TOXICOLOGICAL SCIENCES 98(2), 332–347 (2007)
doi:10.1093/toxsci/kfm008
Advance Access publication January 25, 2007
FORUM
Human Health and Endocrine Disruption: A Simple Multicriteria
Framework for the Qualitative Assessment of End Point–Specific
Risks in a Context of Scientific Uncertainty
Olwenn V. Martin,*,† John N. Lester,‡ Nikolaos Voulvoulis,* and Alan R. Boobis†,1
*Centre for Environmental Policy, Imperial College London, London SW7 2AZ, UK; †Experimental Medicine, Toxicology & Gastroenterology
Group, Faculty of Medicine, Imperial College London, London W12 0NN, UK; and ‡School of Water Sciences, Cranfield University, Cranfield,
Bedforshire MK43 0AL, UK
Received October 18, 2006; accepted January 10, 2007
INTRODUCTION
Endocrine disruption remains one of the most controversial
contemporary environmental issues. While the desired level of
protection is ultimately a societal choice, endocrine toxicity could
result in a wide spectrum of adverse health effects. Although the
application of the causal framework of weight-of-evidence approaches to complex toxicological issues has incited much interest,
no international criteria or guidance have yet been developed. In
this context, the evidence on end point–specific risks to human
health contained in the International Program on Chemical Safety
Global assessment of the State-of-Science on Endocrine Disruptors report was updated and assessed qualitatively using three
simple criteria relevant to the practical application of the precautionary principle (PP): incidence trends, association, and
consequence. The current degree of knowledge was then ranked
according to ignorance, uncertainty, and risk. The main sources of
scientific uncertainty in relation to incidence trends were associated with the evolution of diagnostic criteria or diagnostic tests,
while genetic susceptibility is often proposed as an explanation for
the wide geographic variations in the incidence of some diseases.
Such genetic polymorphisms are also offered as a potential
explanation for some of the inconsistent findings or lack of clear
dose-response gradients described under the association criterion.
The methodology yielded a relative paucity of data addressing
directly the impact for adverse human health effect from both
individual and public health perspectives. Results are discussed
within the context of the application of the PP. Within a participatory context, this simple framework could provide a useful
decision-making tool to both communicate scientific uncertainty
to the wider public and manage uncertain risks.
Key Words: endocrine toxicology—endocrine disruptors; risk
assessment–regulatory policy; risk assessment—epidemiology.
1
To whom correspondence should be addressed at Experimental Medicine,
Toxicology & Gastroenterology Group, Faculty of Medicine, 9S10B Commonwealth Building, Hammersmith Campus, Du Cane Road, London W12
0NN, UK. Fax: þ44 (0)20 8383 2066. E-mail: [email protected].
Ten years after the Weybridge European Workshop defined
endocrine disrupters as ‘‘an exogenous substance that causes
adverse health effects in an intact organism, or its progeny,
subsequent to changes in endocrine function’’ (Birkett and
Lester, 2003), endocrine disruption (ED) remains one of the
most controversial contemporary environmental issues. The
interaction of multiple agents with the host’s endocrine system
could result in a wide spectrum of potential adverse health
effects, and much scientific uncertainty remains regarding the
risks from endocrine-disrupting compounds (EDCs). News
reports about the potential effects of ‘‘gender-bending chemicals’’ on unborn male fetuses are now regularly being printed
in the general media, and scientific uncertainty is a well-known
‘‘dread factor’’ increasing the public’s perception of risk
(Slovic, 1987). If the level of acceptable risk is ultimately
a societal choice, such choice should nonetheless be informed
by science while precautionary actions (PAs) also need to be
justified by a sufficiency of evidence (Gee, 2006). Moreover,
the quantitative risk assessment of EDCs will necessarily be
end point specific (Chen, 2001). There is therefore a need to
find effective ways of assessing and communicating scientific
evidence and identifying disease end points for which evidence
may sanction a quantitative assessment of risks.
Although the concept of ED first emerged when it was
observed that some environmental chemicals were able to
mimic the action of the sex hormones (estrogens and androgens), some chemicals were subsequently shown to be able to
block such actions, and a variety of other mechanisms involved
in the control of circulating hormone levels that could be
disrupted by exogenous chemicals have now been identified.
Although EDCs are often thought of as referring only to
synthetic chemicals, natural agents occurring in the diet such
as phytoestrogens or mycoestrogens are also known to be
endocrinally active.
Ó The Author 2007. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.
For Permissions, please email: [email protected]
HUMAN HEALTH AND ENDOCRINE DISRUPTION
Scientific evidence on associations between exogenous
agents and health effects is traditionally derived from epidemiological and toxicological studies. Both types of methods
have their respective advantages and disadvantages. Exposure
assessment is one of the most significant limitations of
epidemiological studies as it often relies on questionnaires
and is therefore particularly prone to recall bias and potential
over- or underestimation of effects. If industrial exposures can
help assessing risks from many exogenous chemicals, the
‘‘healthy worker effect’’ can also obscure interpretation.
Although the validation of toxicological testing methods for
ED end points is well underway (Owens and Koeter, 2003),
much controversy remains over the lack of reproducibility of
observed low-dose effects (Witorsch, 2002). Interspecies
extrapolation is a source of uncertainty, particularly in the
case of developmental toxicology as the timing of developmental events may substantially differ between species and an
adverse response can result from an effect on the mother, the
fetus itself, or placental transfer (Hogan et al., 1987). In recent
years, rapid advances of biochemical sciences and technology
have resulted in the development of bioassay techniques that
can contribute invaluable information regarding toxicity mechanisms at the cellular and molecular level. However, the
extrapolation of such information to predict effects in an intact
organism for the purpose of risk assessment is still in its
infancy (Gundert -Remy et al., 2005).
Even when all these limitations have been taken into
account, the population samples for both epidemiological and
toxicological studies may be unrepresentative just by chance.
The chance effects of sampling variability are routinely
accounted for in statistical analysis and hypothesis testing. It
is important to note that failure to reject the null hypothesis
(i.e., no association) does not equate with its acceptance but
only allows reasonable confidence that if any association exists,
it would be smaller than an effect size determined by the power
of the study. The issues surrounding power and effect size
should normally be addressed at the design stage of a study,
although this is rarely reported (Rushton, 2000).
Multicriteria frameworks are particularly useful in environmental decision problems where the temporal scale adds to the uncertainty and intrinsic complexity of the issue (Balasubramaniam
and Voulvoulis, 2005) and have also been found to help achieve
overall consensus in a group of decision-makers or stakeholders (Le Gales and Moatti, 1990; Sell and Scholz, 2003). In
the context of the application of the precautionary principle
(PP) to the hazards posed by EDCs, different states of knowledge
justify different types of PAs (EEA, 2001). Furthermore, the Rio
Declaration produced in 1992 by the United Nations Conference
on Environment and Development clearly states that such actions
should be triggered by ‘‘threats of serious or irreversible damage’’
and introduces dimensions of the severity and temporal scale of
the consequences of a potential association. Moreover, the
interest incited by the weight-of-evidence approach originally
developed in the International Program on Chemical Safety
333
(IPCS) Global Assessment of the State-of-the-Science on
Endocrine Disruptors (International Program on Chemical
Safety, 2002) has highlighted the need to develop criteria and
methods for the application of such frameworks to complex
toxicological issues and the communication of such results
(International Programme on Chemical Safety, 2006). Since
1965, evaluations of the association between environment and
disease have often been based on the nine ‘‘Bradford Hill
criteria’’ (Hill, 1965). Drawing on these aspects of association
and consequence thereof, a simple multicriteria framework by
which to assess the current state of knowledge used to justify PAs
has been developed. The methodology applied to update the
scientific evidence may be more likely to yield evidence of effects
than the lack thereof, and therefore prone to publication bias.
Additionally, it introduces a subjective dimension by focusing on
evidence deemed to add to current knowledge and as such cannot
be qualified as a systematic or exhaustive review, but rather
attempts to demonstrate the use of the framework as a decisionmaking tool.
MATERIALS AND METHODS
Literature
Chapter 5 of the IPCS Global Assessment of EDCs (International Program
on Chemical Safety, 2002) gave a detailed summary of the state of science on
an end point–by–end point basis under four main headings: reproductive health,
neurobehavior, immune system, and cancer. For the purpose of this study, this
was considered an accurate account of the state of science on the potential
human health effects of environmental exposure to EDCs up to a year prior to
its publication, 2001. In order to update such evidence, relevant literature
published in the last 5 years has been reviewed. The literature search was
carried out in PubMed with an Entrez date limit of 5 years and using search
algorithms combining the specific health end point considered with a word such
as ‘‘endocrine,’’ ‘‘risk,’’ ‘‘incidence,’’ ‘‘trend,’’ or ‘‘environment.’’ The updated
evidence was then summarized under the three criteria described below for each
health end point.
Criteria
Incidence trends. The first criterion was defined as ‘‘incidence trends’’ and
relates to evidence derived from ecologic studies, in the epidemiological sense
of the term, that is, the unit of observation is the population rather than the
individual. The relation between disease rate and exposure in each of a series of
populations can offer insights into two of the features of association as defined
by Bradford Hill. Geographical studies and migrant studies, as well as
providing some insights into the consistency of observations, offer a way of
discriminating genetic from environmental causes of geographical variation in
disease and may also indicate the age at which an environmental cause exerts its
effect. While time trend analysis is directly related to the temporality aspect of
a putative association, it is also indicative of potential long-term consequences.
This is particularly relevant in the context of the application of the PP to EDCs
as the observation of intergenerational effects in laboratory animals (Newbold
et al., 1998, 2000) raises concerns of ‘‘threats of irreversible damage.’’
Association. The second criterion integrates the Bradford Hill tenets of
association between environment and disease, aside from temporality mentioned above. It is related to evidence derived from cohort, cross-sectional, or
case-control studies of not only environmental but also accidental, occupational, nutritional, or pharmacological exposure. Results from toxicological
studies were also considered. Such studies often include some measure of the
334
MARTIN ET AL.
strength of the association under investigation and its statistical significance.
This was cited when it described the association between environmental EDCs
and human health. Occasionally, studies are designed to investigate the
existence of a biological gradient or dose-response. Plausibility is related to
the state of mechanistic knowledge underlying a putative association, while
coherence can be related to what is known of the etiology of the disease and in
this case known risks factors related to endocrine parameters. Experimental
evidence can be derived both from toxicological studies and natural epidemiological experiments following occupational or accidental exposure. In the
context of ED, analogy with the effects of pharmaceutical doses of hormones or
dietary sources of phytoestrogens is of particular interest. Consistency of
observations between studies, with chemicals exhibiting similar properties or
between species was also considered under this criterion. As this study included
multiple disease end points, specificity was not expected to be always pertinent,
although if any evidence of specific environment-disease association was
found, it would also be taken into account.
employed to signify the absence of evidence of any association; /þ,
inconclusive evidence of an association with environmental EDCs; þ, sparse
evidence of an association based on animal data or mechanistic studies;
and þþ, some epidemiological evidence of an association with at least
some environmental EDCs. For the consequence criterion, ? was used
when consequences of health end point were not clearly defined; (?),
when there were suggestions of yet unrealized but potentially serious or
irreversible consequences; /þ, to indicate the end point had been proposed as
a biomarker; þ, a significant impact on personal health; þþ, a significant
impact on both personal and public health; and þþþ, a severe impact on
both personal and public health.
Ratings were attributed according to internal consistency between the end
points under consideration and relative to the current overall level of proof
underlying the environmental ED hypothesis.
Consequence. The third criterion aims to give an appreciation of the
adverse consequences of a particular health end point from both an individual
perspective, such as the impact of disease on quality of life or life expectancy,
and a public health perspective or the burden of disease or treatment on society
as a whole. The distinction between personal health and public health is
important. A small increase in the relative risk of developing a relatively benign
condition may be of little relevance to personal health. However, it may still be
of importance from a public health perspective if the disease is common as
small effects in a large population can have large attributable risks (Steenland
and Savitz, 1997; Coggon et al., 2003). Prevalence of the disease end point is
therefore particularly relevant, as well as any health risks associated with the
health concern under consideration. In this context, biomarkers or other
potential indications of yet undefined but potentially serious consequences also
needed to be taken into consideration.
The European Environment Agency (EEA) ‘‘Late Lessons’’ report (2001)
provides a working definition of the PP:
Ratings
The updated evidence summarized under each criterion was rated by
assigning to it a qualitative relative measure integrating the existence of
evidence and magnitude and direction of any effects as þ, þþ, or þþþ. No
negative ratings were defined reflecting the fact that negative results cannot
constitute proof of the validity of the null hypothesis or the lack of an effect.
Rather evidence of absence of an effect, or inconclusive evidence, was rated
as /þ, while ? was used to denote the absence of evidence. For the incidence
trends criterion, ? was used to denote the absence of evidence of any trend; /þ,
inconclusive evidence; þ, sparse evidence of an increasing trend or environmental influences; and þþ, some evidence of increasing trend of incidence
and environmental influences. Similarly, for the association criterion, ? was
Evidence Assessment
The Precautionary Principle provides justification for public policy actions
in situations of scientific complexity, uncertainty and ignorance, where there
may be a need to act in order to avoid, or reduce, potentially serious or
irreversible threats to health or the environment, using an appropriate level
of scientific evidence, and taking into account the likely pros and cons of
action and inaction. (EEA, 2001)
It specifies complexity, uncertainty, and ignorance as contexts where the PP
may be applicable and makes explicit mention that PAs need to be justified by
a sufficiency of scientific evidence. The report also offers a clarification of the
terms risk, uncertainty, and ignorance and corresponding states of knowledge
with some examples of proportionate actions. The evidence gathered for the
present study was assessed according to these working definitions reproduced
in Table 1, where ‘‘I’’ stands for ignorance; U, uncertainty; and R, risk. Rating
evidence between I and U depended therefore on the level of evidence or
knowledge concerning the impact or consequence of a specific disease end
point. The R rating applies to well-characterized and quantifiable risks and was
therefore not expected to apply to environmental exposure to EDCs.
RESULTS
The end point–specific assessment of scientific evidence and
ratings attributed under each of the three criteria are given in
TABLE 1
Clarification of Terms Used for the Assessment of Scientific Evidence (reproduced from EEA, 2001)
Situation
Risk
Uncertainty
Ignorance
State and dates of knowledge
‘‘Known impacts’’; ‘‘known probabilities,’’ for example,
asbestos causing respiratory disease and lung and
mesothelioma cancer; 1965–present
‘‘Known’’ impacts; ‘‘unknown’’ probabilities, for example,
antibiotics in animal feed and associated human resistance
to those antibiotics; 1969–present
‘‘Unknown’’ impacts and therefore ‘‘unknown’’ probabilities,
for example, the ‘‘surprises’’ of CFCs and ozone layer
damage prior to 1974; asbestos mesothelioma cancer
prior to 1959
Note. Chlorofluorocarbons, CFCs.
Examples of action
Prevention: action taken to reduce known hazards, for example, eliminate
exposure to asbestos dust
Precautionary prevention: action taken to reduce potential risks, for example,
reduce/eliminate human exposure to antibiotics in animal feed
Precaution: action taken to anticipate, identify, and reduce the impact of
‘‘surprises,’’ for example, use of properties of chemicals such as persistence
or bioaccumulation as ‘‘predictors’’ of potential harm; use of the broadest
possible sources of information, including long-term monitoring; promotion
of robust, diverse, and adaptable technologies and social arrangements to meet
needs, with fewer technological ‘‘monopolies’’ such as asbestos and CFCs
335
HUMAN HEALTH AND ENDOCRINE DISRUPTION
Table 2. The scientific evidence on which ratings and evidence
assessment were based is summarized for individual end points
in the following section. Unless otherwise referenced, information was extracted from chapter 5 of the IPCS ‘‘Global
Assessment of the State-of-Science on Endocrine Disruptors’’
report (2002). It is important to note that the ratings attributed
are given here to illustrate the use of the framework. These only
reflect the consensus achieved between the authors and are
open to discussion. Such qualitative ratings are subject to value
judgments, and indeed, the use of a multicriteria framework in
a participative context can help make those explicit.
Reproductive Health
Sperm Quality
Incidence trends. Due to seasonal, geographical, genetic
or age-related, or abstinence-related variations, it remains
unclear whether there is a trend for a worldwide decline in
sperm quality. Potential recruitment bias has been a recurrent
issue for this type of study (Slama et al., 2004). There is,
however, some evidence of a time-related decrease within
some regions.
Association. Epidemiological evidence remains the subject
of much controversy, due in part to conflicting results following
in utero exposure to diethylstilbestrol (DES). More recently,
a dose-response relationship between monobutylphthalate in
urine and sperm motility was found in males of subfertile
couples (Duty et al., 2003), while a cross-sectional study found
a statistically significant trend between 2,2#,4,4#,5,5#-hexachlorobiphenyl (CB-153), as a proxy of the total polychlorinated biphenyls (PCBs) burden, in serum and sperm DNA
damage for European males (adjusted odds ratio ¼ 1.51) but
not for Inuits, indicating that differences in genetic background
and/or lifestyle habits may play a role. No such association was
found with dichlorodiphenyldichloroethylene (DDE) serum
levels (Spanò et al., 2005). Current research points to an
association between high body mass index with reduced sperm
quality and quantity (Jensen et al., 2004; Kort et al., 2006).
Nonetheless, similar effects on sperm quality have been
observed in rodents following chronic exposure to low
concentrations of octylphenol in drinking water and xenoestrogen-contaminated fish (Aravindakshan et al., 2004; Blake
et al., 2004).
Consequence. The potential fertility in the male is reduced
only when sperm concentration is very low. However, a study
TABLE 2
Assessment of the Evidence of Human Health Effects of EDCs under the Multiple Criteria Framework
Health concern or end point
Reproduction
Sperm quality
Fecundity and fertility
Spontaneous abortion
Sex ratio
Male reproductive tract abnormalities
Prostatitis
Endometriosis
Precocious puberty
PCOS
Shortened lactation
Neurobehavior
Developmental neurobehavior
Adult nervous system
Gender dependent behavior
Immune system
Impaired immune function
Cancer
Breast
Endometrial
Testicular
Prostate
Thyroid
Incidence trends
Association
Consequence
Evidence assessment
þ
/þ
?
/þ
/þ
þþ
þ
þ
þ
þ
þ (?)
þþ
þ
/þ (?)
þ
U (I)
U
U
I
U
?
þþ
þ
?
?
þ
þþ
þ
þ
þþ
þþ
þþ
/þ (?)
þþ
þþ
U
U
I
U
U
/þ
?
?
þþ
þþ
þ
?
?
?
I
I
I
þ
þ
þþ
U
þþ
/þ
þþ
þþ
þ
/þ
/þ
þ
þþ
/þ
þþþ
þþ
þþ
þþþ
þ
U
U
U
U
U
Note. ?, absence of evidence of any incidence trend or association or consequences of health end point not clearly defined; /þ, inconclusive evidence of trend
or association, or end point proposed as a biomarker of population health; þ, sparse evidence of an increasing trend or environmental influences, of an association
based on animal data or mechanistic studies, or significant impact on personal health; þþ, some evidence of increasing trend of incidence and environmental
influences, epidemiological evidence of an association with at least some environmental EDCs, or significant impact on both personal and public health; þþþ,
a severe impact on both personal and public health; I, ignorance; and U, uncertainty.
336
MARTIN ET AL.
of Danish military conscripts reported a median sperm count
close to the value below which increased time to pregnancy
may occur (Toft et al., 2004). The presence of chromosome
abnormalities in sperm increases the risk of recurrent abortion
(Egozcue et al., 2003).
Evidence assessment. The large research effort on this
topic has yielded at least some epidemiological evidence that
specific EDCs could have an effect on sperm parameters, even
if an ED mechanism has not been demonstrated. If potential
impacts on male fertility give reason for precaution, it is
probably the unknown and potentially irreversible effects that
DNA damage could have on the gene pool that have warranted
most of the concern relating to this end point.
Fertility and Fecundity
Incidence trends. There is no clear picture with respect to
time trends in couple fertility. In the United Kingdom, an
increase rather than decrease has been seen. In Sweden, the
decrease in subfertile women has been linked to a lower
incidence of sexually transmissible disease. Geographical
differences in time to pregnancy remain unexplained and do
not appear to be linked to difference in semen quality. Changes
in the availability and use of effective contraception and
induced abortion bias the direct study of time trends such that
it is probably impossible to identify biological changes in
fertility over recent decades (Sallmén et al., 2005).
Association. There are some indications that exposure to
pesticides may contribute to female infertility in some
occupationally exposed groups. However it was not accompanied by any effect on sex hormones (Hanke and Jurewicz,
2004). Data on high parental consumption of contaminated
fish and time to pregnancy remain inconclusive, however
effects may be masked by a protective effect from some of
the constituents of fatty fish (Axmon et al., 2001, 2002;
McGuinness et al., 2001).
Consequence. In general, 1 in 10 couples experiences delay
or inability to conceive. It is a source of social and psychological suffering and can lead to stigmatization, while assisted
reproduction procedures are generally costly (WHO, 2001).
Evidence assessment. If the consequences of subfertility
are serious from both public and personal health perspectives,
the sparse epidemiological evidence remains inconclusive and
an endocrine mechanism has not been demonstrated.
Spontaneous Abortion
Incidence trends. Incidence is estimated to be about 50%
of all pregnancies, based on the assumption that many
pregnancies abort without clinical recognition, which in turn
makes this outcome particularly difficult to monitor. A slight
increase in incidence between 1990 and 1995 was clinically
recognizable in Italy, while known risk factors for early
pregnancy loss include maternal age and gravidity (Rozzi
et al., 2000).
Association. There is some evidence of an effect following
parental occupational exposure to pesticides, but an endocrine
mechanism has not been established. Compounds acting on
progesterone function are of most interest. High dose of
estrogens can also prevent implantation, and one of the health
consequences identified as a result of in utero DES exposure
has been poor pregnancy outcome in ‘‘DES daughters,’’ but
dose considerations cast doubt on the potential involvement of
environmental estrogens. Animal data from studies with
hexachlorobenzene and ketoconazole provide evidence of
a potential ED mechanism.
Consequence. One abortion is associated with an increased
risk of subsequent (recurrent) abortions (Cramer and Wise,
2000).
Evidence assessment. While an association may not have
been demonstrated, it is plausible, coherent, and supported by
some experimental evidence and analogy with pharmaceutical
hormones. Lack of clinical recognition may, however, hamper
any estimation of the real scale of any potential impact of EDC
exposure.
Sex Ratio
Incidence trends. There is an apparent decline in some
regions and an increase in others, assuming the sex ratio
normally remains stable over time. Racial differences were
found within countries, as well as variations between regions
(Canada and United States) and between metropolitan and
nonmetropolitan areas (Italy). Exogenous stress has been found
to lower the sex ratio (Catalano, 2003; Catalano et al., 2006). It
has also been suggested that changes in sex ratio antedated any
exposures to environmental chemicals.
Association. There is evidence of external influence including medical factors consistent with an influence of the
hormonal status of the parents at time of conception. Evidence
from occupational or accidental exposure to organochlorines is
consistent with an effect, but an ED mechanism has not
demonstrated. More recently, it has been argued that the
direction of an effect of EDCs on sex ratio may be different
depending on whether the mother or father had been exposed
(James, 2006).
Consequence. The sex ratio has been proposed as a potential sentinel marker for population health, although a big
change in ratio would in itself be expected to have an impact on
demographics.
Evidence assessment. Available experimental evidence
appears to be coherent with current medical knowledge.
However, consequences are poorly defined, from sentinel
marker to the social and demographic impact of a significantly
skewed ratio.
HUMAN HEALTH AND ENDOCRINE DISRUPTION
Male Reproductive Tract Abnormalities
Incidence trends. Some report temporal increases; however, statistics for birth prevalence vary widely, probably due to
differing diagnostic and reporting criteria and to ethnic/genetic
differences. A rising trend may simply reflect more frequent or
earlier diagnosis of minor forms over time or an increasing
tendency to report them to birth defect registries. The higher
prevalence of congenital cryptorchidism in Denmark compared
with Finland suggests genetic factors are at play, although
environmental or lifestyle factors may play a role in milder
forms of testis maldescent (Boisen et al., 2004). In England and
Wales, the incidence of hypospadias appears to be decreasing
after a steady increase between 1965 and 1983, and rates of
surgical treatment for cryptorchidism remained unchanged
between 1992 and 1998 (Safe, 2005). There are indications
of a tendency toward a decline in this pathology after 1985.
Association. Considerable evidence from laboratory animals of the adverse effects on the male reproductive tract of
estrogenic and antiandrogenic chemicals during critical periods
of development exists. Some epidemiological studies point to
a slightly increased risk of cryptorchidism and hypospadias
associated with exposure to pharmacological doses of estrogens in utero, including DES. Wildlife exposed experimentally
to EDCs is susceptible to some of the same developmental
abnormalities (Edwards et al., 2006). Whether environmental
concentrations of EDCs could have a similar effect remains
controversial, and more research is needed to elucidate the
combination effects of complex chemical mixtures (Fisher,
2004).
Consequence. Cryptorchidism is by far the commonest
congenital abnormality (2–4% incidence) and hypospadias the
second commonest (0.3–0.7% at birth) (Sharpe and Skakkebaek, 2003). There is a need for corrective surgery in severe but
not all cases. Cryptorchidism is also a known risk factor for
testicular cancer.
Evidence assessment. Evidence is consistent, plausible,
and coherent, although the analogy with pharmaceutical
estrogens raises doubts over the strength of a potential
association. If prevalence is relatively low compared to that
of other end points considered, male reproductive abnormalities are nonetheless common.
Prostatitis
Incidence trends. Prostatitis is relatively poorly defined
and there is little epidemiological data available to date. There
is no diagnostic test, and classification has recently been
reviewed (Batstone et al., 2003; Schaeffer, 2003; Schaeffer
et al., 2002). Chronic pelvic pain syndrome is of unknown
etiology and accounts for more than 90% of prostatitis
diagnoses (Habermacher et al., 2006).
337
Association. The rat prostate can be imprinted by endogenous and exogenous estrogens during the neonatal period and
is associated with an increased incidence of hyperplasia,
dysplasia, and adenocarcinoma with aging. Estrogen or prolactin treatments are known to induce inflammation in the rat.
Additionally, exposure to certain pesticides and other environmental compounds during the perinatal, lactational, or early
pubertal periods can lead to lateral prostate inflammation or
prostatitis in the adult male rat. In humans, research has pointed
to an autoimmune mechanism (Batstone et al., 2003).
Consequence. Prostatitis-like symptoms were estimated to
affect nearly 10% of men aged 20–74 years in a Canadian study
(Nickel et al., 2001). Nonbacterial prostatitis is fairly common
and can affect fertility and may also be a risk factor for prostate
cancer (Palapattu et al., 2004).
Evidence assessment. Evidence is derived primarily from
animal studies and is coherent with an association. Although
only recently defined, consequences of this common condition
are thought to be serious from both personal and public health
perspectives.
Endometriosis
Incidence trends. Although determining the incidence of
endometriosis is complicated by issues surrounding diagnosis,
it has been suggested that incidence may have increased over
the past 50 years. There also appears to be a decrease in the age
of onset, when this condition has been traditionally associated
with delayed childbearing (Birnbaum and Cummings, 2002).
Association. The administration of exogenous estrogens
has been shown to aggravate the disease in a dose-dependent
fashion, and it is hypothesized that the imbalance between
progesterone and estrogen may be involved in the pathogenesis
and pathophysiology of the disease. In humans, statistically
significant associations have been reported with blood levels of
PCBs and toxic equivalency quotient for dioxins, furans, and
dioxin-like PCBs in serum (Heilier et al., 2005; Porpora et al.,
2006). Such an association with serum levels of organochlorine
compounds was not observed in infertile Japanese women
(Tsukino et al., 2005), while long-term treatment with high
doses of phytoestrogen supplements was associated with
endometrial hyperplasia (Unfer et al., 2004).
Animal data seem to support the potential contribution of
tetrachlorodibenzo-p-dioxin (TCDD), a potent aryl hydrocarbon receptor (AhR) agonist, in the pathobiology of endometriosis, and an immune mechanism has been proposed.
Consequence. A common gynecologic disorder as well as a
major cause of infertility, endometriosis has also recently been
associated with an increased cancer risk (Brinton et al., 2005).
Evidence assessment. There is a clear biological gradient
between severity of the illness and exogenous hormones, and
evidence is coherent with ED. While an estrogenic etiology is
338
MARTIN ET AL.
plausible, the strength of the association with environmental
EDCs is weak. Although estimates for the prevalence of
endometriosis vary, it is considered both common and the
cause of significant morbidity, while rising incidence trends
warrant concern.
Precocious Puberty
Incidence trends. Between the mid-19th and the mid-20th
century, the average age at menarche (first menstruation)
decreased steadily from 17 to under 14 years old in United
States and in some countries in Western Europe. This has been
attributed to the evolution of living standards and supports the
role of nutrition, health status, or socioeconomic status, and
there are indications that this has now stabilized. There are,
however, concerns over a decrease in age at thelarche (breast
development). More recently published studies point to an
increase in the proportion of female patients with idiopathic
rather than organic forms of precocious puberty (Parent et al.,
2003), while the geographical distribution of incidence of
premature breast development and gynecomastia (excessive
breast development in males) seem to be consistent with
environmental influences (Massart et al., 2005b).
Association. Several reports have indicated an association
between premature breast development and exposure to DES,
ethinyl estradiol, and mestranol.
Consequence. Symptoms of precocious puberty are being
proposed as possible sentinel markers of environmental
exposure. The overall incidence of sexual precocity is estimated to be between 1 in 5000 and 1 in 10,000, with females
being 10 times more likely to be affected than males (Partsch
and Sippell, 2001). If untreated, the prematurity of the pubertal
growth spurt can result in reduced adult height. The psychological consequences of asynchrony between biological development and social expectations will depend on the
sociocultural context (Waylen and Wolke, 2004).
Evidence assessment. Analogy with pharmaceutical estrogens is coherent with an estrogenic influence; however, an
association with EDCs and the wider significance of precocious
puberty still appear unclear.
Polycystic Ovaries Syndrome
Incidence trends. Polycystic ovaries syndrome (PCOS)
has only recently been defined. Signs vary and include
chronic anovulation, hyperandrogenism, insulin resistance,
and obesity.
Association. Genetic studies suggest a strong familial
component. Currently, PCOS is considered to result from
the interaction of genetic traits with environmental factors
(Diamanti-Kandarakis and Piperi, 2005). In animals, there is
evidence that testosterone exposure affects the female developing brain, including induction of acyclicity and anovulation.
Polycystic ovaries can also be induced in a variety of animals
through estrogen administration. However, there is no generally
acceptable animal model of the pathophysiology of PCOS.
Consequence. PCOS is probably the most common endocrine disorder in women of childbearing age with a prevalence
of up to 10%, and according to an ultrasound study, polycystic
ovaries may be a feature in 22% of young women (Hart et al.,
2004). PCOS can be a cause of hirsutism, diabetes, and
infertility (Azziz et al., 2005) and has also been associated
with endometrial cancer in premenopausal women (Pillay
et al., 2006).
Evidence assessment. Animal data following hormone
treatment are coherent with hormonal influences. While
evidence of an association may be scarce, the disorder is both
common and serious.
Shortened Lactation
Incidence trends. Studying the incidence of poor lactation
performance is likely to be complicated by many socioeconomic factors. Influences on lactation initiation and duration
that have received some attention include the use of pacifiers,
maternal obesity, maternal and fetal stresses, as well as
environmental chemicals (Dewey, 2001; Lovelady, 2005; Ullah
and Griffiths, 2003).
Association. Pharmacological doses of combined estrogen:progestagen oral contraceptives (OCs) are known to affect
lactation. In areas of Northern Mexico where dichlorodiphenyltrichloroethane (DDT) has been heavily used, fat levels of
DDE in human breast milk were found to be negatively
correlated with duration of lactation. More recently, it was
found to be also associated with lactation initiation; the
incidence ratio for breast-feeding initiation in nonsmoking
mothers with maternal DDE concentrations over 5 lg/L was
under 0.5, while no such associations were found for PCBs in
breast milk (Karmaus et al., 2005). In rodents, lactation can be
impaired by exposure to the pesticide atrazine through inhibition of prolactin, while the delayed mammary differentiation following TCDD treatment during pregnancy is thought
more likely to be mediated via AhR signaling than the result of
impaired hormone production (Vorderstrasse et al., 2004).
Consequence. The potential risks posed by the chemicals
found in breast milk need to be balanced against the proven
benefits of breast-feeding for both mother and infants (Massart
et al., 2005a), while bottle feeding has also been associated
with the risk of contamination by plasticizers (Onn Wong et al.,
2005; Ozaki et al., 2002).
Evidence assessment. An association between lactation
and EDCs is biologically plausible and coherent, while
evidence points to a modest association with some EDCs. An
impact could be defined as the reduction in the benefits that
would have been gained had breast-feeding been initiated/
339
HUMAN HEALTH AND ENDOCRINE DISRUPTION
continued and risks related to the chemical legacy of mothers’
exposure to DDT estimated.
Neurobehavior
Developmental Neurobehavior
Incidence trends. Diagnosis of conditions such as autism,
attention deficit/hyperactivity disorder, and learning difficulties
are on the rise, but whether this is due to a true increase in
incidence or increased public awareness and changes in
diagnostic criteria remains unclear (Szpir, 2006).
Association. The brain is now considered the target organ
most sensitive to damage. A lowering of the IQ of children
exposed neonatally to PCBs was reported after two mass
poisoning events in Japan in 1968 (Yusho) and in Taiwan in
1979 (Yu Cheng). Environmental PCBs have been reported to
have a negative impact and linked to delays in postnatal
psychomotor, neurological, or cognitive development in most
but not all studies (Gray et al., 2005; Schantz et al., 2003). The
degree of persistence of such effects remains controversial,
and the mechanistic basis of the effects is unclear, although
hypothyroid action has been proposed as a potential mechanism by which neurobehavioral dysfunction is mediated
(Colborn, 2004). Neurobehavioral effect assessment is still
in its infancy and the exposure assessment of potential neurotoxicants impeded by confounding (Bellinger, 2004; Dietrich
et al., 2005). In animals, recent evidence indicates that
environmental concentrations of estrogenic chemicals could
affect brain development and effects could be mediated via the
estrogen receptor ER-b expression and also suggests a possible
link between endocrine disrupters and hyperactivity (Laviola
et al., 2005; Masuo et al., 2004; Salama et al., 2003).
Consequence. In addition to the conditions already mentioned, it has been suggested that some neurodevelopmental
disorders may be associated with mental illness, suicide,
substance abuse, and a higher propensity to commit crimes in
later life (Szpir, 2006).
Evidence assessment. There is consistent, coherent, and
plausible evidence of a potential effect of EDCs on the
developing brain. Effects are potentially serious, but specific
end points remain to be clearly defined.
lifetime but not when exposed briefly in adulthood (Patisaul,
2005). Bisphenol A has been found to disrupt sexual behavior
even in adulthood (Funabashi et al., 2003). While genistein
induced dopamine release by an estrogenic mechanism, such
effects were not found following exposure to nonylphenol
(Ferguson et al., 2002; Funabashi et al., 2003).
Consequence. Endocrine disruptors may alter hormonal
actions in neuroendocrine processes, thereby rendering the
nervous system more susceptible to harmful events.
Evidence assessment. There is consistent, coherent, and
plausible evidence of a potential influence of environmental
EDCs, but the specific impacts of a potential effect need to be
clarified.
Gender-Dependent Behavior
Incidence trends. No relevant data were found.
Association. This generally relates to responses that are
differentially expressed in both sexes but not directly related to
sexual functions and reproduction. Female progeny of DEStreated mothers appear to be masculinized or defeminized in
terms of sexual orientation and language laterality, while male
progeny display less evidence of behavioral change. In animals,
there is mounting evidence of behavioral disruption following
interaction of xenobiotics with gonadal steroids, but studies
have included a wide spectrum of behavioral end points.
Consequence. No relevant data were found.
Evidence assessment. Analogy with DES daughters and
experimental data with rodents provide evidence that EDCs
could influence behavior, but the specific adverse impacts of
a potential effect are not yet clearly defined.
Immune System
Increase in disease incidence is the ultimate outcome of
effects on the immune system, and the current methodology did
not yield any literature on the integrated incidence trend of the
multitude of individual health end points that this would
encompass. The methodology was refined by including key
words relating to more specific health conditions mentioned in
the IPCS report.
Impaired Immune Function
Adult Nervous System
Incidence trends. No relevant data were found.
Association. A few studies suggest that occupational or high
environmental exposure to polyhalogenated aromatic hydrocarbons could affect the thyroid and influence adult neurobehavior. Estrogens are known to influence dopamine release.
Phytoestrogens affect stress response and social behavior as well
as sexual behavior, when animals are exposed over their entire
Incidence trends. An impaired immune function can result
in immunosuppression and allergic and autoimmune diseases.
Allergies are the most common chronic childhood diseases in
Western Europe and the United States. Asthma in particular has
a prevalence varying between 10% and 20% between regions
(Kabesch, 2006), and this has steadily increased over the last
few decades, although a decrease has been observed in some
regions over the last 10 years. Established risk factors cannot
account for the increase in global prevalence that is generally
340
MARTIN ET AL.
linked to environmental or lifestyle factors (Pearce and Douwes,
2006). Interactions between genes and the environment are the
current focus of research into etiological factors (Kabesch,
2006; Marks, 2006). There is, however, some evidence of
a decline in hospitalization and General Practioner consultation
rates for asthma in Australia (Poulos et al., 2005), and the trend
appears to have stabilized in Italy (Ronchetti et al., 2004).
Association. Increases in disease in humans, the ultimate
outcome of immune dysfunction, are detectable end points, but
causality is difficult to attribute. To date, the majority of
immunotoxic reactions to chemicals where the mechanism is
known do not involve endocrine effects. However, several
autoimmune diseases are far more common in women than in
men, and there is also a strong association between endometriosis and female-predominant autoimmune disorders (Seery,
2006). Reviews of the immunological effects of DES exposure
concluded that, in general, humans exposed prenatally to DES
do not exhibit severe defects in basic immune function, but their
propensity to develop autoimmune disease and other diseases
associated with defects in immune regulation appears to be
increased. Immunotoxicity of exogenous estrogens was mainly
manifested as changes in the thymus and thymus-dependent
immunity and occurred at pharmacological dose levels, raising
the question whether immunotoxic effects can be expected at all
with low exposures to weak estrogens, in contrast to TCDD,
which is a potent immunotoxic compound at low exposure level.
Absence of immune effects in rodents following perinatal or
juvenile exposure to methoxychlor supports this. Substantial
amount of animal data demonstrate numerous immune alterations following in utero exposure to DES.
Consequence. Impaired immune function can result in both
altered resistance to infection and/or tumors as well as the
exarcerbation of allergic and autoimmune disease.
Evidence assessment. While analogy with DES prenatal
exposure and other hormonal risk factors are coherent with
a potential association with autoimmune disorders, dose and
potency considerations as well as experimental data from
animal studies do not support this. Autoimmune disorders have
a serious impact on the life of sufferers and are also significant
from a public health perspective.
Cancer
Breast
Incidence trends. Breast cancer incidence has increased
steadily from the 1940‘s through the 1990’s in many industrialized countries. While increased screening can at least
partly explain the increase in the 1980’s, it does not explain the
increase in risk for younger women and older women in
developing countries (Brody and Rudel, 2003). Incidence is
rising most rapidly in previously low-risk populations. Great
variations in relative frequencies are observed between countries, and following migration, breast cancer rates become
similar to those of the host country within two to three
generations, strongly suggesting the influence of environmental
or lifestyle factors. It has been suggested that altered reproductive patterns have increased estrogen surges in modern
women resulting in increased cancer risk. Additionally,
pharmacological hormone replacement therapy (HRT) and
OCs contraceptives are now commonly used. However, known
risk factors predict only poorly breast cancer risk, and other
etiological links need to be identified including exposure to
endocrine disrupters (Sasco, 2003).
Association. Human studies support the etiological role for
estrogens in breast cancer related to increased lifetime
exposure to endogenous estrogens. There is a small increase
in the relative risk for breast cancer in current and recent HRT
and/or OC users, not seen 10 years after cessation. Studies
examining the potential association between exposure to
specific environmental contaminants such as organochlorines
and/or PCBs remain inconclusive. A twofold increase in risk,
which displayed a dose-response relationship, was found for
dieldrin, while a more recent study found no increase in breast
cancer incidence in areas of recent, high agricultural pesticide
use (Reynolds et al., 2004). The perinatal period and the period
between age at menarche and age at first full-term pregnancy
may be particularly important for breast tumor development
and latency. Preliminary results on the risk of breast cancer in
daughters of DES-exposed women found a twofold increase in
relative risk in women over 40 arising exclusively from the ERpositive cases subset (Palmer et al., 2002). While this is
consistent with a carcinogenic role for estrogens, a role for
environmental estrogens should be considered within the
context of their weaker potencies and lower exposure levels.
Similarly, in rodents, experimental studies have clearly shown
that hormone treatment can alter tumor sensitivity and that the
observed effects are dose dependent. With the exception of
DDT, organochlorines and other suspected EDCs have not been
associated with mammary gland cancer in animals. However,
the long latency and variable incidence of mammary tumors in
rodent strains limit the utility of studying such tumors as
a model of human disease.
Consequence. As one of the most common cancers in
humans, affecting almost exclusively women (male breast
cancer accounts for approximately 1% of breast cancer cases),
it is now being considered an epidemic (Sasco, 2003) and is
responsible for over 500,000 deaths a year worldwide (WHO,
2006).
Evidence assessment. An environmental etiology is supported by temporality, while a role for EDCs is coherent with
endogenous estrogen exposure and the analogy with pharmaceutical uses of synthetic estrogenic compounds. However,
experimental data with rodents and epidemiological studies
HUMAN HEALTH AND ENDOCRINE DISRUPTION
have consistently failed to show an association between breast
cancer and known EDCs. It has been argued that life stage at
time of exposure may be of importance in the etiology of the
disease, and this is supported by emerging data following
prenatal DES exposure.
Endometrium
Incidence trends. There is no evidence of an increasing
temporal trend. Incidence is highest in the United States and
Canada and lowest in Asia and Africa. An intermediate
incidence is found for Asian women who have migrated to
the United States, and this again suggests that lifestyle factors
play an etiological role; however, to date, diet has received
most of the scientific attention (Tao et al., 2005).
341
mycotoxin ochratoxin A, also associated with impaired
spermatogenesis, could better explain geographical differences
has also been put forward (Schwartz, 2002). Interaction
between genetic and environmental and lifestyle factors has
been put forward as a potential explanation for the high
variability in incidence, while interpreted by others as consistent with an epidemic in different phases in different countries
(Bray et al., 2006; Giwercman et al., 2006). Some evidence
suggests that the incidence of cryptorchidism and hypospadias
may show similar geographic differences to the incidence of
testicular cancer and, therefore, possibly share the same
etiology. These conditions together with reduced sperm quality
could possibly be components of a testicular dysgenesis
syndrome.
Association. Unopposed estrogens are known to be the
major risk factor for endometrial cancer. However, no association was found between PCBs, organochlorine, and endometrial cancer. In animals, neonatal treatment of mice with DES
has been shown to cause uterine carcinoma within 18 months;
however, there is a paucity of data on whether long-term
exposure at low doses can result in neoplasms. As with other
cancers, the timing of exposure is critical to the potential
development of uterine cancer. Developmental exposure of
mice to DES results in uterine neoplasms, whereas treatment of
adult mice with comparable levels of DES does not induce such
malignancies.
Association. Most of the tumors occurring in young men
are seminomas of germ cell origin and cryptorchidism is
a known risk factor, both suggesting a prenatal etiology.
A metaanalysis of available studies of DES sons has shown
a statistically significant twofold increase in testicular cancer.
The familial component of the risk of testicular cancer may be
the product of shared childhood environment and heritable
causes (Hemminki and Chen, 2006). The type of testicular
cancer commonly found in humans, seminomas, is preceded by
atypical intratubular germ cells termed carcinoma in situ (CIS)
and is extremely rare in laboratory animals. CIS lesions have
been induced in rabbits exposed in utero and/or infancy to
octylphenol, DDT/DDE, or zeranol.
Consequence. Endometrial cancer is the most common
gynecologic cancer. In the United Kingdom, nearly 6000 cases
of uterine cancer are diagnosed and more than 1500 deaths due
to uterine cancer are recorded each year.
Consequence. Testicular cancer is the most common
malignancy among young men, 25–34 years old. Overall,
prognosis is good, with high cure rates being achieved even
with metastatic disease (Kopp et al., 2006).
Evidence assessment. A possible role for EDCs is coherent
with what is known of the etiology of the disease. While no
association has been shown with known EDCs, experimental
data from rodents exposed neonatally to DES suggest timing of
exposure is critical.
Evidence assessment. Temporality and existing experimental toxicological evidence is coherent with a prenatal
etiology of the disease and the analogy with in utero DES
exposure. The strength of the latter association with DES does,
however, raise issues concerning potency and dose.
Testicular Cancer
Incidence trends. Incidence has increased across all continents; however, the strongest and most consistent rise was
seen in Europe, even if profound geographical variations across
countries were found (Purdue et al., 2005; Richiardi et al.,
2004). In countries with a long history of cancer registration,
the start of the increase can be traced back to around 1920.
There are marked differences in incidence levels of testicular
cancer between countries and between races. Generational
influences appear largely responsible for the increasing trends
in Europe, and a reduced incidence for the cohort born during
the Second World War was observed in a number of countries.
This is consistent with the potential etiological role of
environmental exposures of the developing embryo (Bray
et al., 2006). However, the hypothesis that exposure to the
Prostate
Incidence trends. An abrupt increase in age-adjusted incidence rates since the mid-1980’s can be largely attributed to
improved screening and diagnostic tests. However, a long-term
rise in both incidence and mortality remains unexplained.
There are racial and geographical differences in susceptibility.
Again, incidence risk of migrant population from low- to highrisk countries or vice versa tend to match those of the host
population (Grover and Martin, 2002), while some low-risk
regions are reporting a more rapid increase than high-risk
countries (Sim and Cheng, 2005), suggestive of an etiological
role for environmental and/or lifestyle factors.
Association. Prostate cancer is known to be both hormone
dependent and able to be modulated by hormone treatment.
342
MARTIN ET AL.
The epidemiological evidence has been mostly limited to
occupational exposure. No association was found between
PCB or DDE exposure and prostate cancer. No significant
increases in prostate cancer were reported following accidental
exposure to TCDD or occupational exposure in Italy, while
other studies on workers in Germany and the United States
demonstrated a small but significant increase in prostate cancer
mortality. There was a weak but statistically significant
association between acres sprayed with herbicides and prostate
cancer deaths. A limitation of studies investigating occupational exposure is the known effects of sun exposure as well as
a potential effect of physical activity (Aydin, 2005). More
recently, a positive association and dose-response trend was
found between two PCB groupings, moderately chlorinated
PCBs and PCBs with phenobarbital-like activities, and prostate
cancer risk, and cytochrome CYP450–mediated metabolism
was the mechanism proposed (Ritchie et al., 2005).
In contrast to its frequent occurrence in humans, prostate
cancer is rare in laboratory animals, but the frequency can
increase significantly with hormone treatment depending on
the strain of rat, dose, and duration of treatment. It is known,
however, that the development of the prostate gland or the
propensity of this organ to develop cancer can be affected by
perinatal/postnatal exposure to estrogens, phytoestrogens, and
possibly androgens and AhR agonists.
Consequence. Prostate cancer is the most commonly diagnosed cancer among men in developed countries and killed
over a quarter million men worldwide in 2002 (WHO, 2004).
Evidence assessment. Temporality suggests a potential
role for an environmental etiology, while ED is coherent with
current medical knowledge of the disease. There is little evidence of a positive association following occupational exposure; however, a biological gradient was observed with PCBs
possessing specific chemical characteristics. Although an ED
mechanism is biologically plausible, it has not been demonstrated. Experimental toxicological evidence with rodents
indicates that timing of exposure may be critical.
Thyroid
Incidence trends. Incidence rates are two to three times
higher in females than in males. Positive time trends have been
observed in several countries, particularly in younger men and
women (under 35 years old), women of childbearing age being
most at risk (Nagataki and Nyström, 2002; Remontet et al.,
2003; van der Horst et al., 2006). Differences in iodine intake
are one factor explaining geographic variations (Nagataki and
Nyström, 2002). The increase in incidence in Western European countries can be dated back to the mid-1970’s and
epidemiological evidence does not favor a link with the
Chernobyl incident, while the increase seems to have leveled
off in Sweden and Norway since the 1980’s. In France, the rise
in incidence was linked to a change in diagnostic practices for
the management of thyroid diseases (Leenhardt et al., 2004).
Association. The thyroid gland plays a key role in numerous endocrine, metabolic, and physiological functions, particularly growth and development. It is also involved in the
carcinogenic process and can affect tumor formation, growth,
and metastasis. Some environmental chemicals, such as PCBs
or polybrominated diphenyl ethers, have been shown to possess
antithyroidal activity; however, the only known human thyroid
carcinogens are x-rays and ionizing radiations. For papillary
thyroid cancer, the most prevalent histological type, prevalence
peaks for women within the reproductive period, whereas for
men, it increases steadily with age, suggesting a potential
influence of sex hormones on carcinogenesis. Studies have only
found weak associations with risk factors associated with
endogenous hormones (Sakoda and Horn-Ross, 2002).
Consequence. Thyroid cancer is an uncommon and largely
nonfatal tumor.
Evidence assessment. Evidence is coherent with an influence of endogenous hormones, although it does not support
the strength of such an association, if biologically plausible.
Incidence trends may be explained by changes in diagnosis,
and the consequences of the disease although serious are
relatively benign when compared with other tumor sites.
DISCUSSION
Much of the controversy surrounding the association between environmental EDCs and the various end points
considered is probably related to the lack of convincing
evidence related to the strength of such an association. While
for all the end points considered, concern arises from the
analogy either with hormonally active pharmaceuticals or risk
factors related to endogenous hormone levels, there is little
evidence that environmental levels of weakly active compounds elicit similar effects. Ecotoxicology, aquatic toxicology
in particular, and molecular biology are furthering the current
understanding of the effects of combined exposure. Such
research seems to support the pharmacological principles of
dose/concentration addition for substances acting via the same
mechanism and independent action for substances acting via
different mechanisms (Faust et al., 2003; Silva et al., 2002;
Zhu et al., 2006). Exposure assessment continues to limit the
interpretation of results from environmental epidemiology
studies regarding associations with common diseases with
multifactorial etiologies following low-level multiple exposure.
When statistically significant effects have been found, generally
in populations at risks following occupational or accidental
exposure to hormonally active compounds or even following the
prescription of pharmacological doses of hormones, estimates
tend to be relatively small, and, with the exception of DES, of an
order of magnitude comparable to the effects of risk factors
associated with endogenous hormone levels such as twinship.
They may still be relevant from a public health perspective,
HUMAN HEALTH AND ENDOCRINE DISRUPTION
however, when compared for example with the 30% reduction in
mortality associated with regular mammogram screening
(Brody and Rudel, 2003). PAs may therefore prove to be costeffective measures, especially when consequences are severe,
but should nevertheless be balanced by an evaluation of the
potential social, environmental, and economical costs of such
actions.
There is, however, a fairly consistent and recurrent finding
that timing of exposure may be the decisive factor of the nature
and temporal scale of effect. In animals, the period of development in utero is often observed to be the critical life stage
during which organisms are most sensitive to external influence
and during which exposure can lead to irreversible organizational effects and is proposed as one of the main limitations on
the power of epidemiological studies. This is consistent with
current scientific thinking in cancer research as the possibility
that many cancers may originate during the perinatal period is
receiving increasing attention (Anderson, 2004).
Dose considerations were rarely included in the reviewed
literature, and when present, results were most often characterized by a lack of any clear or consistent dose-response
relationship. Biological systems are characterized by complex
interactions and feedback loops that may not always be readily
described by linear cause-effect relationships. This is particularly relevant to a communication system such as the endocrine
system, and the array of end points considered suggests that
multiple stresses can result in a variety of disease states in
a continuum of severity within a population rather than
a dichotomous all or nothing phenomenon. One such example
is the ‘‘testicular dysgenesis syndrome’’ hypothesis stating that
low sperm count, abnormalities such as cryptorchidism and
hypospadias, and testicular cancer share the same etiology in
fetal life and are all disorders of one syndrome, in respective
order of accruing severity. Additionally, there are still difficulties in interpreting the data with regards to low dose-responses.
While some argue for the use of hormesis as the default
assumption for risk assessment (Calabrese and Baldwin, 2001),
several sources of uncertainty have been put forward to explain
the lack of reproducibility of the phenomenon: genetic factors,
some ED responses that have been found to be strain specific,
the intrinsic variability of some assay parameters such as organ
weights or intrauterine position, strain-specific ED responses,
and other environmental factors such as the role of the diet in
modifying outcomes (Ashby, 2001; Markaverich et al., 2002;
Witorsch, 2002). The fact that the nonhomogeneity of a population can give rise to a U-shaped dose-response curve is of
interest as it could be related to genetic polymorphisms and
variations in susceptibility (Chen, 2001).
The choice of the probabilities of making Type I and Type II
errors and the statistical correction applied in frequentist
hypothesis testing when considering multiple end points to
avoid inflating the Type I error rate have also been source of
contention. In some circumstances, such as the estrogenicity
in vitro bioassays applied to EDCs screening programs,
343
receiver operating characteristic analysis may be applied to
diagnostic methods to represent graphically the trade-offs
between false-positive and false-negative rates (Martin et al.,
2005). Because, in practice, end points are often correlated; the
main drawback of the Bonferroni correction, for example, is to
confer greater power to the detection of alternative hypotheses
where only one end point is statistically significant than when
several end points indicate a clinical effect that falls short of
statistical significance after correction (Pocock et al., 1987). In
a Bayesian setting, no such adjustment is necessary, but results
depend on the choice of an appropriate prior distribution and
the amount of belief assigned to it (van der Tweel, 2005). While
there is therefore a need to develop statistical tools better able
to address such complexity, the complexity of such methods
also increases the probability of results being misinterpreted by
the nonstatistician.
For both time trends and geographical studies, the analysis of
incidence trends may be biased by differences in the availability and evolution of diagnostic tests. Cryptorchidism, for
example, varies in severity, and the subtypes of undescended
testes positions included in registry data may vary between
geographical areas and over time, while age at diagnosis will
greatly influence rates as spontaneous descent occurs in many
cases in the first few months of life. Additionally, the lack of
baseline data against which to validate such secular changes is
another recurrent issue. This is already being addressed by
various national and international environmental monitoring
programs, such as the National Children’s Health Program,
a large multiyear prospective epidemiological study on early
life exposures to EDCs (Landrigan et al., 2003). Such programs
should also help define the disease end points most relevant to
developmental exposure. The wide geographical and racial
variations in the incidence of hormonally sensitive cancers and
other conditions such as endometriosis or asthma are suggestive of genetic susceptibility, an emerging issue for environmental epidemiologists. This concept is not entirely new and
constitutes the element of the ‘‘host’’ in the paradigm that
divides causes of disease into environment, host, and agent. Current technologies now permit the study of gene-environment
interactions, and the potential identification of susceptible
subgroups could lead to higher specificity and precision in
estimating the effects of exposure (Hertz-Picciotto, 1998).
Migrant studies of cancer incidence in particular are consistent with an influence from environmental or lifestyle
factors. However, the evolution of age at menarche highlights
the importance of nutritional, health, and socioeconomic factors. While exploratory studies can prove useful at generating
etiological hypotheses, the usefulness of ecologic analysis
is limited for testing such hypotheses. The central problem
results from making causal inferences about individuals on
the basis of group observations and is commonly referred to
as ‘‘the ecological fallacy’’ (Morgenstern, 1982).
The methodology employed retrieved data concerned with
incidence but did not yield many studies particularly
344
MARTIN ET AL.
addressing the social impact of a particular health end point,
and when necessary, additional information was sought from
the gray literature. It is unclear whether this stems from the
methodology itself or if this reflects a disciplinary bias and
illustrates the problem of compartmentalization of science or
lack of interdisciplinarity.
The organizational effects of hormones on the developing
brain are an emerging area of study, and neurobehavior is
a relatively new ED end point. Novel neuroendocrine pathways, such as the membrane-bound ER-X (Toran-Allerand
et al., 2002), are now being uncovered and raise both hopes of
finding potential targets for treatment and fears as potential
mechanisms of injury. Neurobehavior assessment is still in its
infancy and often measuring subtle end points, whose impacts
are still poorly understood. Exposure assessment is particularly
prone to confounding by socioeconomic factors arguably
leading to spurious associations, while adjustment thereof
may lead to erroneous estimates. Effects on adult behavior
would appear to be transitory, and the persistence of developmental effects is also controversial. However, even if the
consequences of potential effects are not well defined or even
likely to be routinely diagnosed or yet be recognized as health
concerns, they may still carry significant social costs. In such
a context of near ‘‘ignorance,’’ long-term environmental and
health-monitoring programs as well as funding research into
early warnings have been proposed as appropriate PAs by the
European Environment Agency. Actions of precautionary
prevention cautioned by a U rating should be justified by some
degree of estimation of risk, and such risks need to be balanced
by the benefits of specific chemical agents as well as the risks
associated with alternatives on a case-by-case basis.
It has been argued that a series of incidents such as ‘‘mad
cow disease’’ in Britain and HIV in blood transfusions in
France have undermined public confidence in governments’
use of science. The public’s willingness to accept risks are key
factors in determining a society’s chosen level of protection. The
increase in measles susceptibility following adverse publicity
concerning the measles, mumps and rubella vaccine illustrates
the importance of communication of risks as an integral part of
risk management (Friederichs et al., 2006). Public participation
has been proposed as a way to make environmental decisionmaking more transparent and increase trust in risk management
and in turn be associated with improved risk perception. It may,
however, be considered a cue that the risk is real, as was
suggested by a recent experiment on the risk perception of the
potential effects of electromagnetic fields (Wiedemann and
Schütz, 2005), although the study only examined the effect of
announcing a participatory exercise as a precautionary measure
rather that that of participation itself. In a deliberative democracy context, the framework used for this review could
prove a useful decision-making tool to communicate scientific
uncertainty to the wider public and manage uncertain risks. The
‘‘incidence trends’’ and ‘‘consequence’’ criteria in addition to
‘‘association’’ aim to assist risk management decisions, where
‘‘potentially serious or irreversible threats’’ give reason to
invoke the PP. Balancing the current state of knowledge or
evidence with trends of incidence, the probability of an
association and its potential consequence not only helps decide
which PA is appropriate and proportional both to the chosen
level of protection and available scientific evidence, but also
offers a framework by which to assess whether it is consistent
with similar measures that have been previously taken and by
which to review such risk management decisions in the light of
new scientific data.
ACKNOWLEDGMENTS
This study was supported by an Economic & Social Research Council
award (PTA-030-2003-00136). The authors are grateful to Dr Mark Scrimshaw
for constructive comments during preparation of the manuscript.
REFERENCES
Anderson, L. M. (2004). Introduction and overview. Perinatal carcinogenesis:
Growing a node for epidemiology, risk management, and animal studies.
Toxicol. Appl. Pharmacol. 199, 85–90.
Aravindakshan, J., Gregory, M., Marcogliese, D. J., Fournier, M., and Cyr, D.
G. (2004). Consumption of xenoestrogen-contaminated fish during lactation
alters adult male reproductive function. Toxicol. Sci. 81, 179–189.
Ashby, J. (2001). Testing for endocrine disruption post-EDSTAC: Extrapolation of low dose rodent effects to humans. Toxicol. Lett. 120, 233–242.
Axmon, A., Rylander, L., Stromberg, U., Dyremark, E., and Hagmar, L. (2001).
Polychlorinated biphenyls in blood plasma among Swedish female fish
consumers in relation to time to pregnancy. J. Toxicol. Environ. Health 64,
485 (Abstract).
Axmon, A., Rylander, L., Stromberg, U., and Hagmar, L. (2002). Female
fertility in relation to the consumption of fish contaminated with persistent
organochlorine compounds. Scand. J. Work Environ. Health 28, 124
(Abstract).
Aydin, Z. D. (2005). Sun exposure may confound physical activity-prostate
cancer association. Arch. Intern. Med. 165, 2538–2539.
Azziz, R., Marin, C., Hoq, L., Badamgarav, E., and Song, P. (2005). Health
care-related economic burden of the polycystic ovary syndrome during the
reproductive life span. J. Clin. Endocrinol. Metab. 90, 4650–4658.
Balasubramaniam, A., and Voulvoulis, N. (2005). The appropriateness of
multicriteria analysis in environmental decision-making problems. Environ.
Technol. 26, 951–962.
Batstone, G., Richard, D., Doble, A., and Batstone, D. (2003). Chronic
prostatitis. Curr. Opin. Urol. 13, 23–29.
Bellinger, D. C. (2004). Assessing environmental neurotoxicant exposures and
child neurobehavior: Confounded by confounding? Epidemiology 15,
383–384.
Birkett, J. W., and Lester, J. N. (2003). Endocrine Disrupters in Wastewater and
Sludge Treatment Processes. Lewis Publishers/IWA Publishing, London.
Birnbaum, L. S., and Cummings, A. M. (2002). Dioxins and endometriosis:
A plausible hypothesis. Environ. Health Perspect. 110, 15–21.
Blake, C. A., Boockfor, F. R., Nair-Menon, J. U., Millette, C. F., Raychoudhury,
S. S., and McCoy, G. L. (2004). Effects of 4-tert-octylphenol given in
drinking water for 4 months on the male reproductive system of Fischer 344
rats. Reprod. Toxicol. 18, 43–51.
Boisen, K. A., Kaleva, M., Main, K. M., Virtanen, H. E., Haavisto, A. M.,
Schmidt, M., Chellakooty, M., Damgaard, N., Mau, C., Reunanen, M., et al.
HUMAN HEALTH AND ENDOCRINE DISRUPTION
(2004). Difference in prevalence of congenital cryptorchidism in infants
between two Nordic countries. Lancet 363, 1264–1269.
Bray, F., Richiardi, L., Ekbom, A., Pukkala, E., Cuninkova, M., and Moller, H.
(2006). Trends in testicular cancer incidence and mortality in 22 European
countries: Continuing increases in incidence and declines in mortality. Int. J.
Cancer 118, 3099–3111.
Brinton, L. A., Westhoff, C. L., Scoccia, B., Lamb, E. J., Althuis, M. D., Mabie,
J. E., and Moghissi, K. S. (2005). Causes of infertility as predictors of
subsequent cancer risk. Epidemiology 16, 500–507.
Brody, J. G., and Rudel, R. A. (2003). Environmental pollutants and breast
cancer. Environ. Health Perspect. 111, 1007–1019.
Calabrese, E. J., and Baldwin, L. A. (2001). Hormesis: U-shaped dose
responses and their centrality in toxicology. Trends Pharmacol. Sci. 22,
285–291.
Catalano, R., Bruckner, T., Marks, A. R., and Eskenazi, B. (2006). Exogenous
shocks to the human sex ratio: The case of September 11, 2001 in New York
City. Hum. Reprod. 21, 3127–3131.
Catalano, R. A. (2003). Sex ratios in the two Germanies: A test of the economic
stress hypothesis. Hum. Reprod. 18, 1972–1975.
Chen, C. W. (2001). Assessment of endocrine disruptors: Approaches, issues,
and uncertainties. Folia Histochem. Cytobiol. 39, 20–23.
Coggon, D., Rose, G., and Barker, D. J. P. (2003). Epidemiology for the
Unititiated. BMJ Publishing Group. London.
Colborn, T. (2004). Neurodevelopment and endocrine disruption. Environ.
Health Perspect. 112, 944–949.
Cramer, D. W., and Wise, L. A. (2000). The epidemiology of recurrent
pregnancy loss. Semin. Reprod. Med. 18, 331 (Abstract).
Dewey, K. G. (2001). Maternal and fetal stress are associated with impaired
lactogenesis in humans. J. Nutr. 131, 3012S–3015S.
Diamanti-Kandarakis, E., and Piperi, C. (2005). Genetics of polycystic ovary
syndrome: Searching for the way out of the labyrinth. Hum. Reprod. Update
11, 631–643.
Dietrich, K. N., Eskenazi, B., Schantz, S., Yolton, K., Rauh, V. A., Johnson,
C. B., Alkon, A., Canfield, R. L., Pessah, I. N., and Berman, R. F. (2005).
Principles and practices of neurodevelopmental assessment in children:
Lessons learned from the Centers for Children’s Environmental Health
and Disease Prevention Research. Environ. Health Perspect. 113,
1437–1446.
Duty, S. M., Silva, M. J., Barr, D. B., Brock, J. W., Ryan, L., Chen, Z. Y.,
Herrick, R. F., Christiani, D. C., and Hauser, R. (2003). Phthalate exposure
and human semen parameters. Epidemiology 14, 269–277.
Edwards, T. M., Moore, B. C., and Guillette, L. J. (2006). Reproductive
dysgenesis in wildlife: A comparative view. Int. J. Androlog. 29, 109–121.
European Environment Agency (EEA) (2001). Late lessons from early
warnings: The precautionary principle 1896–2000. European Environment
Agency, Copenhagen.
Egozcue, J., Blanco, J., Anton, E., Egozcue, S., Sarrate, Z., and Vidal, F. (2003).
Genetic analysis of sperm and implications of severe male infertility—
A review. Placenta 24(Suppl 2), S62–S65.
Faust, M., Altenburger, R., Backhaus, T., Blanck, H., Boedeker, W., Gramatica,
P., Hamer, V., Scholze, M., Vighi, M., and Grimme, L. H. (2003). Joint algal
toxicity of 16 dissimilarly acting chemicals is predictable by the concept of
independent action. Aquat. Tox. 63, 43–63.
Ferguson, S. A., Flynn, K. M., Delclos, K. B., Newbold, R. R., and Gough, B. J.
(2002). Effects of lifelong dietary exposure to genistein or nonylphenol on
amphetamine-stimulated striatal dopamine release in male and female rats.
Neurotoxicol. Teratol. 24, 37–45.
Fisher, J. S. (2004). Environmental anti-androgens and male reproductive
health: Focus on phthalates and testicular dysgenesis syndrome. Reproduction 127, 305–315.
345
Friederichs, V., Cameron, J. C., and Robertson, C. (2006). Impact of adverse
publicity on MMR vaccine uptake: A population based analysis of vaccine
uptake records for one million children, born 1987-2004. Arch. Dis. Child.
91, 465–468.
Funabashi, T., Sano, A., Mitsushima, D., and Kimura, F. (2003). Bisphenol
A increases progesterone receptor immunoreactivity in the hypothalamus in
a dose-dependent manner and affects sexual behaviour in adult ovariectomized rats. J. Neuroendocrinol. 15, 134–140.
Gee, D. (2006). Late lessons from early warnings: Toward realism and
precaution with endocrine-disrupting substances. Environ. Health Perspect.
114, 152–160.
Giwercman, A., Rylander, L., Hagmar, L., and Giwercman, Y. L. (2006). Ethnic
differences in occurrence of TDS—Genetics and/or environment? Int. J.
Androl. 29, 291–297.
Gray, K. A., Klebanoff, M. A., Brock, J. W., Zhou, H., Darden, R., Needham,
L., and Longnecker, M. P. (2005). In utero exposure to background levels of
polychlorinated biphenyls and cognitive functioning among school-age
children. Am. J. Epidemiol. 162, 17–26.
Grover, P. L., and Martin, F. L. (2002). The initiation of breast and prostate
cancer. Carcinogenesis 23, 1095–1102.
Gundert-Remy, U., Kremers, P., Renwick, A., Kopp-Schneider, A., Dahl, S. G.,
Boobis, A., Oberemm, A., and Pelkonen, O. (2005). Molecular approaches to
the identification of biomarkers of exposure and effect—Report of an expert
meeting organized by COST action B15. November 28, 2003. Toxicol. Lett.
156, 227–240.
Habermacher, G. M., Chason, J. T., and Schaeffer, A. J. (2006). Prostatitis/
chronic pelvic pain syndrome. Annu. Rev. Med. 57, 195–206.
Hanke, W., and Jurewicz, J. (2004). The risk of adverse reproductive and
developmental disorders due to occupational pesticide exposure: An overview of current epidemiological evidence. Int. J. Occup. Med. Environ.
Health 17, 223 (Abstract).
Hart, R., Hickey, M., and Franks, S. (2004). Definitions, prevalence and
symptoms of polycystic ovaries and polycystic ovary syndrome. Best Pract.
Res. Clin. Obstet. Gynaecol. 18, 671–683.
Heilier, J. -F., Nackers, F., Verougstraete, V., Tonglet, R., Lison, D., and
Donnez, J. (2005). Increased dioxin-like compounds in the serum of women
with peritoneal endometriosis and deep endometriotic (adenomyotic)
nodules. Fertil. Steril. 84, 305–312.
Hemminki, K., and Chen, B. (2006). Familial risks in testicular cancer as
aetiological clues. Int. J. Androl. 29, 205–210.
Hertz-Picciotto, I. (1998) Environmental Epidemiology. In Modern Epidemiology. Rothman, K. J. and Greenland, S. (eds), 2nd ed. pp.555–583.
Lippincott-Raven Publishers, Philadelphia.
Hill, A. B. (1965). The environment and disease: Association or causation?
Proc. R. Soc. Med. 58, 295–300.
Hogan, M. D., Newbold, R. R., and McLachlan, J. A. (1987). Extrapolation of
Teratogenic Responses Observed in Laboratory Animals to Humans. Branbury Report 26: Development Toxicology: Mechanisms and Risk. Cold
Spring Harbor Laboratory, New York.
International Programme on Chemical Safety (2002). Global Assessment of the
State-of-the-Science of Endocrine Disruptors. World Health Organisation,
Geneva, Switzerland. 180 p.
International Programme on Chemical Safety (2006). Coordinator’s Report on
IPCS Progress and Activities—June 2004 to December 2005. WHO, Geneva,
Switzerland. p. 29.
James, W. H. (2006). Offspring sex ratios at birth as markers of paternal
endocrine disruption. Environ. Res. 100, 77–85.
Jensen, T. K., Andersson, A. -M., Jorgensen, N., Andersen, A. -G., Carlsen, E.,
Petersen, J. o. H., and Skakkebaek, N. E. (2004). Body mass index in relation
to semen quality and reproductive hormones among 1,558 Danish men.
Fertil. Steril. 82, 863–870.
346
MARTIN ET AL.
Kabesch, M. (2006). Gene by environment interactions and the development of
asthma and allergy. Toxicol. Lett. 162, 43–48.
Karmaus, W., Davis, S., Fussman, C., and Brooks, K. (2005). Maternal
concentration of dichlorodiphenyl dichloroethylene (DDE) and initiation
and duration of breast feeding. Paediatr. Perinat. Epidemiol. 19, 388–398.
Kopp, H. G., Kuczyk, M., Classen, J., Stenzl, A., Kanz, L., Mayer, F., Bamberg,
M., and Hartmann, J. T. (2006). Advances in the treatment of testicular
cancer. Drugs 66, 641 (Abstract).
Kort, H. I., Massey, J. B., Elsner, C. W., Mitchell-Leef, D., Shapiro, D. B., Witt,
M. A., and Roudebush, W. E. (2006). Impact of body mass index values on
sperm quantity and quality. J. Androl. 27, 450–452.
Landrigan, P., Garg, A., and Droller, D. B. J. (2003). Assessing the effects of
endocrine disruptors in the National Children’s Study. Environ. Health
Perspect. 111, 1678–1682.
Laviola, G., Gioiosa, L., Adriani, W., and Palanza, P. (2005). D-amphetaminerelated reinforcing effects are reduced in mice exposed prenatally to
estrogenic endocrine disruptors. Brain Res. Bull. 65, 235–240.
Leenhardt, L., Grosclaude, P., and Chérié-Challine, L. (2004). Increased
incidence of thyroid carcinoma in France: A true epidemic or thyroid nodule
management effects? Report from the French Thyroid Cancer Committee.
Thyroid 14, 1056–1060.
Newbold, R. R., Hanson, R. B., Jefferson, W. N., Bullock, B. C., Haseman, J.,
and McLachlan, J. A. (2000). Proliferative lesions and reproductive tract
tumors in male descendants of mice exposed developmentally to diethylstilbestrol. Carcinogenesis 21, 1355–1363.
Nickel, J. C., Downey, J., Hunter, D., and Clark, J. (2001). Prevalence of
prostastitis-like symptoms in a population based study using the national
institutes of health chronic prostatitis symptom index. J. Urol. 165,
842–845.
Onn Wong, K., Woon Leo, L., and Leng Seah, H. (2005). Dietary exposure
assessment of infants to bisphenol A from the use of polycarbonate baby
milk bottles. Food Addit. Contamin. 22, 280–288.
Owens, W., and Koeter, H. B. W. M. (2003). The OECD program to validate the
rat uterotrophic bioassay: An overview. Environ. Health Perspect. 111,
1527–1529.
Ozaki, A., Yamaguchi, Y., Okamoto, A., and Kawai, N. (2002). Determination
of alkylphenols, bisphenol A, benzophenone and phthalates in containers of
baby food, and migration into food simulants. Shokuhin Eiseigaku Zasshi
43, 260 (Abstract).
Palapattu, G. S., Sutcliffe, S., Bastian, P. J., Platz, E. A., De Marzo, A. M.,
Isaacs, W. B., and Nelson, W. G. (2004). Prostate carcinogenesis and
inflammation: Emerging insights. Carcinogenesis 26, 1170–1181.
Le Gales, C., and Moatti, J. P. (1990). Searching for consensus through multicriteria decision analysis. Assessment of screening strategies for hemoglobinopathies in southeastern France. Int. J. Technol. Assess. Health Care 6,
430 (Abstract).
Palmer, J. R., Hartge, P., Noller, K. L., Kaufman, R. H., Hatch, E. E.,
Rosenberg, C. L., Rao, R. S., Titus-Ernstoff, L., and Herbst, A. L. (2002).
Risk of breast cancer in women exposed to diethylstilbestrol in utero:
Preliminary results (United States). Cancer Causes Control 13, 753–758.
Lovelady, C. A. (2005). Is maternal obesity a cause of poor lactation performance. Nutr. Rev. 63, 352 (Abstract).
Markaverich, B., Mani, S., Alejandro, M. A., Mitchell, A., Markaverich, D.,
Brown, T., Velez-Trippe, C., Murchison, C., O’Malley, B., and Faith, R.
(2002). A novel endocrine-disrupting agent in corn with mitogenic activity in
human breast and prostatic cancer cells. Environ. Health Perspect. 110,
169–177.
Parent, A. S., Teilmann, G., Juul, A., Skakkebaek, N. E., Toppari, J., and
Bourguignon, J. P. (2003). The timing of normal puberty and the age limits of
sexual precocity: Variations around the world, secular trends, and changes
after migration. Endocr. Rev. 24, 668–693.
Partsch, C. J., and Sippell, W. G. (2001). Pathogenesis and epidemiology of
precocious puberty. Effects of exogenous oestrogens. Hum. Reprod. Update
7, 292–302.
Marks, G. B. (2006). Environmental factors and gene-environment interactions
in the aetiology of asthma. Clin. Exp. Pharmacol. Physiol. 33, 285–289.
Patisaul, H. B. (2005). Phytoestrogen action in the adult and developing brain.
J. Neuroendocrinol. 17, 57–64.
Martin, O. V., Lai, K. M., Scrimshaw, M. D., and Lester, J. N. (2005). Receiver
operating characteristic analysis for environmental diagnosis. A potential
application to endocrine disruptor screening: in vitro estrogenicity bioassays.
Environ. Sci. Technol. 39, 5349–5355.
Pearce, N., and Douwes, J. (2006). The global epidemiology of asthma in
children. Int. J. Tuberc. Lung Dis. 10, 125 (Abstract).
Massart, F., Harrell, J. C., Federico, G., and Saggese, G. (2005a). Human breast
milk and xenoestrogen exposure: A possible impact on human health.
J. Perinatol. 25, 282–288.
Massart, F., Seppia, P., Pardi, D., Lucchesi, S., Meossi, C., Gagliardi, L.,
Liguori, R., Fiore, L., Federico, G., and Saggese, G. (2005b). High
incidence of central precocious puberty in a bounded geographic area of
northwest Tuscany: An estrogen disrupter epidemic? Gynaecol. Endocrinol. 20, 92–98.
Masuo, Y., Morita, M., Oka, S., and Ishido, M. (2004). Motor hyperactivity
caused by a deficit in dopaminergic neurons and the effects of endocrine
disruptors: A study inspired by the physiological roles of PACAP in the
brain. Regul. Pept. 123, 225–234.
McGuinness, B. M., Buck, G. M., Mendola, P., Sever, L. E., and Vena, J. E.
(2001). Infecundity and consumption of polychlorinated biphenylcontaminated fish. Arch. Environ. Health 56, 250–253.
Morgenstern, H. (1982). Uses of ecologic analysis in epidemiologic research.
Am. J. Public Health 72, 1336–1344.
Nagataki, S., and Nyström, E. (2002). Epidemiology and primary prevention of
thyroid cancer. Thyroid 12, 889–896.
Newbold, R. R., Hanson, R. B., Jefferson, W. N., Bullock, B. C., Haseman, J.,
and McLachlan, J. A. (1998). Increased tumors but uncompromised fertility
in the female descendants of mice exposed developmentally to diethylstilbestrol. Carcinogenesis 19, 1655–1663.
Pillay, O. C., Te Fong, L. F. W., Crow, J. C., Benjamin, E., Mould, T.,
Atiomo, W., Menon, P. A., Leonard, A. J., and Hardiman, P. (2006). The
association between polycystic ovaries and endometrial cancer. Hum.
Reprod. 21, 924–929.
Pocock, S. J., Geller, N. L., and Tsiatis, A. A. (1987). The analysis of multiple
endpoints in clinical trials. Biometrics 43, 487–498.
Porpora, M. G., Ingelido, A. M., Domenico, A. d., Ferro, A., Crobu, M.,
Pallante, D., Cardelli, M., Cosmi, E. V., and De Felip, E. (2006). Increased
levels of polychlorobiphenyls in Italian women with endometriosis. Chemosphere 63, 1361–1367.
Poulos, L. M., Toelle, B. G., and Marks, G. B. (2005). The burden of asthma in
children: An Australian perspective. Paediatr. Respir. Rev. 6, 20–27.
Purdue, M. P., Devesa, S. S., Sigurdson, A. J., and McGlynn, K. A. (2005).
International patterns and trends in testis cancer incidence. Int. J. Cancer
115, 822–827.
Remontet, L., Esteve, J., Bouvier, A. -M., Grosclaude, P., Launoy, G.,
Menegoz, F., Exbrayat, C., Tretare, B., Carli, P. -M., Guizard, A. -V., et al.
(2003). Cancer incidence and mortality in France over the period 1978-2000.
Rev. Epidemiol. Sante Publique 51, 3 (Abstract).
Reynolds, P., Hurley, S. E., Goldberg, D. E., Yerabati, S., Gunier, R. B., Hertz, A.,
Anton-Culver, H., Bernstein, L., Deapen, D., Horn-Ross, P. L., et al. (2004).
Residential proximity to agricultural pesticide use and incidence of breast
cancer in the California Teachers Study cohort. Environ. Res. 96, 206–218.
Richiardi, L., Bellocco, R., Adami, H. -O., Torrång, A., Barlow, L., Hakulinen,
T., Rahu, M., Stengrevics, A., Storm, H., Tretli, S., et al. (2004). Testicular
HUMAN HEALTH AND ENDOCRINE DISRUPTION
cancer incidence in eight northern European countries: Secular and recent
trends. Cancer Epidemiol. Biomark. Prev. 13, 2157–2166.
Ritchie, J. M., Vial, S. L., Fuortes, L. J., Robertson, L. W., Guo, H., Reedy, V.
E., and Smith, E. M. (2005). Comparison of proposed frameworks for
grouping polychlorinated biphenyl congener data applied to a case-control
pilot study of prostate cancer. Environ. Res. 98, 104–113.
347
Spanò, M., Toft, G., Hagmar, L., Eleuteri, P., Rescia, M., Rignell-Hydbom, A.,
Tyrkiel, E., Zvyezday, V., Bonde, J. P., and INUENDO (2005). Exposure to
PCB and p, p’-DDE in European and Inuit populations: Impact on human
sperm chromatin integrity. Hum. Reprod. 20, 3488–3499.
Steenland, K., and Savitz, D. A. (1997). Topics in Environmental Epidemiology.
Oxford University Press, New York.
Ronchetti, R., Villa, M. P., and Barreto, M. (2004). L’asma in eta pediatrica.
Minerva Pediatr. 56, 133 (Abstract).
Szpir, M. (2006). New thinking on neurodevelopment. Environ. Health
Perspect. 114, A100–A107.
Rozzi, C. C., Benassi, L. L., Parazzini, F. F., Chatenoud, L. L., La Vecchia, C.
C., and Benzi, G. G. (2000). Trends of spontaneous abortions in Italy 1990–
1995. Epidemiol. 11, 229–230.
Tao, M. H., Xu, W. H., Zheng, W., Gao, Y. T., Ruan, Z. X., Cheng, J. R., Xiang,
Y. B., and Shu, X. O. (2005). A case-control study in Shanghai of fruit and
vegetable intake and endometrial cancer. Br. J. Cancer 92, 2059–2064.
Rushton, L. (2000). Reporting of occupational and environmental research: Use
and misuse of statistical and epidemiological methods. Occup. Environ. Med.
57, 1–9.
Toft, G., Hagmar, L., Giwercman, A., and Bonde, J. P. (2004). Epidemiological
evidence on reproductive effects of persistent organochlorines in humans.
Reprod. Toxicol. 19, 5–26.
Safe, S. (2005). Clinical correlates of environmental endocrine disruptors.
Trends Endocrinol. Metab. 16, 139–144.
Toran-Allerand, C. D., Guan, X., MacLusky, N. J., Horvath, T. L., Diano, S.,
Singh, M., Connolly, E. S. J., Nethrapalli, I. S., and Tinnikov, A. A. (2002).
ER-X: A novel, plasma membrane-associated, putative estrogen receptor
that is regulated during development and after ischemic brain injury. J.
Neurosci. 22, 8391–8401.
Sakoda, L. C., and Horn-Ross, P. L. (2002). Reproductive and menstrual history
and papillary thyroid cancer risk: The San Francisco Bay Area thyroid cancer
study. Cancer Epidemiol. Biomark. Prev. 11, 51–57.
Salama, J., Chakraborty, T. R., Ng, L., and Gore, A. C. (2003). Effects of
polychlorinated biphenyls on estrogen receptor-beta expression in the anteroventral periventricular nucleus. Environ. Health Perspect. 111, 1278–1282.
Sallmén, M., Weinberg, C. R., Baird, D. D., Lindbohm, M. -L., and Wilcox, A. J.
(2005). Has human fertility declined over time?: Why we may never know.
Epidemiology 16, 494–499.
Sasco, A. J. (2003). Breast cancer and the environment. Horm. Res. 60, 50.
Schaeffer, A. J. (2003). Epidemiology and demographics of prostatitis.
Andrologia 35, 252–257.
Schaeffer, A. J., Alexander, R. B., Nadler, R. B., Propert, K. J., Nickel, J. C.,
Landis, J. R., Knauss, J. S., Litwin, M. S., and O’Leary, M. P. (2002).
Demographic and clinical characteristics of men with chronic prostatitis: The
national institutes of health chronic prostatitis cohort study. J. Urol. 168,
593–598.
Schantz, S. L., Widholm, J. J., and Rice, D. C. (2003). Effects of PCB exposure
on neuropsychological function in children. Environ. Health Perspect. 111,
357–576.
Schwartz, G. G. (2002). Hypothesis: Does ochratoxin A cause testicular
cancer? Cancer Causes Control 13, 91–100.
Seery, J. P. (2006). Endometriosis associated with defective handling of
apoptotic cells in the female genital tract is a major cause of autoimmune
disease in women. Med. Hypotheses 66, 945–949.
Sell, J., and Scholz, R. W. (2003). Consensus building and knowledge
integration methods applied to technology implementation in the Swiss waste
incineration industry. Int. J. Environ. Technol. Manag. 3, 368 (Abstract).
Sharpe, R. M., and Skakkebaek, N. E. (2003). Male reproductive disorders and
the role of endocrine disruption: Advances in understanding and identification of areas for future research. Pure Appl. Chem. 75, 2023–2038.
Silva, E., Rajapakse, N., and Kortenkamp, A. (2002). Something from
‘‘nothing’’—Eight weak estrogenic chemicals combined at concentrations
below NOECs produce significant mixture effects. Environ. Sci. Technol. 36,
1751–1756.
Sim, H. G., and Cheng, C. W. S. (2005). Changing demography of prostate
cancer in Asia. Eur. J. Cancer 41, 834–845.
Slama, R., Bouyer, J., Remontet, L., and Spira, A. (2004). Epidemiology of
male reproductive function: A field searching for tools. Rev. Epidemiol.
Sante Publique 52, 221 (Abstract).
Slovic, P. (1987). Perception of risk. 236, 280–285.
Tsukino, H., Hanaoka, T., Sasaki, H., Motoyama, H., Hiroshima, M., Tanaka,
T., Kabuto, M., Niskar, A. S., and Rubin, C. (2005). Associations between
serum levels of selected organochlorine compounds and endometriosis in
infertile Japanese women. Environ. Res. 99, 118–125.
Ullah, S., and Griffiths, P. (2003). Does the use of pacifiers shorten
breastfeeding duration in infants? Br. J. Community Nurs. 8, 458 (Abstract).
Unfer, V., Casini, M. L., Costabile, L., Mignosa, M., Gerli, S., and Di Renzo, G.
C. (2004). Endometrial effects of long-term treatment with phytoestrogens:
A randomized, double-blind, placebo-controlled study. Fertil. Steril. 82,
145–148.
van der Horst, M., Falck Winther, J., and Olsen, J. H. (2006). Cancer incidence
in the age range 0-34 years: Historical and actual status in Denmark. Int. J.
Cancer 118, 2816–2826.
van der Tweel, I. (2005). Repeated looks at accumulating data: To correct or not
to correct? Eur. J. Epidemiol. 20, 205–211.
Vorderstrasse, B. A., Fenton, S. E., Bohn, A. A., Cundiff, J. A., and Lawrence,
B. P. (2004). A novel effect of dioxin: Exposure during pregnancy severely
impairs mammary gland differentiation. Toxicol. Sci. 78, 248–257.
Waylen, A., and Wolke, D. (2004). Sex ‘n’ drugs ‘n’ rock ‘n’ roll: The meaning and social consequences of pubertal timing. Eur. J. Endocrin. 151,
U151–U159.
Wiedemann, P. M., and Schütz, H. (2005). The precautionary principle and risk
perception: Experimental studies in the EMF area. Environ. Health Perspect.
113, 402–405.
Witorsch, R. J. (2002). Low-dose in utero effects of xenoestrogens in mice and
their relevance to humans: An analytical review of the literature. Food Chem.
Toxicol. 40, 905–912.
World Health Organisation (WHO) (2001). Current Practices and Controversies in Assisted Reproduction. Medical, Ethical and Social Aspects of
Assisted Reproduction. World Health Organisation, Geneva, Switzerland.
World Health Organisation (WHO) (2006). Fact sheet No297. World Health
Organisation, Geneva, Switzerland.
World Health Organisation (WHO) (2004). World Health Report
2004—Changing History. World Health Organisation, Geneva, Switzerland.
96 p.
Zhu, Z., Boobis, A. R., and Edwards, R. J. (2006). Use of protein profiles to
characterise concentration-effect curves of mixtures of estrogenic compounds in human breast cell lines. Toxicol. Lett. 164, S165–S166.