risk was particularly high
for individuals 465 years of age. However, the study case
numbers were very small.
In mammalian systems, the principal metabolites of TNT
are 2-amino-4,6-dinitrotoluene (2ADNT) and 4-amino-2,6dinitrotoluene (4ADNT); smaller amounts of other metabolites
are also formed (4,10,11). In humans mainly 2ADNT and
4ADNT were found in urine (4,12--14). Two studies showed
that the urine metabolite levels were too high compared with
expected levels calculated from the air measurements (12,13).
It has been deduced from these data, that absorption through
the skin must be the main pathway of exposure. Therefore,
biological monitoring of workers is preferable as the method of
assessing exposure rather than environmental monitoring. The
metabolites of urine usually indicate only recent exposures up
to 48 h post exposure (15). Hemoglobin (Hb) adducts are an
indicator of exposure over the last 4 months, assuming that the
adduct is stable and that the lifetime of the erythrocytes is not
affected. The mechanism of adduct formation between aromatic amines or nitroarenes with Hb involves the reaction of
the metabolite—nitrosoarene—with cysteine residues to form
a sulfinic acid amide (16--20). Evidence for such reaction
in vitro has been obtained by Ringe et al. (18). Sulfinic
acid amide adducts are readily hydrolyzed under mild conditions, yielding the parent amine. Hb adducts are dosimeters for
the internal and possibly of the target dose leading to toxic
effects. Covalent Hb, plasma protein and protein adducts in
various tissues were found in rats dosed with radiolabeled TNT
1272
Biomonitoring studies in TNT exposed workers
(21). Therefore, Hb-adducts of TNT might be a biomarker
of toxic effects. The presence of the two Hb-adducts
indicates that 4-(N-hydroxyamine)-2,6-dinitrotoluene and
2-(N-hydroxyamine)-4,6-dinitrotoluene are biologically available. These intermediates can be formed by reduction of TNT
in the gut, in the liver and in the erythrocytes (21,22). In vitro
experiments with TNT showed that 2% of the dose is bound to
Hb (21). Therefore, it is conceivable that a small amount of
TNT adducts found in vivo might result from the reactions of
TNT in the erythrocytes (21,22). This is not the case for the
less reducible nitroarenes, such as nitrobenzene (23). Hbadducts have been found previously in workers (24,25).
Toxic damage from TNT might not be caused by protein
adducts alone but also from oxidative stress (26--28). Extensive biochemical and morphological changes in rats after a
single injection of TNT were indicative of peroxidation
(27,28), possibly caused by the cyclic oxidation and reduction
of the nitro group (26). It has been postulated that cataracts
result from oxidative damage of the lens (29).
For the present study, blood samples, work place description, air measurements and medical examinations were collected from Chinese workers and non-matching controls. The
external dose, the urine metabolites and Hb-adducts were
determined and compared with the health effects.
Methods
Chemicals and reagents
2ADNT and 4ADNT were obtained from Promochem (Wesel, Germany).
Dichloromethane (Pestanal) was obtained from Fluka (Buchs, Switzerland),
pentafluoropropionic anhydride (PFPA) was obtained from Supelco
(Taufkirchen, Germany), diethylether (stabilized with EtOH), sodium hydroxide (No. 6498) and water (No. 15333) from Merck (Darmstadt, Germany).
d9-4-Aminobiphenyl (d9-4ABP) (30), 2,4-d6-dinitrotoluene (31), d5-2ADNT
and d5-4ADNT were synthesized in our laboratory as described below.
b-Glucuronidase type HP-2 from Helix Pomatia (100 000 U/ml) was obtained
from Sigma (Taufkirchen, Germany).
Synthesis of d5-TNT. To 2,4-d6-dinitrotoluene (150 mg) (31), 96% D2SO4
(900 ml) and 100% HNO3 (300 ml) were added. After 1 h at 80 C and 2.5 h
at 90 C, the reaction mixture was cooled to room temperature. Cold D2O was
added and extracted with benzene (2 3 ml). The collected organic phases
were washed with 2 ml of saturated NaHCO3 and then with saturated NaCl
solution (2 3 ml). The organic phase was dried through a pasteur pipet filled
with anhydrous Na2SO4 and evaporated under reduced pressure. Slightly
yellow crystals of d5-TNT were obtained (170 mg, 77% yield). Electron ion
ionization-mass spectrometry (EI-MS) (70 eV), m/z (%): 215 (9), 214 (100),
213 (24), 196 (15), 184 (11), 168 (11), 152 (11), 138 (14), 94 (48), 93 (17), 92
(14), 66 (46), 64 (12), 54 (20). Negative chemical ionization-mass spectrometry (NCI-MS): 233 (10), 232 (100), 231 (8), 230 (4), 214 (19), 202 (28).
Synthesis of d5-4ADNT and d5-2ADNT. To d5-TNT (12 mg) in methanol
(0.6 ml), palladium on carbon (2 mg) and ammonium formate were added at
room temperature. After 25 min in a shaking bath, the mixture was separated
on silicagel plates without fluorescence indicator and dichloromethane as
solvent. The silicagel zones were cut out and extracted with ethyl acetate in
a glass tube. The extracts contained d5-4ADNT and d5-2ADNT as determined
by gas chromatography and mass spectrometry (GC--MS). EI-MS (70 eV),
m/z (%): d5-4ADNT: 203 (6), 202 (63), 201 (6), 185 (11), 184 (100), 166 (14),
154 (16), 126 (15), 124 (16), 110 (44), 108 (100). d5-2ADNT: 203 (8), 202
(81), 201 (5), 184 (100), 156 (19), 138 (29), 136 (27), 126 (16), 124 (13), 110
(50), 109 (32), 108 (74).
Questionnaire and medical examination
The workers for this study were recruited from Tuoli Chemical Factory (in the
surburb of Beijing). The workers (n ¼ 78) were mostly involved in the
production of satchel charges for mining. The workers were classified as
mixers, loaders, grinders and packers. The time of exposure was 8 h per day
and 5 days per week. The controls (n ¼ 25) were employed in the same factory.
Of the 25 controls, 8 were exposed 8--20 years to TNT up to 2 years ago.
The exposed workers were 76 males and 2 females (1 female did not have a
job classification). The control group consisted of 9 females and 16 males.
The mean age of the controls and exposed group were 38.4 and 39.6 years,
respectively. The mean workyears of the controls and exposed group were 15.1
and 9.8, respectively. The study was performed in accordance with the principles embodied in the declaration of Helsinki (www.wma.net/e/
policy/b3.htm). Informed consent was obtained from each worker. Blood
collection, urine collection, medical examination and questionnaire were
all performed in the same week. The blood was collected by the medical
department of the factory. The plasma, the erythrocytes and the urine samples
were transported on dry ice to Europe. Each participant was interviewed with a
questionnaire about their general status, exposure history, smoking habits,
previous medical record and present symptoms. The medical department
of the Chinese Academy of Preventive Medicine performed the following
examinations.
(i) Physical examinations: blood pressure, cardiovascular system and
heart rate.
(ii) Routine blood tests, liver function test: serum glutamic pyruvic transaminase (SGPT), alkaline phosphatase, total protein, albumin and total
bilirubin.
(iii) Electrocardiogram (ECG).
(iv) Ultrasonic type B examination for liver and spleen.
(v) Serological assays of hepatitis B antigens and antibodies were performed,
because hepatitis B is rather common in China, and liver damage can also
be caused by TNT.
(vi) Ophthalmological examinations.
Isolation of Hb from blood and determination of Hb adducts
EDTA blood (1--2 ml) was centrifuged for 5 min at 2000 g. After removal
of plasma, red blood cells were washed three times in equal volumes of
0.9% NaCl solution and lysed by the addition of 4 vols of water. The cell
debris were removed by centrifugation. Hb was precipitated with ethanol from
the lysed erythrocytes. The precipitate was washed with ethanol--water (8:2),
ethanol, ethanol, ethanol--diethyl ether (3:1) and diethyl ether. The dried Hb
(50 mg) was dissolved in 0.1 M NaOH (3 ml) in new screw top glass tubes
(16 mm 100 mm) fitted with teflon liners. The internal standard mixture
d5-2ADNT/d5-4ADNT (2 ng, 10 ml) in ethyl acetate was added to the Hb
solution. After 1 h in a shaking bath at room temperature, dichloromethane
(5 ml) was added. The mixture was vortex-mixed for 1 min, centrifuged for
5 min at 3000 g and frozen in liquid nitrogen. The thawed organic layer
was transferred to a graduated tapered tube (98 mm 15 mm) and concentrated down to 200 ml in a speed evacuator. The residue was then transferred
to a microinsert (200 ml) for 12 mm 32 mm autosampler vial and evaporated
very carefully under a stream of nitrogen at 25 C. At the disappearance of
the last drop, the stream of nitrogen was stopped. The residue was then
taken up in 10 ml ethyl acetate containing 2 ng of the PFPA derivative of
d9-4ABP.
Identification and quantification by GC--MS
The analyses were performed on a Hewlett Packard chromatograph
(HP 5890II) equipped with an autosampler (HP 7673) and interfaced to
a mass spectrometer (HP 5989A). The amines were analyzed by splitless
injection onto a fused silica capillary column (Rtx-5MS, I.D. 0.25 mm; length
15 m, 0.5 mm film thickness: Restek corporation, Bellefonte, PA) with a
0.25 mm 1 m methyl--silyl retention gap (Supelco). In all cases the initial
oven temperature, the injector temperature and the transfer line temperature
were set at 50, 220 and 220 C, respectively. The oven temperature was
increased at a rate of 50 C/min to 200 C, held for 1.8 min and then heated at
30 C/min to 300 C. Helium was used as carrier gas with a flow rate of 1.5 ml /
min. The elution order of the amines is represented in Figure 1. For negative
chemical ionization (NCI), with methane as the reagent gas, the source pressure was typically 160 Pa, the electron energy was 160 eV, the emission
current was 300 mA and the source temperature was 200 C. For the identification and quantification of 2ADNT, 4ADNT, d5-2ADNT, d5-4ADNT the
molecular ions 197, 198, 201 and 202 were monitored. The molecular ion
m/z ¼ 304 of d9-4ABP--PFPA was determined in order to control the recovery
of each extraction. The dwell time for all ions was set at 50 ms. The retention
times of d9-4ABP--PFPA, 4ADNT and 2ADNT were 5.83, 6.72 and 7.03 min,
respectively.
In the NCI mode the detection limits of the standard compounds are in the
femtogram range. The determination limit from the present experiments was
found to be 40 pg for 2ADNT and 4ADNT per analysis of 50 mg Hb. The
samples were quantified against calibration curves obtained from Hb (50 mg)
spiked with a constant amount of the surrogate internal standard d5-2ADNT/
d5-4ADNT and with different levels of 2ADNT and 4ADNT (0.2, 1, 5 and
10 ng).
1273
7.03
500
0
6.00 6.10 6.20 6.30 6.40 6.50 6.60 6.70 6.80 6.90 7.00 7.10 7.20 7.30
Abundance
1600
m/z 202
6.71
d5-2ADNT
1400
d5-4ADNT
7.02
75
50
25
0
E-Hb-4ADNT
1200
100
1000
800
600
400
C-U-gl-2ADNT
1000
125
E-U-gl-2ADNT
2ADNT
C-U-gl-4ADNT
1500
150
E-U-gl-4ADNT
2000
C-U-2ADNT
2500
175
E-U-2ADNT
3000
200
C-U-4ADNT
4ADNT
3500
225
E-U-4ADNT
4000
6.72
C-Hb-2ADNT
m/z 197
E-Hb-2ADNT
4500
C-Hb-4ADNT
Abundance
Hb-adducts [ng/g Hb], urine metabolites [ng/100µL]
G.Sabbioni et al.
200
0
6.00 6.10 6.20 6.30 6.40 6.50 6.60 6.70 6.80 6.90 7.00 7.10 7.20 7.30
Time
Fig. 1. Chromatogram of an Hb extract from an exposed worker. Single
ions were monitored in the NCI mode, with methane as the reagent gas
(see Methods).
Fig. 2. Box plot showing the median level of 4ADNT extracted from
hydrolyzed Hb, from enzyme treated urine (U-gl) and from raw urine (U).
Workers were grouped as exposed (E) and controls (C). The 10th, 25th, 75th
and 90th percentiles have been presented as vertical boxes with error bars,
and the outliers 5 and 95% have been highlighted as circles. The 95% outliers
of E-U-gl-4ADNT is out of scale.
likelihood ratio statistic based on the maximum partial likelihood estimate was
used for variable entry. The criterion for entry was set at P 5 0.05.
Urine analysis: extraction of raw urine
Urine sample (100 ml) and 10 ml of internal standard (2.5 ng/ml of
d5-2ADNT and d5-4ADNT) were pipetted into screw capped glass tubes
(25 mm 11 mm) with teflon liners. The samples were vortex mixed for
30 s. The samples were extracted with hexane (250 ml), by vortex mixing for
1 min. The hexane-phase was transferred to a microinsert (200 ml capacity) and
analyzed by GC--MS.
Urine analysis: b-glucuronidase treatment
Urine sample (100 ml), b-glucuronidase buffer (100 ml, 0.4 M NaOAc, pH 4.5)
and 10 ml of internal standard (2.5 ng/ml of d5-2ADNT and d5-4ADNT)
were pipetted into screw capped glass tubes (25 mm 11 mm) with teflon
liners. The samples were vortex mixed for 30 s then 10 ml b-glucuronidase
(1000 U from H.pomatia Type HP-2) was added and the samples were incubated in a shaking bath overnight at 37 C. After cooling to room temperature,
the samples were extracted with hexane (250 ml), by vortex mixing for 1 min.
Following centrifugation, at 3000 g for 10 min, the samples were frozen in
liquid nitrogen, then thawed at room temperature to aid phase separation. The
hexane-phase was transferred to a microinsert (200 ml capacity).
Quantitation of urine metabolites
The same instrumental conditions were used as described above for the
Hb-adduct analyses. Calibration curves were obtained from urine treated in
the same way as the in vivo samples and spiked with the internal standard and
increasing amounts of 2ADNT/4ADNT: 0, 0.01, 0.1, 0.5, 1, 2 mg.
Statistical analyses
Statistical analyses were performed with the program SPSS 10.0, Sigma Stat
2.0 and Sigma Plot 3.0. The results of the questionnaire and of the medical
examination were not known to the scientists performing the Hb-adduct
and urine analyses. All results were disclosed at the end of the analyses.
All tests of statistical significance were two sided. The distribution of the
metabolite levels and of the Hb-adduct levels were markedly skewed; therefore, the data were transformed logarithmically. The logarithmically transformed data showed normal distribution only for the exposed worker group
according to the one sample Kolmogorov--Smirnov test. Inclusion of the
control group yielded a non-normal distribution. Therefore, parametric and
non-parametric tests were performed on the data. Health effects were compared with the Hb-adduct and urine metabolite levels using the Mann--Whitney
test and logistic regression analysis. In the first step univariate logistic regression analysis was used to compare disease with the log-transformed Hb-adduct
data. The inclusion of a constant was allowed. Confounding factors like age,
smoker status, smoke years, cigarettes per day, workyears, and hepatitis B status
were introduced stepwise using a multivariate logistic regression analysis. The
1274
Results
The biological samples were worked up following a standard
method from our laboratory (25,30,31). The samples were
hydrolyzed with sodium hydroxide according to methods for
the hydrolysis of sulfinamide adducts of arylamines (25,20
and literature cited therein). For the determination of Hbadducts a method used previously was applied. However,
instead of 3,5-dinitroaniline the deuterated d5-2ADNT and
d5-4ADNT were used as internal standards. The samples
were analyzed by GC--MS in the NCI mode. Chromatograms
of an Hb extract from an exposed worker is presented in
Figure 1.
The Hb adducts of 4ADNT and 2ADNT were found in 100
and 81% of the exposed workers (78 of 78 for 4ADNT and 63
of 78 for 2ADNT). 4ADNT and 2ADNT were found in 16 and
8% of the factory controls (4 in 25 for 4ADNT, 2 in 25 for
2ADNT), respectively. The mean and median levels in the
exposed workers were significantly higher, as determined by
ANOVA and Mann--Whitney test, respectively (Figure 2). The
median levels of 4ADNT (59 ng/g Hb) were ~24 times higher
than the levels of 2ADNT. Therefore, 4ADNT was found in
most workers but not 2ADNT. However, the levels of 4ADNT
and 2ADNT correlate well (r ¼ 0.81, Table I).
Urine samples were obtained from the same workers. For the
determination of the metabolites of urine, a short and fast
method was developed. Urine samples were analyzed
with and without ß-glucuronidase treatment. Urine was not
collected from all workers. In the group of the exposed workers, 100% (71 of 71) and 97% (69 of 71) urine samples were
positive for 4ADNT and 2ADNT, respectively, in the extracts
of raw urine or in the extracts of enzyme treated urine. In the
group of the control workers, 50% (4 of 8) of the urine samples
were positive for 4ADNT and 2ADNT after both work-up
Job related Hb-adduct and urine metabolite levels
The workers were classified into five different categories:
factory controls, mixers, loaders, grinders and packers. The
Hb-adduct and the urine metabolite levels were compared
between the different groups (Figure 3). The Hb-adduct levels,
increase in the following order: factory controls 5 packers 5
mixers 5 loaders 5 grinders. The order is the same for the
urine metabolite levels except for the packers. The median Hbadduct levels are significantly different between all possible
group pairs (n ¼ 10) except for the group pairs: loaders versus
grinders, and mixers versus loaders and packers. The median
urine metabolite levels are significantly different between all
possible group pairs (n ¼ 10) except for the group pairs:
loaders versus grinders; packers versus mixers, loaders and
grinders. Thus, specific work tasks are critical in regards to
the TNT body burden of the workers.
Relationship between internal dose of TNT and health effects
in Chinese workers
A full medical examination was performed on each of the
Chinese workers. Each worker was examined for the specific
adverse health effects linked to exposure to TNT. Of particular
interest was the investigation of dose--response relationships in
the Chinese workers exposed to TNT. We were interested in
defining whether biomarkers of recent exposure or chronic
exposure correlated with the risk of suffering from one or
more of the health conditions, symptomatic of toxic exposure
to TNT. In the first analysis, the prevalence of cataract, splenomegaly, hepatomegaly, cardiovascular disease (ECG1 ¼ sinus
tachycardia, sinus bradycardia, ECG2 ¼ arrhythmia, ECG3 ¼
conduction), hepatitis B, cigarettes per day and smoke-years
100
50
grinders-U-gl
0
loaders-U-gl
procedures. The levels of 4ADNT and 2ADNT in urine samples without enzyme treatment were much lower than after
enzyme treatment. The median levels of 4ADNT and 2ADNT
were 4.5 and 2.4 times lower in raw urine than in enzyme
treated urine. The urine levels of 4ADNT and 2ADNT in the
control samples were much lower than in the exposed workers
(Figure 2). The urine levels of 4ADNT in enzyme treated urine
correlates with the levels of 4ADNT in raw urine (r ¼ 0.77).
The levels of 4ADNT and 2ADNT correlate for enzyme
treated (r ¼ 0.86) and raw urine (r ¼ 0.85). The levels of
urine metabolites and Hb-adducts do not correlate well (r ¼
0.34--0.55). Urine metabolites reflect the exposure of
the last 48 h, Hb adducts reflect the exposure up to the last
120 days. Therefore, this moderate correlation between urine
levels and Hb-adduct levels is not unexpected.
150
mixers-U-gl
Correlation coefficient r was determined from logarithmic transformed
data (Pearson correlation). The statistical significance of each correlation
was given by the two-sided P value.
a
P 5 0.001.
b
P 5 0.01.
200
packers-U-gl
0.52a
0.52a
0.71a
0.78a
0.86a
controls-U-gl
0.55a
0.52a
0.77a
0.65a
grinders-Hb
0.34b
0.37b
0.85a
loaders-Hb
0.41a
0.44a
250
mixers-Hb
Hb-4ADNT 0.81a
Hb-2ADNT
U-4ADNT
U-2ADNT
U-gl-4ADNT
300
packers-Hb
Hb-2ADNT U-4ADNT U-2ADNT U-gl-4ADNT U-gl-2ADNT
350
controls-Hb
Table I. Exposed TNT-workers: correlation of Hb adducts (Hb-2ADNT,
Hb-4ADNT) with urine metabolites obtained from raw urine (U-4ADNT,
U-2ADNT) and from ß-glucuronidase treated urine (U-gl-4ADNT,
U-gl-2ADNT)
Hb: 4ADNT [ng/g Hb], urine: 4ADNT [ng/100 µL]
Biomonitoring studies in TNT exposed workers
Fig. 3. Box plot showing the median level of 4ADNT determined in Hb and
enzyme treated urine from workers exposed to TNT. Workers were grouped
according to their job description. The 10th, 25th, 75th and 90th percentiles
have been presented as vertical boxes with error bars, and the outliers 5 and
95% have been highlighted as circles. The 95% outlier of the Hb-adduct from
the loading workers is out of scale. Hb-4ADNT (U-gl-4ADNT) was
determined in 25 (8) controls, 6 (5) packers, 35 (33) mixers, 26 (23) loaders
and 10 (10) grinders.
Table II. Comparison between the prevalence of adverse health effects
and confounding factors in exposed and factory controls
Controls (n ¼ 25) (%)
Cataract
Hepatomegaly
Splenomegaly
ECG1a
ECG2b
ECG3c
Hepatitis B
Smoker
d
26
12
0
12
100
24
20
64
Exposed (n ¼ 78) (%)
e
60
22f
13g
22h
79
17h
29
84
P
0.004
0.25
0.013
0.28
0.002
0.45
0.34
0.033
The number of exposed workers where a symptom was recorded has been
calculated as a percent of the total number of exposed workers analyzed. The
number of factory control workers where a symptom was recorded has been
calculated as a percent of the total number of factory control workers
analyzed. The statistical difference of the prevalences between exposed and
control workers was tested with the x2-test (likelihood ratio).
Electrocardiogram (ECG): asinus tachycardia, sinus bradycardia, barrhythmia,
c
abnormal conduction, d-hThe data were available for: dn¼23, en ¼ 68,
f
n ¼ 72, gn ¼75, hn ¼ 65.
were compared between controls and exposed. The results are
summarized in Table II. Except for cardiovascular effects the
other diseases were more prevalent in the exposed workers.
Smokers were more prevalent in the exposed group.
The health effects were compared with the adduct levels
(Table III) using the Mann--Whitney test. In the first analysis,
only the exposed workers were included in the calculations.
The adduct levels in the exposed people with cataracts were
significantly higher than in people without cataracts. The significance level for the difference of people with and without
cataracts increased by multiplying the adduct levels with age
or with the amount of years employed in the factory.
Hepatomegaly was significantly present at higher adduct
levels. For hepatitis B infected people, the adduct level
difference was not significant. For workers without hepatitis B
1275
G.Sabbioni et al.
Table III. Hb adduct levels of TNT (¼4ADNT), age and workyears of Chinese workers, compared with hepatomegaly, cataract and splenomegaly
Hepatomegaly
hepatitis, yes
Hepatomegaly
hepatitis, no
Hepatomegaly
all
Cataract
Splenomegaly
Exposed workers
Hb-4ADNT
Age
Workyears
Hb-4ADNTworkyearsd
Hb-4ADNTaged
54.6a, 61.4b, 0.88c
38.0, 39.9, 0.69
9.3, 9.9, 0.38
402.4, 875.7, 0.34
2101.2, 3022.1, 0.55
54.5, 106.6, 0.009
41.0, 41.9, 0.76
9.5, 10.8, 0.64
468.4, 1401.8, 0.018
2016.9, 4005.7, 0.004
54.5, 88.3, 0.02
41.0, 41.9, 0.66
9.5, 10.6, 0.33
431.7, 1015.7, 0.011
2031.9, 3721.9, 0.01
49.6, 68.9, 0.021
34.5, 43.1, 50.001
5.0, 11.0, 50.001
211.3, 1007.3, 50.001
1733.1, 3031.3, 50.001
55.6, 100.1, 0.023
41.0, 41.4, 0.56
9.0, 10.4, 0.48
418.1, 1006.3, 0.033
2167.7, 3751.1, 0.022
Exposed þ controls workers
Hb-4ADNT
Age
Workyears
52.6, 47.5, 0.74
35.4, 35.0, 0.54
7.5, 9.0, 0.37
36.9, 88.3, 0.015
42.8, 41.9, 0.97
10.6, 11.0, 0.96
42.0, 75.3, 0.018
41.0, 41.0, 0.72
10.0, 10.7, 0.61
20.7, 63.3, 50.001
34.9, 43.4, 50.001
6.4, 11.5, 0.001
42.8, 100.1, 0.002
41.0, 41.4, 0.55
10.0, 10.4, 0.81
Except for workers with hepatitis B the median Hb-adduct levels were significantly higher in workers with hepatomegaly, cataract and splenomegaly
(Mann--Whitney test). In case of hepatomegaly, the workers were split in a category with hepatitis B and without hepatitis B. A significant difference for
age and workyears was found for people with cataract. For workers with hepatitis B only splenomegaly is significantly prevalent (x2-test, data not shown).
a
Median level (ng 4ADNT/g Hb) of the workers without disease.
b
Median level (ng 4ADNT/g Hb) of the workers with the health effect.
c
Two sided significance level P.
d
Hb adduct levels multplied with workyears.
Table IV. Logistic regression analysis of Hb-adducts and disease in Chinese workers
Hb-adduct: 4ADNT
Hepatomegaly ORa (CIb, Pc)
Cataract OR (CI, P)
Splenomegaly OR (CI, P)
Exposed workers
7.6 (1.3--43.7, 0.024)
Exposed þ control workers
2.1 (0.96--4.4, 0.063)
6.4
25.6
2.9
8.3
9.6
12.8
11.1
13.7
(1.4--29.6, 0.018)
(2.8--232, 0.004)d
(1.6--5.3, 0.001)
(2.9--23.2, 50.001)f
(1.1--85.3, 0.042)
(1.3--125, 0.029)e
(1.5--83.8, 0.019)
(1.6--118, 0.017)e
a
OR ¼ odds ratio, bCI ¼ confidence interval, ctwo sided significance level.
The ORs were adjusted for: dage, ehepatitis B status and fage and workyears.
infection, the Hb-adduct levels were higher in workers with
hepatomegaly. In workers with cardiovascular defects, the
adduct levels were not significantly different. In order to
determine the confounding factors, the same analyses were
peformed with age, workyears, cigarettes per day and smoke
years. Only for age and workyears was there a difference
between people with disease. For workers with cataract, the
median age and median workyears were significantly higher.
For smoked cigarettes no such difference was seen. The same
analyses were performed including the control group in the
analyses. The results obtained were very similar (Table III).
Thus Hb-adduct levels are a good indicator for the biological
effect. The same statistical analyses were performed for the
urine metabolites. All these analyses were statistically not
significant P 4 0.35. Therefore, only Hb-adducts are a predictor for the health outcome.
In order to calculate the odds ratios (ORs), logistic regression analyses were performed with the data. The logistic
regression was performed using the log-transformed data of
Hb-cleavage products and urine metabolites. The obtained
ORs are listed in Table IV. The OR indicates the odds of
suffering for the various health effect of subjects with one
log-unit more of each adduct compared with the odds of subjects with one log-unit less. The results presented in Table IV
show that there is a significant dose--response relationship
between the Hb-cleavage product, 4ADNT and the risk of
suffering from cataract, hepatomegaly or splenomegaly. No
such relationship was found for the urine metabolites. The
odds of suffering from cataract were 6.4 times higher when
1276
the level of 4ADNT Hb-adducts increased by one log-unit
(P 5 0.05). Similar ORs were observed with hepatomegaly
(7.6) and splenomegaly (9.6). These results were tested for
confounding factors like age, workyears, smoker status, smoke
years, cigarettes per day and hepatitis B status using stepwise
forward logistic regression analysis. In the case of splenomegaly, hepatitis B status is a confounder. In the case of cataract,
age is a confounder. The hepatitis B adjusted OR for splenomegaly was 12.8 and the age adjusted OR for cataract was
25.6. All ORs have been listed in Table IV. These results
inferred that quantitation of TNT Hb-adducts provided an
effective biomarker of toxicity and could be used to estimate
the risk associated with a particular exposure to TNT.
The clinical blood parameters of the exposed male workers
were correlated with the Hb-adduct levels using the Spearman
rank order methods. There was a significant correlation
between the Hb-adduct levels and the Hb-concentration (r ¼
0.51, P 5 0.001), hematocrit (r ¼ 0.25, P 5 0.05) and
serum glutamic pyruvic transaminase (SGPT) levels (r ¼
0.25, P 5 0.05). The other parameters did not correlate. For
a further statistical analysis, the Hb-adduct levels of the
exposed male workers were split in two categories: values
below the median level and values above the median level.
These two categories were compared with the clinical blood
parameters using the Mann--Whitney test. Workers with a
higher adduct level have a significantly lower hematocrit,
red blood cell count level, Hb concentration and total blood
protein levels. In addition, the SGPT levels and the neutrophil
white blood count cell were significantly lower in workers with
Biomonitoring studies in TNT exposed workers
higher Hb-adduct levels. The changes in the other parameters
were not significant, although the biliriubin level, the
albumin--globulin ratio and the lymphocyte blood cell count
was higher in the group of workers with high adduct levels.
The other parameters did not differ between the two Hbadduct categories.
Discussion
Biomonitoring
For TNT, this is the first study combining different markers of
exposure with biological effects. The external dose, skin
exposure and air measurement was performed only on about
one-third of the workers (n ¼ 35). The data and experiments
are not included in this paper since the worker number is too
small to observe significant correlations with the health
effects. A Spearman rank analysis of the external dose with
urine metabolites and/or Hb-adducts yielded correlation coefficients 50.41. For skin exposure the correlation was moderately better (r 5 0.54).
In all the other previous studies about workers exposed to
TNT, urine metabolites or the Hb-adducts were measured
separately. In German workers dismantling old munition, the
median 4ADNT level found in raw urine was 0.23 mg/ml
(51 workers) (32) and 0.71 mg/ml (9 workers) (14). We found
similar 4ADNT levels (median ¼ 0.21 mg/ml) in raw urine of
exposed Chinese workers. Therefore, the same disease
noticed in Chinese workers could appear in other populations
also, if chronic exposure is given over the years. The Hbadduct levels found in the present study were similar to an
earlier study performed in a Chinese factory (25). In
Germany, Hb-adducts were found in environmentally
exposed people (33). Populations living on a formerly used
industrial area for the production of TNT were compared
with matched controls living in non-contaminated areas.
The median levels found in people of the contaminated
areas was 0.24 ng/g Hb 4ADNT. This is 262 times lower
than the median levels found in Chinese workers. The maximum value found was up to 3.4 ng/g Hb. In control population, the values were similar, 0.08 ng/g Hb (median) and
1.9 ng/g Hb (maximum). Unfortunately, no experimental
details were given for the determination of Hb-adducts.
Therefore, the levels found are difficult to discuss and the
method cannot be repeated.
Medical examination
The clinical blood parameters regarding the hematological
effects, decrease of hematocrit and Hb concentration, corresponded to earlier findings (1). The other changes observed in
the present study—hepatic, immunological and lymphoretical
effects—have not been observed in earlier studies in humans,
though observed in animal studies (1). In the present group of
workers, the SGPT levels decreased significantly with increasing levels of Hb-adduct. The same was observed in mice, rats
and dogs; in these animals, the SGPT levels decreased and the
liver size increased (34). The serum glutamic-oxaloacetic
transaminase remained the same in those animals. To monitor
immunological and lymphoretical effects, white blood cell
counts (lymphocytes, eosinophils and neutrophils) and
spleen size was measured in the present group of Chinese
workers. The amount of lymphocytes was higher in workers
exposed to higher doses as observed earlier (1). The significant
decrease in neutrophil white blood cell counts was not
observed in earlier studies. Splenomegaly has never
been reported in humans although reported in animals (1).
Hepatomegaly was found both in animals and humans in the
past. Cataract is a common disease among TNT workers.
Cancer risk
In a recent study, higher incidences of cancer were found in
TNT workers. How many additional cancer cases are expected
for people exposed to the levels of TNT found in the present
study? For a risk assessment, the daily dose of the Chinese
workers has to be estimated. With the knowledge from animal
data, it is possible to estimate the daily dose from the measured
Hb-adduct levels. However, it is necessary to make the following assumptions: (i) The steady-state level of Hb-adducts is
estimated with the following formula (35,36): steady-state
adduct level ¼ daily-adduct level 0.5 lifetime of the
erythrocytes. Thus, to calculate the single dose, the adduct
level has to be divided by 60. (ii) Modified Hb has the same
life span as unmodified Hb and the adducts are stable to repair
mechanisms. (iii) The pharmacokinetics of the xenobiotic
compound is comparable in rats and humans. For 4ABP the
percentage of the dose bound to Hb in rats was very similar to
the percentage of the dose bound to Hb of smokers (36),
although two additional cysteine groups are present in
the alpha-chain of the Hb of rats (37) but not in humans or
mice. The Hb-adduct levels for five arylamines were 2--28
times lower in mice than in rats (38). Therefore, the estimation
of the daily dose from Hb-adducts in rats could be an
underestimation.
The absorbed dose per day was calculated from the Hbadduct levels using the equation for the steady-state level of
Hb-adducts. With the Hb-adducts, the daily dose was estimated assuming that the same amount of the dose binds in
rats and in humans. In rats ~0.09% of the dose binds to Hb
(39). The median adduct levels of the workers was 62.1
(mean ¼ 93.9 ng/g Hb, 95th percentile ¼ 226.3 ng/g Hb).
The daily dose for the median Hb-adduct level was estimated
to be 14.1 mg/kg/day. An excess lifetime cancer risk (ELCR)
for exposure to TNT was estimated using the formula published by the US regulatory agencies (40,41) (http://risk.
lsd.ornl.gov): ELCR ¼ (cancer slope factor) (human dose).
Workers are not exposed 7 days a week and not all 52 weeks of
the year to the work place contaminant. Since we deduced the
external dose from the internal dose, days without exposure
were included. Therefore, the dose was not corrected for the
days without exposure. Since workers are not exposed for a
lifetime of 70 years to the occupational pollutants, but only
for 40 years of work life, the dose was corrected with the factor
40/70. The cancer slope factors for TNT is 0.03 (mg/kg/day)1
(www.epa.gov/iris). Therefore the excess cancer risk for
lifetime exposure is 23 in 100 000 for the median adduct levels
found in workers. For the 95th percentile adducts level the
estimated daily dose was 51.4 mg/kg/day. This dose translates
into an estimated lifetime cancer risk of 88 in 100 000. Thus,
for workers exposed to this dose only an excess of 0.09% more
cancers is expected. However, based on an approximation
using standard mortality rates, it was recently reported that
TNT-exposed Chinese workers have a higher relative risk of
2.3 for all cancers (8). The average age of death was 51.7 years
for the TNT-exposed workers and 54.1 years for the unexposed
controls in this study. We calculated a cumulative lifetime risk
of 7% for this particular age group (15--64 years) in China
using age-specific cancer mortality rates (42,43). Based on this
1277
G.Sabbioni et al.
calculation, a relative risk of 2.3 would yield a 9.1% excess of
cancers for the exposed workers in the earlier study (8). This
estimate exceeds our estimate. There are several possible
explanations for the discrepancy: (i) The dose estimated from
Hb-adducts is too low if less TNT binds to Hb in humans than
in rats. (ii) The cancer slope factor provided by the Environmental Protection Agency is inaccurate and too low for TNT.
(iii) The previous study from China does not differentiate
between an ammunition factory, where only TNT is used,
and other TNT factories where other intermediates are the
more prevalent chemicals in the work environment. 2,4Dinitrotoluene, 2,6-dinitrotoluene and 2-nitrotoluene are the
major occupational pollutants in most TNT factories, and
dinitrotoluenes have a cancer slope factor that is 23 times
larger than the factor for TNT. We previously found a 1.5%
excess of lifetime cancer in another TNT factory (44) using the
same approximation methods that are described herein. This
risk estimate is in the same order of magnitude as the estimate
deduced from the other study (8).
In conclusion, this is one of the rare biomonitoring studies
where a dose-relationship could be found between biological
effects and biomarkers in people exposed to occupational
pollutants. The findings of the present study provide a useful
tool for future studies involving the risk assessment of people
involved in the clean up of contaminated sites, for people
living in contaminated areas (www.atsdr.cdc.gov) (6,45) and
for workers of TNT factories.
Acknowledgements
We acknowledge the technical assistance of Renate Hartley and the financial
support by the European Commission, ERB-IC-CT97-0221.
Conflict of Interest Statement: None declared.
References
1. Agency for Toxic Substances and Disease Registry (ATSDR) (1995)
Toxicological Profile for 2,4,6-Trinitrotoluene. U.S. Department of Health
and Human Services, Public Health Service, Atlanta, GA.
2. U.S. Environmental Protection Agency. Integrated Risk Information
System (IRIS) (1991) Risk Estimate for Carcinogenicity and Reference
Dose for Oral Exposure for 2,4,6-Trinitrotoluene. Online file. Office of
Health and Environmental Assessment, Cincinnati, OH.
3. Rosenblatt,D.H., Burrows,E.P., Mitchell,W.R. and Parmer,D.L. (1991)
Organic explosives and related compounds. In Hutzinger,O. (ed.)
Handbook of Environmental Chemistry. Springer-Verlag Berlin,
Heidelberg, pp. 195--234.
4. Yinon,J. (1990) Toxicity and Metabolism of Explosives. CRC Press, Boca
Raton, Ann Arbor, Boston, pp. 1--67.
5. Hathaway,J.A. (1985) Subclinical effects of trinitrotoluene: a review of
epidemiological studies. In Rickert,D.E. (ed.) Toxicity of Nitroaromatic
Compounds. Hemisphere Publishing Corporation, Washington, DC,
pp. 255--274.
6. Feltes,J., Levsen,K., Volmer,D. and Spiekermann,M. (1990) Gas chromatographic and mass spectrometric determination of nitroaromatics in
water. J. Chromatogr., 518, 21--40.
7. Qu,G.C. (1984) Nation-wide investigation of occupational poisonings by
lead, benzene, mercury, organic phosphorus, and trinitrotoluene, with an
analytical study of their aetiology. J. Ind. Hyg. Occup. Dis., 2, 25--30 (in
Chinese).
8. Yan,C., Wang,Y., Xia,B., Li,L., Zhang,Y. and Liu,Y. (2002) The
retrospective survey of malignant tumor in weapon workers exposed to
2,4,6-trinitrotoluene. Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi,
20, 184--188 (in Chinese).
9. Kolb,G., Becker,N., Scheller,S., Zugmaier,G., Pralle,H., Wahrendorf,J.
and Havemann,K. (1993) Increased risk of acute myelogenous leukemia
(AML) and chronic myelogenous leukemia (CML) in a county of Hesse,
Germany. Soz. Praventivmed., 38, 190--195.
1278
10. Rickert,D.E. (1987) Metabolism of nitroaromatic compounds. Drug
Metab. Rev., 18, 23--53.
11. Yinon,J. and Hwang,D.G. (1985) Identification of urinary metabolites of
2,4,6-trinitrotoluene in rats by liquid chromatography--mass spectometry.
Toxicol. Lett., 26, 205--209.
12. Woollen,B.H., Hall,M.G., Craig,R. and Steel,G.T.B. (1986) Trinitrotoluene:
assessment of occupational absorption during manufacture of explosives.
Br. J. Ind. Med., 43, 465--473.
13. Ahlborg,G.Jr, Einist€o,P. and Sorsa,M. (1988) Mutagenic activity and
metabolites in the urine of workers exposed to trinitrotoluene (TNT). Br. J.
Ind. Med., 45, 353--358.
14. Bader,M., Goen,T., Muller,J. and Angerer,J. (1998) Analysis of nitroaromatic compounds in urine by gas chromatography--mass spectrometry for
the biological monitoring of explosives. J. Chromatogr. B, 710, 91--99.
15. van Welie,R.T.H., van Dijck,R.G.J.M., Vermeulen,N.P.E. and
van Sittert,N.J. (1992) Mercapturic acids, protein adducts, and DNA
adducts as biomarkers of electrophilic chemicals. Crit. Rev. Toxicol., 22,
271--306.
16. Neumann,H.G. (1984) Analysis of hemoglobin as a dose monitor for
alkylating and arylating agents. Arch. Toxicol., 56, 1--6.
17. Green,L.C., Skipper,P.L., Turesky,R.J., Bryant,M.S. and Tannenbaum,S.R.
(1984) In vivo dosimetry of 4-aminobiphenyl in rats via a cysteine adduct
in hemoglobin. Cancer Res., 44, 4254--4259.
18. Ringe,D., Turesky,R.J., Skipper,P.L. and Tannenbaum,S.R. (1988)
Structure of the single stable hemoglobin adduct formed by
4-aminobiphenyl in vivo. Chem. Res. Toxicol., 1, 22--24.
19. Kazanis,S. and McClelland,R.A. (1992) Electrophilic intermediate in the
reaction of glutathione and nitrosoarenes. J. Am. Chem. Soc., 114,
3052--3059.
20. Sabbioni,G. and Jones,C.R. (2002) Biomonitoring of arylamines and
nitroarenes. Biomarkers, 7, 347--421.
21. Liu,Y.Y., Lu,A.Y.H., Stearns,R.A. and Chiu,S.H. (1992) In vivo covalent
binding of [14C]trinitrotoluene to proteins in the rat. Chem. Biol. Interact.,
82, 1--19.
22. Leung,K.H., Yao,M., Stearns,R. and Chiu,S.H. (1995) Mechanism of
bioactivation and covalent binding of 2,4,6-trinitrotoluene. Chem. Biol.
Interact., 97, 37--51.
23. Sabbioni,G. (1994) Hemoglobin binding of nitroarenes and quantitative
structure activity relationships. Chem. Res. Toxicol., 7, 267--274.
24. Liu,Y.Y., Yao,M., Fang,J.L. and Wang,Y.W. (1995) Monitoring human
risk and exposure to trinitrotoluene (TNT) using hemoglobin adducts as
biomarkers. Toxicol. Lett., 77, 281--287.
25. Sabbioni,G., Wei,J. and Liu,Y.Y. (1996) Determination of hemoglobin
adducts in workers exposed to 2,4,6-trinitrotoluene. J. Chromatogr. B,
682, 243--248.
26. Biaglow,J.E., Jacobson,B., Greenstock,C.L. and Raleigh,J. (1977) Effect
of nitrobenzene derivatives on electron transfer in cellular and chemical
models. Mol. Pharmacol., 13, 269--282.
27. Kong,L.Y., Jiang,Q.G. and Qu,Q.S. (1989) Formation of superoxide by
trinitrotoluene in rat liver, brain, kidney, and testicle in vitro and monkey
liver in vivo. Biomed. Environ. Sci., 2, 72--77.
28. Zitting,A., Szumanska,G., Nickels,J. and Savolainen,H. (1982) Acute
toxic effects of trinitrotoluene in rat brain, liver and kidney; role of radical
production. Arch. Toxicol., 51, 53--64.
29. Spector,A. (1995) Oxidative stress-induced cataract: mechanism of action.
FASEB J., 9, 1173--1182.
30. Sabbioni,G. and Beyerbach,A. (1995) Determination of hemoglobin
adducts of arylamines in humans. J. Chromatogr. B, 667, 75--83.
31. Sabbioni,G. and Beyerbach,A. (2000) Hemoglobin adducts of aromatic
amines: diamines and polyaromatic amines. J. Chromatogr. B, 744,
377--387.
32. Letzel,S., Goen,T., Bader,M., Angerer,J. and Kraus,T. (2003) Exposure to
nitroaromatic explosives and health effects during disposal of military
waste. Occup. Environ. Med., 60, 483--488.
33. Ewers,U., Zwirner-Baier,I., Neumann,H.G., Futtig,E., SeurenKronenberg,K.
and
Luken,B.O.
(2000)
H€amoglobin-AdduktKonzentrationen nitroaromatischer Verbindungen im Blut von
Bewohnern von R€ustungsaltstandorten. 1. Studie Hessisch Lichtenau—
Hirschhagen. Umweltmed. Forsch. Prax., 5, 267--275.
34. Dilley,J.V., Tyson,C.A., Spanggord,R.J., Sasmore,D.P., Newell,G.W. and
Dacre,J.C. (1982) Short-term oral toxicity of 2,4,6-trinitrotoluene in mice,
rats and dogs. J. Toxicol. Environ. Health, 9, 565--585.
35. Ehrenberg,L., Hiesche,K.D., Osterman-Golkar,S. and Weinberg,I.
(1974) Evaluation of genetic risks of alkylating agents: tissue doses in
the mouse from air contaminated with ethylene oxide. Mutat. Res., 24,
83--103.
Biomonitoring studies in TNT exposed workers
36. Tannenbaum,S.R., Bryant,M.S., Skipper,P.L. and Maclure,M. (1986)
Hemoglobin adducts of tobacco-related aromatic amines: application to
molecular epidemiology. Banbury Report, 26, 63--75.
37. Maples,K.R., Eyer,P. and Mason,R.P. (1990) Aniline-phenylhydroxylamine-, nitrosobenzene- and nitrobenzene-induced hemoglobin thiyl
free radical formation in vivo and in vitro. Mol. Pharmacol., 37,
311--318.
38. Birner,G. and Neumann,H.G. (1988) Biomonitoring of aromatic amines II:
hemoglobin binding of some monocyclic aromatic amines. Arch. Toxicol.,
62, 110--115.
39. Zwirner-Baier,I., Kordowich,F.J. and Neumann,H.G. (1994) Hydrolyzable
hemoglobin adducts of polyfunctional monocyclic N-substituted arenes as
dosimeters of exposure and markers of metabolism. Environ. Health
Perspect., 102 (Suppl 6), 43--45.
40. U.S. Environmental Protection Agency (2000) Risk Characterization
Handbook, Science Policy council, EPA 100-B-00--002, U. S. Environmental Protection Agency, Washington, DC.
41. U.S. Department of Energy--Oak Ridge Operations (DOE--ORO)
(1999) Guidance for Conducting Risk Assessments and Related Risk
Activities for the DOE--ORO. Environmental Management Program, BJC/
OR-271.
42. Yang,L., Parkin,D.M., Li.,L. and Chen,Y. (2003) Time trends in
cancer mortality in China 1987--1999. Int. J. Cancer, 106,
771--783.
43. Day,N.E. (1992) Cancer incidence in five continents. Cumulative rate and
cumulative risk. IARC Sci. Pub., 120, 862--864.
44. Jones,C.R., Liu,Y.-Y., Sepai,O., Yan,H. and Sabbioni,G. (2005)
Hemoglobin adducts in workers exposed to nitrotoluenes. Carcinogenesis,
26, 133--143.
45. Gowik,P., Dieter,H.H. and Litz,N. (1994) Mess- und Vergleichswerte von
Nitro-/Aminoaromaten und weiterer Sprengstoff-typischer Verbindungen
im Boden, Sicker- und Grundwasser. WaBoLu Hefte 1/1994, Institut f€ur
Wasser-, Boden- und Lufthygiene des Bundesgesundheitsamtes, Berlin,
pp. 1--55.
Received December 21, 2004; revised March 17, 2005;
accepted March 22, 2005
1279