Content uploaded by Regina Santella
Author content
All content in this area was uploaded by Regina Santella
Content may be subject to copyright.
American Journal of Epidemiology
Copyright © 1998 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol.
147,
No.
3
Printed in U.S.A.
Recent Developments in Molecular Epidemiology
A Study of the Effects of Environmental Polycyclic Aromatic Hydrocarbons on Birth
Outcomes in Poland
Frederica P. Perera,1* Robin M. Whyatt,1* Wieslaw Jedrychowski,2 Virginia Rauh,1 David Manchester,3
Regina M. Santella,1 and Ruth Ottman1
The authors briefly review the current potential and limitations of molecular epidemiology. This approach
uses biomarkers to measure the internal and bioeffective dose of toxicants, early biologic effects likely to be
predictive of cancer, and variations in individual susceptibility. The most frequent application of biomarkers
has been in assessment of exposure/dose and susceptibility due to genetic and nongenetic factors. More
research is needed to establish the predictive significance of biomarkers in terms of disease risk. To illustrate
that molecular epidemiology has potential in identifying etiologic factors in disease, this article presents data
from a recent study of the developmental effects of fetal exposure to polycyclic aromatic hydrocarbons (PAH)
via ambient pollution. The study was carried out in an industrialized area of Poland with relatively high levels
of PAH pollution from coal burning. PAH-DNA adducts in leukocytes and plasma cotinine were measured in
umbilical cord blood as dosimeters of transplacental PAH and cigarette smoke, respectively. The study
subjects were 70 newborns from the industrialized city of Krakow and 90 newborns from Limanowa, a rural
town with far greater use of coal for home heating. Newborns whose levels of PAH-DNA adducts were above
the median (3.85/10s nucleotides) had a significantly decreased birth length, weight, and head circumference.
Cotinine was significantly inversely associated with birth weight and
length.
Although preliminary, these results
provide a new molecular link between PAH exposure and developmental effects, generating initial data and
hypotheses for further study. Am J Epidemiol 1998;147:309-14.
biological markers; environmental exposure; epidemiology, molecular; polycyclic hydrocarbons, aromatic;
pregnancy outcome
Since 1982, with the initial article setting out the
conceptual framework for integrating molecular mea-
surements into epidemiologic research on cancer (1),
molecular epidemiologic approaches have been ap-
plied to the study of many other diseases, including
reproductive and developmental disorders (2-4). The
1982 paper discussed the potential of measuring bio-
markers (molecular alterations or variants) in human
samples to document individual exposure or dose,
early biologic response or preclinical effect, and host
susceptibility to carcinogens. It also stressed the need
for validation of methods. Up to the present, the cat-
egorization of biomarkers has remained essentially
•These authors have contributed equally to the work.
Received for publication July 16,1997, and in final form October
16,
1997.
Abbreviations: PAH, polycyclic aromatic hydrocarbons; B(a)P,
benzo(a)pyrene.
1 School of Public Health, Columbia University, New York, NY.
2 College of Medicine, Jagiellonian University, Krakow, Poland.
3 University of Colorado Health Sciences Center, Denver, CO.
Reprint requests to Dr. Frederica Perera, Columbia University
School of Public Health, 60 Haven Avenue, B-1, New York, NY
10032.
unchanged; but the need for adequate validation of
molecular techniques prior to their use in epidemio-
logic studies or in screening has been more widely
recognized (3, 5, 6). During the past 15 years, molec-
ular epidemiology has been variously perceived as a
new field, an advanced form of clinical epidemiology,
or a subdiscipline of public health epidemiology (5, 7).
Although the label has stuck, and many schools of
public health and research institutes now have pro-
grams in "molecular epidemiology," the most com-
mon view is that the approach represents a natural
convergence of molecular biology and epidemiology
and that it should remain broad in scope—owned by
no one discipline—for the furtherance of disease pre-
vention.
There are three major areas of disease prevention to
which this approach can potentially contribute, work-
ing in conjunction with more established methods of
laboratory testing, monitoring, and epidemiology.
They are 1) the identification of etiologic factors and
disease mechanisms, 2) modeling of the distribution
of exposure and susceptibility within the population,
309
at Medical Library (National University of Singapore & affiliated Research Ins on July 15, 2011aje.oxfordjournals.orgDownloaded from
310 Perera, Whyatt, et al.
and 3) the design and monitoring of interventions
(5,
8).
Most of the research carried out to date has used
biomarkers to characterize exposure—or more
specif-
ically the internal or biologically effective dose of
toxicants—or to control for confounding effects. In-
creasingly, biomarkers are being used to investigate
the modulation of exposure and risk by genetic and
other susceptibility factors. These studies have raised a
host of methodological problems of a technological,
epidemiologic, and ethical nature (6). While a goal is
to substitute early preclinical response markers for
disease endpoints in etiologic studies and interven-
tions,
there are few cases where the predictive
signif-
icance of biomarkers has been established through the
needed prospective or nested case-control studies.
This remains a gap in the field.
The study described here illustrates the application
of molecular epidemiology to gain a better understand-
ing of the role of in utero environmental exposures
(tobacco smoke and air pollution) in developmental
impairment. It was intended to generate hypotheses for
further research in an area that has proven elusive. To
complement environmental monitoring and question-
naire data, biomarkers were used to estimate the indi-
vidual dose of toxicants to the fetus (9). The biochem-
ical and molecular dosimeters used to measure
transplacental exposure were cotinine in umbilical
cord blood, reflecting cigarette smoke exposure, and
DNA adducts in cord blood leukocytes formed by
PAH. The research illustrates both the advantages and
the limitations of molecular epidemiology.
Krakow, Poland, is an industrialized city with rela-
tively high levels of PAH in air pollution attributed to
multiple sources, especially coal burning for industrial
purposes and residential heating (10, 11). PAH expo-
sure can result from inhalation of polluted air and from
the diet (ingestion of grilled and smoked foods and
vegetables grown in contaminated areas). Pollution
levels are highest in the city center and decrease to-
ward the periphery. We estimate that in 1991, the year
preceding the birth of the newborns in this study, the
women living in Krakow were exposed to annual
average ambient concentrations of respirable particu-
lates ranging from 37 jug/m3 for the least exposed
group to 78 jug/m3 for the most exposed. The corre-
sponding concentrations of benzo(a)pyrene (B(a)P),
an indicator PAH, were estimated to be 7 ng/m3 and
15 ng/m3, representing approximately 0.02 percent of
particulate matter (12). During 1995, annual average
concentrations of B(a)P were estimated at 4.2 ng/m3
and 9.2 ng/m3 in the areas of Krakow with the lowest
and highest particulate levels, respectively.
Limanowa, Poland, is in a rural agricultural district
70 km southeast of Krakow. In Limanowa, levels of
respirable particulates were estimated to be two- and
fivefold lower than those in the least and most polluted
areas of Krakow, respectively (12). However, use of
coal stoves for indoor home heating (which can emit
substantial amounts of PAH (10)) was twice as fre-
quent in Limanowa as in Krakow (12).
Experimental bioassays have shown a number of
PAH to be transplacental carcinogens and to be asso-
ciated with adverse reproductive outcomes, including
decrements in fetal weight (13-16). For example, ad-
ministration of B(a)P or coal liquefaction products
containing high levels of PAH to pregnant rats caused
significant decreases in fetal weight (15, 16), with
effects capable of persisting into the postnatal period
(14).
PAH-DNA adducts represent the net effect of
exposure, absorption, activation, detoxification, and
repair and have been widely used to measure the
individual biologically effective dose of PAH (8, 17).
The adducts have an estimated half-life of approxi-
mately 9-23 weeks in leukocytes (18). Plasma coti-
nine is a validated internal dosimeter for tobacco
smoke (19).
The current study was part of an ongoing research
project evaluating multiple effects of environmental
PAH on Polish women and newborns (12). As was
reported previously, despite higher exposure to ambi-
ent PAH, mothers and newborns from Krakow did not
differ significantly from those in Limanowa with re-
spect to levels of PAH-DNA adducts in leukocytes,
possibly because of greater coal use for home heating
in Limanowa, after analyses controlled for dietary
sources of PAH (20). However, among Krakow sub-
jects for whom ambient exposure estimates were con-
sidered most reliable (those women not employed
away from home), there was a significant dose-
response relationship between estimated air pollution
concentrations at the women's residences and PAH-
DNA adduct levels in both maternal and newborn
leukocytes (20). Newborn adduct levels were somewhat
higher than paired maternal levels, despite experimental
evidence that transplacental PAH exposures are 10-fold
lower than paired maternal exposures. Here we have
examined associations between fetal development and
PAH-DNA adduct levels in the newborns as a dosimeter
of transplacental PAH. Differences in birth outcomes
between newborns from Krakow and Limanowa were
also investigated.
MATERIALS AND METHODS
Subjects
Subjects were enrolled during January-March 1992
and included 70 mothers and newborns from Krakow
Am J Epidemiol Vol. 147, No. 3, 1998
at Medical Library (National University of Singapore & affiliated Research Ins on July 15, 2011aje.oxfordjournals.orgDownloaded from
PAH Exposure and Birth Outcomes 311
and 90 mothers and newborns from Limanowa. En-
rollment alternated biweekly between Krakow and
Limanowa to control for monthly variation in ambient
pollutant levels, and was restricted to vaginal deliver-
ies and to women who had resided in Krakow or
Limanowa for at least 1 year. Detailed questionnaire
and medical record data were collected on environ-
mental exposures and known or potential risk factors
for developmental impairment (12). A sample of um-
bilical cord blood was collected at delivery (20-60
ml) and was processed and stored as described previ-
ously (12).
Laboratory assays
DNA was extracted from umbilical cord leukocytes,
and PAH-DNA adduct levels were measured by a
competitive enzyme-linked immunosorbent assay with
fluorescence endpoint detection, essentially as de-
scribed previously (21). The antiserum recognizes
multiple, structurally related PAH diol-epoxide-DNA
adducts (22). Values are expressed as the amount of
B(a)P diol-epoxide-DNA that would cause similar
inhibition in the assay. The quantity of DNA was
adequate to measure PAH-DNA adducts in 135 um-
bilical cord blood samples (58 from Krakow and 77
from Limanowa). Plasma cotinine was measured in
cord blood samples of all newborns by gas chroma-
tography, as described elsewhere (12). The estimated
half-life of plasma cotinine in nonsmoking adults is 48
hours;
in the chronic exposure situation, this marker is
a good reflection of daily uptake of nicotine (23, 24).
Statistical analysis
Associations between the outcome variables (birth
weight, birth length, and head circumference) and
newborn PAH-DNA adduct levels were analyzed by
multiple linear regression and analysis of covariance
in the Krakow and Limanowa groups, separately and
combined. Levels of PAH-DNA adducts were analyzed
both as a dichotomous variable—high (>median)
versus low (^median)—and as a continuous measure.
The regression models included covariates represent-
ing known or potential confounders of fetal develop-
ment that were correlated with the birth outcomes
{p < 0.2) and/or varied significantly between new-
borns from Krakow or Limanowa or with high versus
low levels of adducts. Models controlled for maternal
height, age, socioeconomic status (maternal educa-
tional level of ^11 years vs. >11 years), history of
low birth weight (yes/no), gestational age (weeks), sex
of the newborn, and whether or not the mother drank
alcohol regularly during the pregnancy (>1 drink/
week during ^1 trimester). Newborn plasma cotinine
level (ng/ml) was included rather than self-reported
maternal smoking status, since the biomarker ex-
plained more of the variance in fetal development.
When regression analyses were conducted controlling
for maternal prepregnancy body mass index rather
than maternal height, results were essentially un-
changed from those reported here.
RESULTS
Birth outcomes and demographic data for Krakow
and Limanowa subjects are presented in table 1. Com-
pared with newborns from Limanowa, Krakow new-
borns had a lower mean birth weight and a signifi-
cantly lower birth length and head circumference,
before (table 1) and after data were controlled for
maternal height, age, socioeconomic status, history of
low birth weight, alcohol consumption, gestational
age,
sex of the newborn, and plasma cotinine.
Table 2 presents differences in birth outcomes for
newborns with high versus low levels of PAH-DNA
adducts. In both the Krakow and Limanowa groups,
analyzed separately, all three measures (birth weight,
birth length, and head circumference) were decreased
in newborns with high adduct levels, although the
differences were not significant in all cases. Among
Krakow newborns, those with elevated PAH-DNA
adduct levels had a significantly decreased birth length
(by 1.8 cm, p
—
0.02) and head circumference (by 0.9
cm, p = 0.05). Among Limanowa newborns, those
TABLE 1. Distribution of 160 newborns from two Polish
cities according to birth outcomes and demographic
variables, 1992
No.
of mother-newborn pairs
Newborn birth weight
(g)
Newborn birth length
(cm)
Newborn head circumference
(cm)
Maternal age (years)
Newborn plasma cotinine level
(ng/ml),
by
maternal smoking
status
Current smoker
Ex-smoker
Nonsmoker
No.
of
mothers with £11 years of
education (socioeconomic status)
Maternal height
(cm)
Maternal prepregnancy weight
(kg)
Maternal prepregnancy weight
gain
(kg)
No.
of nulDparous mothers
No.
of mothers with
a
prior history
of low birth weight (<2,500
g)
Gestational age (weeks)
Sex
of
newborn
(%
female)
Maternal alcohol consumption (no.
who consumed £1 drink/week
in
ai trimester)
Krakow
70
3,328
(516)»
54.0 (3.0)t
34.1 (1.5)t
27.6 (5.3)t
75.0 (n =
12)t
15.6 (n = 20)
3.7 (n = 38)
23t
164.9 (5.0)t
59.8(8.8)
15.5 (15.1)
31
4
39.8(1.5)
46
12
Umanowa
90
3,378
(464)
55.5 (2.9)
34.7(1.6)
25.4 (4.1)
63.7 (n = 4)
20.1 (n=18)
4.1 (n = 68)
56
162.5 (5.5)
60.1 (10.3)
13.0 (10.4)
41
4
39.7(1.4)
51
11
*
Numbers In parentheses, standard deviation,
t
p £
0.02 for Krakow
vs.
Limanowa (Student's
f test
for birth length, head
circumference, age, and height;
%2
for smoking status and educational level).
t Number
of
smokers.
Am J Epidemiol Vol. 147, No. 3, 1998
at Medical Library (National University of Singapore & affiliated Research Ins on July 15, 2011aje.oxfordjournals.orgDownloaded from
312 Perera, Whyatt, et al.
TABLE 2. Difference in birth outcomes for Polish newboms with high (>median) versus low (^median)
leukocyte levels of PAH*-DNA adducts, 1992f
Krakow newboms (n = 58)
Limanowa newboms (n = 77)
All newboms
Birth weight
(g)
Difference
-205
-129
-147
P
value
0.11
0.16
0.05
Birth length
(cm)
Difference
-1.8
-0.8
-1.1
P
value
0.02
0.17
0.02
Head
circumference
(cm)
Difference
-0.9
-1.2
-0.9
P
value
0.05
0.0004
0.0005
* PAH, polycyclic aromatic hydrocarbons.
t Analyses controlled for maternal height, age, socioeconomic status (educational level), history of low birth
weight, maternal alcohol consumption, gestatjonal age (weeks), sex of the newborn, and plasma cotinine (ng/ml).
Median adduct level, 3.85/10» nudeotides.
with elevated adduct levels had a significantly de-
creased head circumference (by 1.2 cm, p = 0.0004).
Considering all of the newboms combined (Krakow
and Limanowa), those with high levels of PAH-DNA
adducts had a significantly lower birth weight (p =
0.05), birth length (p = 0.02), and head circumference
(p = 0.0005) than newboms with low adduct levels
(table 2).
After combining the two groups and removing cur-
rent smokers in order to investigate the effects of PAH
in the absence of active smoking by the mother, we
found that all three measures of fetal development
remained significantly lower among those with high
adduct levels versus low levels (p ^ 0.03). To remove
the effect of both active and passive smoking, we
further restricted the analyses to the 53 newboms of
nonsmokers without detectable plasma cotinine. The
newboms with elevated adduct levels had a decreased
birth weight (by 217 g, p = 0.1) and length (by 0.8 cm,
p > 0.2) and a significantly decreased head circum-
ference (by 1.3 cm, p = 0.006).
When levels of PAH-DNA adducts among all of the
newboms were included in the regression models as a
continuous variable, they were inversely, though not
significantly, correlated with birth weight and length.
Adduct levels were inversely correlated with head
circumference both before {p = 0.006) and after data
were controlled for birth weight (p - 0.003), suggest-
ing asymmetrical growth retardation.
Consistent with prior research (25), an inverse cor-
relation was seen between newborn plasma cotinine level
(ng/ml) and birth weight {p = 0.0001, after controlling
for place of residence and the other potential con-
founders). In this study, cotinine was also inversely
associated with birth length (/? = 0.003). The effect of
cotinine on birth outcomes was independent of that of
PAH-DNA adducts.
DISCUSSION
These findings suggest that transplacental exposure
to PAH from multiple environmental sources may
compromise fetal development. They are consistent
with previous ecologic data indicating that birth out-
comes are worse in industrialized, contaminated re-
gions of Poland (26) and with experimental studies
indicating that certain PAH are developmental toxi-
cants (13, 15).
The results do not imply that developmental damage
is necessarily mediated by DNA binding, although that
is one possible mechanism. Here the extent of DNA
binding by PAH in newborn leukocytes was used as a
dosimeter of PAH that have reached the fetus. Neither
the mechanisms by which PAH exert developmental
toxicity nor the target sites have been identified. In-
deed, it is possible that PAH act by more than one
mechanism. For example, it has been hypothesized
that B(a)P exposure may interfere with uterine growth
during pregnancy because of its antiestrogenic effects,
thereby disrupting the endocrine system (15). Similar
to polychlorinated biphenyls, which are associated
with deficits in fetal growth and intelligence quotient
(27,28),
PAH bind to the human Ah receptor to induce
P450 enzymes (29). Additionally, the developing cen-
tral nervous system appears to be particularly sensitive
to DNA-damaging agents (30), and it may respond by
activating apoptotic pathways (31, 32). For example,
in humans, fetal microcephaly has been seen following
exposure to ionizing radiation (33) and anticonvulsant
drugs (34). Risk from anticonvulsants was most pro-
nounced in infants deficient in enzymes that detoxify
the DNA-binding intermediate (35, 36). It is of interest
that, in the present study, PAH-DNA adduct levels
were most strongly associated with reduction in head
circumference; and the data were suggestive of asym-
metrical growth retardation related to DNA binding.
The growth retardation associated in this study with
high levels of PAH-DNA adducts (147 g in weight, 1.1
cm in length, and 0.9 cm in head circumference) is in
the range of that which has been linked to adverse
human health sequelae of exposure to other develop-
mental toxicants, including maternal smoking, co-
caine, and polychlorinated biphenyls (19, 27, 28, 37).
Am J
Epidemiol
Vol. 147, No. 3, 1998
at Medical Library (National University of Singapore & affiliated Research Ins on July 15, 2011aje.oxfordjournals.orgDownloaded from
PAH Exposure and Birth Outcomes 313
Weight at birth is a predominant determinant of infant
health, and low birth weight is a major cause of infant
mortality (19). Head circumference has been corre-
lated with brain size, intelligence quotient, and cogni-
tive function (38). A reduction in head circumference
of 1-2 cm at birth has been associated with reduced
mental and psychomotor development in early child-
hood (37, 39).
In the present study, the combination of known risk
factors and PAH-DNA adducts does not explain the
observed difference in birth outcomes between Kra-
kow and Limanowa. It is possible that unmeasured
factors of urbanization, including perhaps constituents
of air pollution other than PAH, are contributing to
developmental deficits in Krakow. For example, in
northern Bohemia (Czech Republic), airborne sulfur
and nitrogen oxides have been associated with low
birth weight and small size for gestational age (40),
while in China, an association was seen with sulfur
oxides and particulates (41).
Strengths of this study include the detailed question-
naire monitoring data, the ability of biomarkers to
measure individual doses of PAH and tobacco smoke,
and the control of known risk factors associated with
fetal development, including smoking and alcohol
drinking by the mother. Limitations include the one-
time-only measurement of biomarkers, the lack of
personal monitoring of the women during pregnancy,
and our inability to draw conclusions about mecha-
nisms. Although we were unable to evaluate the rela-
tion of the molecular dose of PAH to leukocytes
versus the molecular dose of PAH to target tissue,
prior research indicates that, in humans, PAH are
distributed systemically and form adducts in numerous
tissues, including those in the developing fetus (42,
43).
We were also unable to evaluate the effect of
timing of exposure on the associations observed in this
study. Given their estimated half-life in leukocytes of
9-23 weeks (18), PAH-DNA adducts measured in
newborns will largely reflect exposures incurred dur-
ing the latter half of pregnancy.
To our knowledge, this is the first molecular epide-
miologic evidence that transplacental PAH adversely
-affect-fetaldevelopment. However, the-molecular link
seen here does not prove causality and should be
considered preliminary. There exists the potential for
confounding by other unmeasured constituents of air
pollution and diet. Further study is warranted by the
potential implications for public health, given wide-
spread exposure to PAH during pregnancy and earlier
results from this study (20) indicating that infants
receive a high biologic dose of PAH relative to moth-
ers.
PAH levels comparable to those seen in Poland
are found in other industrialized areas of the world, but
in most Western countries they are considerably
lower. Although prior data do not indicate a threshold
for PAH-DNA adduct formation (10), the effect of
lower levels of PAH on human fetal development is
unknown.
In conclusion, although the results are preliminary
and require confirmation, the study illustrates the po-
tential benefits of using biomarkers in the identifica-
tion of environmental etiologic factors. They support
continued commitment to the development and use of
this approach in disease prevention.
ACKNOWLEDGMENTS
This work was supported by grants from the National
Institutes of Health, the National Institute of Environmental
Health Sciences, the US Department of Energy, the March
of Dimes Birth Defects Foundation, the Lucille P. Markey
Charitable Trust, the Colette Chuda Environmental Fund,
and the Gladys and Roland Harriman Foundation.
The authors acknowledge M. C. Randall, Dr. T. L.
Young, T. B. Cooper, Dr. D. Tang, A. Rundle, Y. Hsu, and
R. Lonow for their assistance in the Polish study. The
authors also thank Dr. I. Bernard Weinstein for his invalu-
able contribution to the conceptual framework of molecular
epidemiology.
REFERENCES
1.
Perera FP, Weinstein IB. Molecular epidemiology and
carcinogen-DNA adduct detection: new approaches to studies
of human cancer causation. J Chronic Dis 1982;35:581-600.
2.
Hulka BS. Epidemiological studies using biological markers:
issues for epidemiologists. Cancer Epidemiol Biomarkers
Prev 1991;1:13-19.
3.
Schulte PA, Perera FP, eds. Molecular epidemiology: princi-
ples and practices. New York, NY: Academic Press, Inc,
1993.
4.
National Research Council, Commission on Life Sciences,
Board on Environmental Studies and Toxicology. Biologic
markers in reproductive toxicology. Washington, DC: Na-
tional Academy Press, National Academy of Sciences, 1989.
5.
McMichael AJ. Invited commentary—"molecular epidemi-
ology": new pathway or new travelling companion? Am J
Epidemiol 1994;140:l-ll.
6. Vineis P. Use of biomarkers in epidemiology: the example of
metabolic susceptibility to cancer. Toxicol Lett 1995;77:
163-8.
~ '
7.
Shields PG, Harris CC. Molecular epidemiology and the ge-
netics of environmental cancer. JAMA 1991;266:681-7.
8. Perera FP. Molecular epidemiology: insights into cancer sus-
ceptibility, risk assessment, and prevention. J Natl Cancer Inst
1996;88:496-509.
9. Whyatt RM, Perera FP. Application of biologic markers to
studies of environmental risks in children and the developing
fetus.
Environ Health Perspect 1995;103(suppl 6):105-10.
10.
Perera FP, Hemminki K, Grzybowska E, et al. Molecular and
genetic damage from environmental pollution in Poland. Na-
ture 1992,360:256-8.
11.
Jedrychowski W, Becher H, Wahrendorf J, et al. A case-
control study of lung cancer with special reference to the
Am J Epidemiol Vol. 147, No. 3, 1998
at Medical Library (National University of Singapore & affiliated Research Ins on July 15, 2011aje.oxfordjournals.orgDownloaded from
314 Perera, Whyatt, et al.
effect of air pollution in Poland. J Epidemiol Community
Health 1990;44:114-20.
12.
Whyatt RM, Garte SJ, Cosma G, et al. CYP1A1 messenger
RNA levels in placental tissue as a biomarker of environmen-
tal exposure. Cancer Epidemiol Biomarkers Prev 1995;4:
147-53.
13.
Agency for Toxic Substances and Disease Registry, Environ-
mental Protection Agency, and ICF-Clement. Toxicological
profile of benzo(a)pyrene. Oak Ridge, TN: Oak Ridge Na-
tional Laboratory, 1990.
14.
MacKenzie KM, Angevine DM. Infertility in mice exposed in
utero to benzo(a)pyrene. Biol Reprod 1981;24:183—91.
15.
Bui QQ, Tran MB, West WL. A comparative study of the
reproductive effects of methadone and benzo[a]pyrene in the
pregnant and pseudopregnant rat. Toxicology 1986;42:
195-204.
16.
McKee RH, Pasternak SJ, Traul KA. Developmental toxicity
of EDS recycle solvent and fuel oil. Toxicology 1987;46:
205-15.
17.
Harris CC. Chemical and physical carcinogenesis: advances
and perspectives for the 1990s. Cancer Res 1991;51(suppl):
5023S-44S.
18.
Mooney LA, Santella RM, Covey L, et al. Decline in DNA
damage and other biomarkers in peripheral blood following
smoking cessation. Cancer Epidemiol Biomarkers Prev 1995;
4:627-34.
19.
Office of the Surgeon General, US Public Health Service.
Health benefits of smoking cessation: a report of the Surgeon
General. Washington, DC: US Public Health Service, 1990.
(DHHS publication no. (CDC) 90-8416).
20.
Whyatt R, Santella R, Garte SJ, et al. Effects of environmental
exposures on DNA damage in women and the developing
fetus and its modulation by genetic/metabolic factors. (Ab-
stract).
Proc Am Assoc Cancer Res 1995;36:284.
21.
Perera FP, Hemminki K, Young TL, et al. Detection of poly-
cyclic aromatic hydrocarbon-DNA adducts in white blood
cells of foundry workers. Cancer Res 1988;48:2288-91.
22.
Santella RM, Gasparo F, Hsieh LL. Quantitation of
carcinogen-DNA adducts with monoclonal antibodies. Prog
Exp Tumor Res 1987;31:63-75.
23.
Kemmeren JM, van Poppel G, Verhoef P, et al. Plasma
cotinine: stability in smokers and validation of self-reported
smoke exposure in nonsmokers. Environ Res 1994;66:
235-43.
24.
Sepkovic DW, Haley NJ, Hoffmann D. Elimination from the
body of tobacco products by smokers and passive smokers.
(Letter). JAMA 1986;256:863.
25.
Eskenazi B, Prehn AW, Christianson RE. Passive and active
maternal smoking as measured by serum cotinine: the effect
on birthweight. Am J Public Health 1995,85:395-8.
26.
Siniarska A, Antoszewska A, Dziewiecki C. Urbanization and
industrialization versus biological status of human popula-
tions.
Stud Hum Ecol
1992;
10:335-58.
27.
Fein GG, Jacobson JL, Jacobson SW, et al. Prenatal exposure
to polychlorinated biphenyls: effects on birth size and gesta-
tional age. J Pediatr
1984;
105:315-20.
28.
Jacobson JL, Jacobson SW. Intellectual impairment in chil-
dren exposed to polychlorinated biphenyls in utero. N Engl
J Med 1996;335:783-9.
29.
Manchester DK, Gordon SK, Golas CL, et al. Ah receptor in
human placenta: stabilization by molybdate and characteriza-
tion of binding of 2,3,7,8-tetrachlorodibenzo-/?-dioxin, 3-
methylcholanthrene, and benzo(a)pyrene. Cancer Res 1987;
47:4861-8.
30.
Wood KA, Youle RJ. The role of free radicals and p53 in
neuron apoptosis in vivo. J Neurosci 1995;15:5851-7.
31.
Meyn MS. Ataxia-telangiectasia and cellular responses to
DNA damage. Cancer Res 1995;55:5991-6001.
32.
Nicol CJ, Harrison ML, Laposa RR, et al. A teratologic
suppressor role for p53 in benzo[a]pyrene-treated transgenic
p53-deficient mice. Nat Genet 1995;10:181-7.
33.
Blot WJ, Miller RW. Mental retardation following in utero
exposure to the atomic bombs of Hiroshima and Nagasaki.
Pediatr Radiol
1973;
196:617-20.
34.
Hiilesmaa VK, Teramo K, Granstrom ML, et al. Fetal head
growth retardation associated with maternal antiepileptic
drugs.
Lancet
1981
;2:165-7.
35.
Strickler SM, Dansky LV, Miller MA, et al. Genetic predis-
position to phenytoin-induced birth defects. Lancet 1985;2:
746-9.
36.
Buehler BA, Delimont D, van Waes M, et al. Prenatal predic-
tion of risk of the fetal hydantoin syndrome. N Engl J Med
1990;322:1567-72.
37.
Chasnoff
IJ,
Griffith DR, Freier C, et al. Cocaine/polydrug use
in pregnancy: two-year follow-up. Pediatrics 1992;89:284-9.
38.
Ratcliffe SG, Masera N, Pan H, et al. Head circumference and
IQ of children with sex chromosome abnormalities. Dev Med
Child Neurol 1994;36:533-44.
39.
Azuma SD, Chasnoff IJ. Outcome of children prenatally ex-
posed to cocaine and other
drugs:
a path analysis of three-year
data. Pediatrics 1993;92:396-402.
40.
Sram RJ, Dejmek JI, Topinka J, et al. Impact of environmental
pollution on human health in Northern Bohemia. In: Carcino-
genesis from environmental pollution: assessment of human
risk and strategies for prevention. Proceedings of the Joint
Conference of the American Association for Cancer Research,
the International Agency for Research on Cancer, and the
Hungarian Cancer Society. Philadelphia, PA: American As-
sociation for Cancer Research, 1996.
41.
Wang F, Ding H, Ryan L, et al. Association between air
pollution and low birth weight: a community-based study.
Environ Health Perspect 1997;105:514-20.
42.
Shamsuddin AK, Gan R. Immunocytochemical localization of
benzo(a)pyrene-DNA adducts in human tissue. Hum Pathol
1988;19:309-15.
43.
Hatch MC, Warburton D, Santella RM. Polycyclic aromatic
hydrocarbon-DNA adducts in spontaneously aborted fetal tis-
sue.
Carcinogenesis 1990;ll:1673-5.
Am J Epidemiol Vol. 147, No. 3, 1998
at Medical Library (National University of Singapore & affiliated Research Ins on July 15, 2011aje.oxfordjournals.orgDownloaded from