A review of environmental and human health risk assessment frameworks of multiple stressors: the case of endocrine disruptors


This review is a compilation of the recommended approaches and methods for the development of a risk assessment framework of multiple stressors. Some of the driving forces contributing to address this subject are the current demands of stakeholders like the drinking water industry, the society and regulators of evaluating the risks of mixtures of contaminants that may harm human beings and the environment. Therefore, our work aims at exploring the possibility of integrating within the risk assessment, environmental, human and societal aspects, acknowledging new international regulations and policies for the safe use of chemicals which enforce the integrative study of the hazards of multiple chemicals on humans and the environment throughout their life-cycle. We chose the group of compounds called endocrine disruptors as an example of multiple stressors because of their emerging relevance and the fact that they usually integrate complex mixtures, originate from multiple sources and exist in several environmental compartments, generating adverse effects on receptors through different routes and pathways. Their actions may be severe chronic and long-term modifications of the normal development and reproduction patterns of the individual and/or its progeny, eventually representing systemic risks at the population level which may affect sustainability and biodiversity. Due to the ubiquity of these chemicals, it is necessary to address the inclusion of human beings as potential receptors and deal with risk assessment in an integrated manner. As endocrine disruptors may provoke differentiated responses relative to the developmental stage, acting through varied mechanisms and at very low concentrations, with the particularity that their toxicokinetics may present sometimes unusual dose-response curves, might challenge long-term predictions and hazard characterization, adding to risk assessment uncertainties. References to the current methodologies including the applicable bioassays, chemical analysis, modeling, statistics tools and equations to calculate joint effects considering the interactions of toxicants within a mixture are also discussed in this review.


risk assessment, endocrine disruptors, multiple stressors

1. Introduction

The aim of this review is to analyze the risk assessment frameworks and current practices, the integration of environmental and human health methodologies, the effects evaluation and how to assess the risk of complex mixtures of chemicals. Assessing the risks of multiple stressors for human health and the environment arises from the realization that biological and physical stressors usually coexist in complex mixtures in the natural and constructed environment, sometimes generating impacts on living creatures.

Historically, since the publication in 1962 of the book “Silent Spring”, by Rachel Carson, a warning on the deleterious actions on wildlife of persistent toxicants, such as DDT, which caused a decrease of birds in Pennsylvania, became presentwas recognised among the scientific world, the media and the society. More recently, the research by Dr. Theo Colborn on the reproductive disorders of predators in the Great Lakes of North America and her book, published in 1996 “Our Stolen Future” co-authored by herself with Dianne Dumanoski and John Peterson Myers, was able to generate public awareness on the threats that EDCs might pose to human beings and the environment.

Some of the driving forces for writing this review are the demands of stakeholders represented by the community, the NGOs, the health and environmental regulators, the manufacture industrial sector and the drinking water companies. For instance, the European Environment and Health Strategy emphatically stresses the gaps in knowledge related to risk assessment methodologies that include foetuses, infants and children and calls for the precautionary principle within the strategy for environmental contaminants, for which there is a sufficient level of scientific evidence at the effect level (e.g. molecular, cellular, or tissue-related) to show the likelihood of health impacts. Not enough information exists on the link between emissions of dioxins and PCBs and other substances and their accumulation in ecosystems and foodstuffs. The need for research on the understanding of the links between environmental factors and certain diseases is recognized, but, due to the complexity of the issue, the immediate action is to gather evidence in order to concrete measures to protect human health and the environment.

Many reports are able to demonstrate through laboratory experiments and field surveys that exposure of animals to chemicals released into the environment exert reproductive or developmental effects on the individual and its off-spring, affecting the viability of the species at a population level (Colborn and Smolen 1997). As a matter of fact, these kind of adverse effects have been observed in wildlife and many of them can be attributed to the exposure to man-made chemicals. The cause-effect relationship is still controversial in human beings, but still a matter of concern due to the severity of probable harm that could affect individuals and populations. Thus, regulatory decisions must be informed by risk assessment on this important topic (Fenner-Crisp 2000).

Endocrine-mediated effects may be more relevant in populations rather than in individuals. As there is still not consensus about a cause-effect relationship, it is recommended a science-based precautionary approach to protect public health. Further research is needed to demonstrate effects and carry out birth defect registries and epidemiology studies designed to track delayed effects of environmental exposures (Solomon and Schettler 2000).

The classical paradigm of human health risk assessment authored by the National Research Council (NRC, 1983) is composed of four steps: hazard identification, dose-response assessment, exposure assessment and risk characterization. This paradigm was modified in 1994 to include characterization of each component. One of the approaches considered by some authors as best suited for developing a risk assessment of multiple stressors is a cumulative risk assessment framework, which may include societal aspects with participatory involvement of stakeholders (Gentile and Harwell 2001; Callahan and Sexton 2007; USEPA 2003).

The global trend towards a policy-driven integration applied to risk assessment, demanded by regulations on safety of chemicals and industrial operations should ideally include a multiplicity of stressors, compartments, geographical scales, and end-points (Assmuth and Hildén 2008). For example, the new European Union Regulation on chemicals and its safe use (REACH, EC 1907/2006) enforces linking risks to human health and the environment for chemicals throughout their life cycle. In United States, the Environmental Protection Agency, still discusses both topics separately because of practical reasons, but acknowledging the need to integrate them.

An overview of the most suitable risk assessment frameworks is described in this literature review, focusing on the case of mixtures of reproductive endocrine disruptors. Matters that differentiate this topic are also presented, regarding particularities in mechanistic and toxicokinetics aspects and some of the sources of uncertainties and confounding factors. Developing a novel approach to the classical risk assessment scheme is also a goal, with the intention of contributing to design a risk assessment framework comprising a choice of tests, models, computational and statistical tools.

2. Hazard identification

In this stage the nature of the hazard is described taking into account toxicity data. The hazard can then be characterized deriving numerical values of acceptability of the risk, based on mechanisms of action, biological extrapolation, dose-response and dose-effect relationships, and related uncertainties (Fenner-Crisp, 2003).

2.1. Nature of the hazard

Endocrine disruptors (EDCs) are substances that interfere with the endocrine system by changing homeostasis and producing developmental effects and/or diminishing the fertility of the organisms. EDCs include a broad range of substances which can be classified according to their effect. The best known are the environmental estrogens, alkylphenol and its ethoxylates, the monomer in polycarbonate manufacture bisphenol A, some pesticides and chlorinated organics.

2.2. Sources of EDCs

Possible sources of watercourses pollution with EDCs are wastewater sewage discharge, industrial effluents, or point and non-point source contamination of watercourses with agrochemicals such as herbicides or pesticides.

Sewage discharge from cities contains residues of domestic products such as personal care products, pharmaceuticals and detergents and excreta with natural and artificial steroidal hormones from contraceptive pill usage. Nonylphenol (NP) is a highly hydrophobic bioccumulating biodegradation by-product of nonylphenol ethoxylate non-ionic surfactants which persists in sewage sludge and river sediments. Its use and production have been banned in EU countries and strictly monitored in many other countries such as Canada and Japan (Soares et al. 2008).

Pulp mills are some of the industries associated to studies on endocrine disruption. Bleached Kraft pulp mill effluents have been linked to both estrogenic and androgenic effects on biota, depending on the process characteristics and wastewater treatment. Several studies have associated the chlorination of organic matter to the estrogenicity of the effluent. Nowadays, the application of elemental-chlorine-free processes has diminished the risk of dioxins and furans formation, but not eliminated it, as other halogenated organics are formed by use of chlorine dioxide as bleaching agent. Various wood-extractive compounds produced in the pulping process, such as rosin acids and phytosterols and found in pulp mills effluents have also been considered potentially responsible of endocrine disruption (Hewitt et al. 2008). The main identified resin acids in pulp mill effluents are: pimaric, isopimaric, sandaracopimaric, palustric, dehydroabietic, abietic and neoabietic acid (Meriläinen and Oikari, 2008). Other compounds found in this type of effluent are phenolic guaiacyl-based lignin degradation products, diterpenoids, and dimethoxy stilbene (Belknap et al, 2006). Modern analytical methods, like full-scan GC-MS have been used to identify wood related extractives in final effluent from a chlorine dioxide bleached pulp mill effluent, including monoterpenes, phenolics, fatty acids, resin acids, resin acid neutrals and sterols (Wartman et al. 2009). Receptor binding bioassays for androgens and estrogens indicated that androgens were most abundant in this effluent and the range of androgens for the various extraction protocols used was 189-283 ng/L as testosterone equivalent concentration.

Some examples of common sources of EDCs and typical environmental concentrations are summarized on Table 1.

Table 1.Sources of EDCs and typical environmental concentrations

Origin, use and occurrence

Source of environmental exposure

EDC group

Example molecule

Typical concentrations


Industrial (pulp and paper mills)

Contaminated fish

Resin acids

pimaric acid

4-140 µg g-1

Owens et al, 1994

Industrial (pulp and paper mills)

Industrial wastewater treatment plant

Chlorinated organics


1.5 µg l-1

Owens et al, 1994

Industrial (pulp and paper mills)

Final stage secondary treatment



58.42 µg l-1

Landman et al, 2008


(contraceptive pills)

Sewage effluent



14-17 ng l-1

Liu et al, 2004

Human and animal excreta

Sewage effluent

Natural steroid hormones


5.0 ng l-1

Koh YKK et al,


Domestic and industrial (laundry detergents, wool scouring processes)

Sewage sludge

Non ionic surfactants


829.3 mg/kg

González et al, 2010

Domestic and industrial (polycarbonate bottles)

Leaching from solid waste, sewage effluent


bisphenol A

0.62 µg l-1

Sánchez-Avila et al, 2009

Agricultural (soil fertilization)

Livestock waste

Male steroid hormones


10-1830 ng l−1

Lange et al, 2002

Agricultural (dairy farming)

Streams contaminated by dairy cow excreta

Female steroid hormones


0.04-3.6 ng l−1

Matthiessen et al, 2006


(weed and grass control in soybean crops)




0.1-0.7 mg l-1

Peruzzo et al, 2008

2.2. Dose-response assessment

There are several methods to demonstrate dose-response relationships, either by in vivo or in vitro tests. Fish reproduction tests, like the ones using the model fish fathead minnow (Pimephales promelas) have shown a decrease in fecundity associated with depressed steroid and vitellogenin (Vtg) production in female specimens (Ankley et al. 2008). Many of the tests rely on the measurement of an increase of Vtg as biomarker of estrogenicity as seen in several publications (Schwaiger et al. 2002; An et al. 2008; Holbech et al. 2006 Panter et al, 1998; Sohoni et al. 2001; Kunz and Fent, 2009).

Tests results on resin acids show different responses in the first generation of fish than in the second (Christianson-Heiska et al. 2007).

In some cases there are not many examples of in vivo tests, like for glyphosate. A fish exposure tests with this compound showed Vtg induction in female fish, indicative of estrogenic activity (An et al, 2008). An investigation working the commercial formulation of the herbicide glyphosate and human placental cells demonstrated its toxicity at concentrations lower than the usual in agricultural practices. The aromatase activity disruption seems to be due not only to glyphosate but also to co-adjuvants (the surfactant nonylphenol or others), which enhance its bioavailability and/or bioaccumulation (Richard et al. 2005; Gasnier et al. 2009). Table 2 shows some examples of dose-response experiments working with fish, crustacea, molluscs and amphibia. Varied protocols exist to develop ecotoxicity tests, in flow-through, static or partly renewal conditions, and with different duration and end-points. Only chronic effects and particularly developmental and reproductive effects were considered.

Table 2. Dose-response for endocrine disruption effects in freshwater organisms exposed to single EDCs

EDC chemical name

Taxonomic group


Dose to produce effect


Test conditions




Rivulus marmoratus


µg l-1

Testicular agenesis and oogenesis inhibition in 100 % fish

Static system, daily renewal

Tanaka and Grizzle, 2002



Oncorhynchus mykiss, rainbow trout

1 -10 µg l-1

10 µg l-1

High Vtg in adult male fish plasma

Low hatching rate

Intermittent exposure of adult fish for 4 months until spawning

Schwaiger et al, 2002



Ceriodaphnia dubia

NOEC for reproduction: 1 µg l-1

Low hatching rate

7 days chronic exposure, static

Isidori et al, 2005



Carassius carassius, crucian carp

100% effluent

Vtg induction in female fish (38.6 +/- 9.8 µg l-1)

3 weeks, continuous exposure

An et al, 2008



Pseudosuccinea columella, aquatic snail

1 mg l-1

10 mg l-1

Faster development of F3 embryos

Hatching inhibition

3 generation continuous

Tate et al, 1997



Danio rerio

LOEC: 14 ng l-1

50 ng l-1

Significant Vtg increase

Higher female ratio

40 days fish sexual development test

Holbech et al, 2006



Danio rerio

LOEC: 54 ng l-1

Significant Vtg increase

Higher female ratio

40 days fish sexual development test

Holbech et al, 2006



Pimephales promelas, fathead minnow

100 ng l-1

Significant Vtg increase

Testicular growth inhibition

21 days male fish exposure

Panter et al, 1998



Danio rerio

LOEC: 0.6 µg l-1

21.7 µg l-1

Significant Vtg increase

Higher female ratio

40 days fish sexual development test

Holbech et al, 2006

Dehydroabietic acid (DHAA), resin acid


Danio rerio, zebra fish

50 µg l-1

Low plasma Vtg in female in F0; high Vtg and affected spermatogenesis in F1 males

2 generations, continuous

Christianson-Heiska et al 2008



Danio rerio

10-20 µg l-1

F1: higher ratio of male fish; F2: higher ratio of female fish

2 generation fish exposure test

Nakari and Erkomaa, 2003

Quercetin, phytoestrogen


Xenopus laevis, frog

200 µg l-1

Higher female ratio

> 10% abnormal testes (some with ovotestes)

Exposure up to 1 month post-metamorphosis

Cong et al, 2006

Phenanthrene, PAH


Oryzias latipes,


NOEL: 100 µg l-1

Developmental, hatching

18 days, renewal

Rhodes et al, 2005

Bisphenol A


Marisa cornuaretis, aquatic snail

NOEC: 640 µg l-1


12 weeks, juvenile snails

Forbes et al, 2007

Bisphenol A


Pimephales promellas

16 µg l-1

640 and 1280 µg l-1

640 µg l-1

1280 µg l-1

Altered spermatogenesis

Growth inhibition and Vtg induction in male fish

Reduced hatchability in F1 generation

Egg production inhibition

3 generation reproduction exposure test

Sohoni et al, 2001

Bisphenol A


Brachydanio rerio, zebrafish

EC50: 2.90 µg l-1

Embryo malformation and low hatchability

72 h exposure

Liu et al, 2007

Benzo-α-pirene (BaP) (PAH)



heteroclitus , common mummichog

10 µg l-1

CYP19A1 expression decreased by about 50% in immature stage I oocytes


for 10 or 15 days by in situ hybridization, several developmental stages

Dong et al, 2008

Polychlorinated biphenyl 126


Danio rerio, zebrafish

LC50: 3.270 mg l-1

Developmental effects through aryl hydrocarbon receptor activation

Dilutions of PCB 126 for 12 weeks

Sişman et al, 2007

Polychlorinated biphenyl 126


Salvelinus namaycush, lake trout

3 μg kg−1body weight

Retinol depletion

Oral exposure for 12 weeks; confirmation with radiolabelled retinol

Palacea et al, 1997



Pimephales promelas

4919 µg l-1

Vtg induction

14 days exposure, semi-static, renewal

Kunz and Fent, 2009

3. Exposure assessment

3.1. Ecosystems and human sub-populations potentially at risk of endocrine disruption effects

Increasing evidence generated by scientists turn endocrine disruption into a recognized risk to the environment. Due to the ubiquity of EDCs and the widespread routes of exposure, most ecosystems and human populations are potentially at risk of endocrine disruption. Notwithstanding this fact, under the scope of a risk assessment of EDCs the potentially most vulnerable risk subgroups are identified corresponding to maternal, fetal and early developmental stages. The concern that prenatal or childhood exposure to EDCs may be responsible for abnormalities in human sexual and reproductive health are still in the hypothetical ground. However, many reports on exposure to high concentrations of recognized EDCs such as DES, certain PCBs, and DDT demonstrate this fact. At low-doses the question remains unanswered whether there could be a critical window where they could harm the fetal development (Hood 2005).

Several reports on human developmental anomalies and reproductive ailments have been raising international concern, such as a seven fold increase risk of testicular cancer since 1969 to 2002 in men from several countries of Europe, United States and New Zealand. Also, the sperm density halved, as rates of cryptorchidism (undescended testicles) and hypospadias (shortened urinary tracts) simultaneously rose. It is thought that human congenital malformation of sex organs, low sperm quality, endometriosis, reduced fertility and some types of cancers of breast and testis could be linked to exposure to EDCs. More than 80000 synthetic chemicals are produced in the world and have still not been fully evaluating with regards to endocrine disruption. In 1996, the U.S. Environmental Protection Agency initiated an Endocrine Disruption Screening Program to evaluate more than 15,000 chemicals calling for a policy based on the “precautionary approach” to be overcautious and protect human health and the environment. A historical example of policies which demanded the banning of a drug due to these after-effects is the case of diethylstilbestrol (DEADES), which used to be prescribed to pregnant women to prevent spontaneous abortions because it produced higher risk of genital deformities and cancer in the offspring, among other effects (Stair 2008).

Internationally there is consensus that the most vulnerable group for EDCs exposure are children. For example, in European countries, the Strategy for Environment and Health known as “SCALE” for Science, Children, Awareness, Legislation and Evaluation, has set as a priority agenda for the evaluation diseases caused by endocrine disruptors in children.

The exposure to insecticides and herbicides used in agricultural practices has been linked to developmental or reproductive effects in wild animals and also in human beings. The occupational exposure to pesticide has received much attention, as for example prolonged time-to-pregnancy was observed in male greenhouse workers exposed to pesticides before conception of their first pregnancy (Bretveld et al 2008). The domestic exposure of children to residues of pesticides in low-level long-term exposures are associated to chronic effects and include routes of exposure such as fruit or breast milk (Goodman and Laverda 2002).

3.2. Evidence of endocrine disruption effects in wildlife around the world

There are reports on impacts on wildlife reproduction and development observed in invertebrates, fish, reptiles, birds and mammals, sometimes confirmed by laboratory tests. In laboratory experiments the impacts to fish populations by EDCs affect reproductive health and persistence of various fish species (Mills and Chichester 2005). Many examples of impacts due to exposure to endocrine disruptors exist in wildlife, such as the seals population decline in the Baltic and North Sea, the high levels of female egg yolk in male fish or snail imposex and intersex around the world. Intersexuality of fish has been demonstrated in several investigations carried out in rivers around the world. The findings of abnormal reproductive female-like ducts and oocytes in male fish were related to the treated sewage discharge from the cities in laboratory experiments measuring induction of plasma vitellogenin in exposed male fish (Jobling et al. 2002). Field studies were carried out using wild roach as a model fish to confirm the incidence and the severity of intersex which correlated with the predicted concentrations of the natural estrogens (E1 and E2) and the synthetic contraceptive pill estrogen (EE2) present (Jobling et al. 2006).

Some case-studies have made clear that the estrogenic activity of municipal wastewater correlates to demographics. The number of inhabitants was found to correlate with changes in estrogenic activities in a research conducted at a university city in US, with seasonal fluctuations in population. The concentrations of synthetic and natural estrogens and other EDCs were measured and effects demonstrated through the application of in vivo and in vitro tests (fish exposure with Vtg induction measurement and the yeast estrogen screen) (Brooks et al. 2003).

The demonstration of effects of pulp mill effluents has also been supported by fish surveys with a sampling design that includes upstream and downstream sites from the discharge pipe of the pulp mill. For instance, Munkittrick et al. (1994) have demonstrated that the absence of chlorine bleaching or the presence of secondary treatment does not eliminate estrogenic responses evidenced by decreased circulating levels of sex steroids, decreased gonadal size, which implies that there may be multiple causative agents. In other cases, androgenic effects have been noticed, such as a biased male to female ratio in fish in Sweden downstream from pulp mills (Larsson and Förlin 2002).

As seen on Table 3, several adverse endocrine effects are evidenced in various animals, from mollusks to amphibian but they also appear in higher animal species.

Table 3. Effects of EDCs in wildlife evidenced through field studies




Postulated mechanism or causative agent



High incidence of deformed frogs in Minnesota, United States

Multiple EDCs

Retinoid signaling

pathways activation

Gardiner et al. 2003

Marine Gastropods

Masculinization of female snails (imposex) occurs worldwide. Females grow accessory sex organs including sperm ducts, seminal vesicles,

external sperm grooves, and penises.

Exposure to low levels of tributyltin (TBT) (1ng/l)

Aromatase inhibition, testosterone inhibition, or neuroendocrine disorder or interaction with retinoid receptors

Novák et al. 2008

Wild roach (Rutilius rutilus)

Intersex, and high plasma Vtg concentration

Multiple EDCs

Sewage effluent from wastewater treatment plant discharging into rivers

Joblin et al. 2006


(Gambusia affinis)

Masculinization (90% affected in number of segments in the longest anal fin ray).

Androgen-dependent gene expression by luciferase test

Kraft pulp mill effluent

Affinity for human androgen receptor (hAR)

Parks et al. 2001

Eastern Mosquitofish, (Gambusia holbrooki)

Androgenic activity measured by androgen receptor transcription assay with human receptor in sediment. Fish masculinization.

Paper mill effluent, river

Pine pulp-derived phytosteroids accumulate in river sediment where they are converted by microbes into progesterone and this into androstenedione and other bioactive steroids

Jenkins et al. 2003

3.3. Conceptual model

Deriving a conceptual model requires knowing the pathways and toxicokinetics of the EDCs identified in the hazard identification step. An effects-based assessment start by identifying the effects and the relevant stressors and geographically located (for example through the use of GIS software). On the other hand, the model used in stressor-based assessments, depicts how stressors affect receptors and it is commonly applied when evaluating risks of environmental pollution. If a river basin is evaluated, the sources of contamination are studied, identifying the pathways, receptors and effects. To develop the human health risk assessment component, the fish consumption of the population and the drinking water intake are two of the main factors to consider especially for the most vulnerable population, which are newborn and lactating infants. The food chain is the main source of exposure, and in particular, fish consumption and drinking water are possible sources for the nursing mother and the pathway of distribution through the milk to the baby, but the direct intake of drinking water is important in the case of formula preparation. The environmental risk assessment should consider fish, crustacean and sediment dwelling organisms within the framework.

During pregnancy maternal fat is moved, releasing to the blood the bioaccumulated compounds, due to their liposolubility and persistence, through all the different exposure routes (foodstuffs, environmental, occupational) throughout her life. Acute exposure should also be considered if it happened previously to gestation or during this period. There are substances that traspass the placental barrier and chemicals reach the offspring. Also, through the breastmilk, explaining the extrangely high levels of some xenobiotics (Fernández et al. 2007).

3.4. Methodologies to determine dose-response in exposure assessment

The analysis of exposure and effect determines the concentration of the EDC on the environment matrixes matrices (water courses, ground water, drinking water, soil, sediment, air, biota), and assesses the potential or actual effects. In order to do so, many tools are recommended and in general a tiered approach is the most suited for this task as it helps to work in a logical order and increasing the specificity of the tests.

One of the main sources of exposure to most chemicals is through the food chain. The bioconcentration of organics in beef, cow milk and vegetation correlates to the octanol-water partition coefficient (Kow) to predict the bioaccumulation in the aquatic and terrestrial food chains (Travis and Arms 1988). There are many models based on the characteristics of the chemicals, such as the fugacity model, which allows to predict the expected concentrations in six environmental compartments (water, air, soil, bottom and suspended sediment and fish) (MacKay et al. 1985).

3.4.1. The use of a tiered methodology to demonstrate endocrine disruptive effects

This type of approach is carried out including different tests, such as bioassays, in vitro tests and field studies as part of the experimental design. The methodologies generally employed are in vivo fish reproduction exposure tests and in vitro receptor binding bioassays, for androgens and estrogens (Wartman et al., 2009). Even though there is an international trend towards diminishing the use of live organisms for experimentation for safety testing, in vivo tests are still of key importance for the confirmation of the findings of in vitro screens. Some of the most utilized tests relay on the use of fish as model experimental organism in various life-stages, as for example the 21 days reproduction fish test with fathead minnow (EPA/600/R-01/067).

3.4.2. In vitro screens and tests

Some of the in vitro assays that can be used as screening tools of estrogenic activity are the following: yeast based assays, cell proliferation assays, binding assays, tranfection assays and stably transfected cell lines. In general they are considered good and convenient screening tools because of their rapid and fairly easy methodology and they may reduce the number of animals needed for exposure testing. The ones utilizing ER stable cell line using T47D human breast cancer cells can be used for screening chemicals for estrogenic and antiestrogenic activities. Several environmental estrogens were tested using this method (T47D-KBluc cells assay), and results were a lowest observed effect concentrations (LOEC) for genistein and 4-NP of 10 nM and 0.5 nM, respectively (Wilson et al. 2004).

Table 4 summarizes the most frequently used in vitro screens and tests to demonstrate estrogenic effects.

Table 4. In vitro screens and assays for estrogenicity testing

Name of method




Yeast estrogen screen (YES)

Yeast strain tranfected with human estrogen receptor DNA sequences and a lacZ reporter gene (plasmid) encoding β-galactosidase enzyme.

Binding of EDC to estrogen receptor triggers a colorimetric reaction with β-galactosidase, and the yellow substrate CPRG turns into a red dye, which is measured at 540 nm

Arnold et al. 1996; Routledge and Sumpter 1996

Transactivation assay ERCALUX®

Human breast adenocarcinoma cells (T47D)

Exposure of cells marked with luciferase. After adding luciferin substrate, reaction with EDC is measured as luciferase activity with luminometer and correlated to calibration curve with E2.

Boever et al, 2001


MCF-7 cell culture

MCF-7 cells stimulate estrogen receptor-dependent transcription and growth promotion of estrogen-dependent cells in culture

Soto and Sonnenschein, 1985

Competitive ER binding assay

Cell lines from uteri of ovariectomized Sprague-Dawley rats

Measurement of the radioactivity to determine equilibrium binding of 3H-17 β-estradiol at various concentrations to EDC in rat cytosolic recombinant ER

Blair et al. 2000

The YES screen uses a genetically engineered strain tranfected with human estrogen receptor DNA sequences. After interacting with the estrogenic substance there is a change in the conformation of the receptor. The dimmers of the estrogen receptor are then located upstream of the lacZ reporter gene (encoding the enzyme b-galactosidase) present on a reporter plasmid. Incubation of these recombinant yeasts with EDCs triggers the expression of β-galactosidase, which causes expression of the reporter gene and this enzyme is secreted into the medium, metabolizes the chromogenic substrate, chlorophenol red-b-D-galactopyranoside (CPRG), initially yellow, turning it into a red dye that can be measured by absorbance at 540 nm (Arnold et al. 1996; Routledge and Sumpter 1996).

Transactivation assays have also been generated that make use of endogenous receptors, such as E-Screen and ERCALUX® assays, using MCF7 and T47D cells, respectively (Boever et al. 2001). The general procedure consists on the exposure of T47D human breast adenocarcinoma cells to the samples and the activity is calculated based on 17β-estradiol (E2) calibration curve. The response is produced after adding luciferin substrate was added to each multiplate well and luciferase activity measured with a luminometer. Estrogenic potency is expressed as estradiol equivalency (EEQ). The application to an investigation of endocrine disruptive activity in the rivers Meuse and Rhine in the Netherlands with fish exposure tests using fathead minnows measured estrogenic responses with this in vitro test. The reference toxicant was 17α-ethinylestradiol (EE2) (Bogers et al. 2007).

Bioactive concentrations in human beings cannot be easily extrapolated from animal studies, but in the case of endocrine disruptors it is recommended to base on the precautionary principle employing data of toxicity in mammals. Screens done using mammals cells (of rat or mice) rate the relative competitive ER binding activities of several EDCs taking DES as 100 %. As an example, the (ER) competitive-binding assay in the case of nonylphenol, using human cell lines shows that the compound has a bioactivity of -1.3 (LogRBA) (Kuiper et al. 1998) and -1.53 using rat cell lines from uteri of ovariectomized Sprague-Dawley rats (Blair et al. 2000). Table 5 shows some examples of relative ER binding activities found in databases for several EDCs that can be used for this purpose of extrapolating toxicity data to human beings. Relative ER binding activities in rat cells considers DES as 100%.

Table 5. Bioactivity data of selected EDCs


Relative ER binding activity in rat

17β-estradiol (E2)






ethynyl estradiol












diethylstilbestrol (DES)


Source: Tong et al (2008) FDA National Center for Toxicological Research Estrogen Receptor Binding Database (NCTRER) (http://pubchem.ncbi.nlm.nih.gov/)

3.4.3. Evaluation of effects with -omics technologies

The risk assessment of mixture effects is aided by the knowledge of mechanisms to diminish uncertainty and help to know if studies results can be extrapolated among species. Some of the technologies used are the following:

i) Genomics/transcriptomics

DNA array technologies allow thousands of genes to be surveyed in parallel, both for expression monitoring under various physiological conditions and in polymorphism analysis (Oberemm et al. 2005). An example of the use of this methodology is a study conducted by Vizziano et al. (2008) using microarray analysis working with rainbow trout, Oncorhynchus mykiss. Fish were exposed to a natural nonaromatizable fish androgen (11β-hydroxyandrostenedione) and to an aromatase inhibitor (1,4,6-androstatriene-3,17-dione), in order to elucidate the steroid-induced masculinization mechanism.

ii) Proteomics, metabolomics and toxicogenetics

These methods serve to study respectively the proteins, metabolites and genetic variability of organisms exposed to toxicants. Other laboratory experiments with fish, such as those employed in the research by Ankley et al. (2009) incorporate to the classical Vtg induction test, gene end-points measured via quantitative real-time polymerase chain reaction (PCR). Predictions of estrogenicity based only on chemical analysis are not accurate, because they do not account for mixture effects and molecular interactions. The measurement of Vtg is done using an ELISA immunochemical method. This technique can be applied to model fish (for example zebrafish, medaka, fathead minnow), for which there is a publically available sequences and primers of the vtg gene can be easily obtained. Some ecologically relevant species of non model fish can be used to evaluate the aquatic ecosystem exposed to EDCs applying a quantitative real-time polymerase chain reaction method with a vtg gene amplification (Biales et al. 2007).

3.4.5. The use of biomarkers in biomonitoring and in vivo bioassays and biomonitoring to assess exposure to EDCs

The use of biomarkers in among one of the most used tools to assess endocrine disruption. The IUPAC defined biomarker as an “indicator signaling an event or condition in a biological system or sample and giving a measure of exposure, effect, or susceptibility.”

They have also been defined as “functional measures of exposure to stressors expressed at the sub-organismal, physiological or behavioral level”.

a) Vitellogenin

The use of biomarkers such as vitellogenin (Vtg) is one of these tools. The measurement can be done using blood plasma or liver extracts and it is based on the sandwich enzyme-linked immunosorbent assay, specific for each fish species selected. Vtg is a female phospholipoprotein synthetized in the liver precursor of egg yolk protein, non-detectable in normal male fish, which production can be induced when exposed to xenoestrogens (IUPAC 2003). It is used as an end-point in field biomonitoring and in laboratory experiments (fish exposure tests). Examples of the first are several Canadian studies finding evidence of higher than normal VTG, decreased gonad size and circulating steroids in fish living downstream from pulp mills and municipal sewage treatment plants (Mac Master et al. 2006). In order to evaluate the efficacy of treatment for reclaimed water from sewage this biomarker was used in China with experiments developed for the fish crucian carp (Carassius carassius) (An et al. 2007).

As an example, in a laboratory experiment, adult fathead minnow (Pimephales promelas) fish were exposed in a classical 21 days reproduction assay to chemicals known to affect the hypothalamic-pituitary-gonadal (HPG) axis. Plasma concentrations of Vtg were measured, among other end-points (Ankley et al. 2009).

The use of biomarkers such as vitellogenin (Vtg) is still one of the most recognized around the world. It can be run with blood plasma or with liver and it is based on the sandwich enzyme-linked immunosorbent assay, specific for each fish species selected. Vtg is a female phospholipoprotein synthetized in the liver precursor of egg yolk protein, non-detectable in normal male fish, which production can be induced when exposed to xenoestrogens (IUPAC 2003). It is used as an end-point in field biomonitoring and in laboratory experiments. Examples of the first are several studies carried out in Canada which found evidence of higher than normal Vtg, decreased gonad size and circulating steroids in fish living downstream from pulp mills and municipal sewage treatment plants (Mac Master et al. 2006).

Fish reproduction tests are done in the laboratory generally based on USEPA or OECD protocols or modifications of these procedures, using several model fish, such as Japanese medaka, fathead minnow or zebrafish. An example of the use of this method is the research conducted to assess the estrogenic effects of the chemical 4-tert-octylphenol. The model fish was Japanese medaka (Oryzias latipes) with end-point Vtg induction, aided by histological studies. Some of the findings were an inhibition of spermatogenesis at concentrations higher than 41 mg/l. The appearance of oocytes in the testes and germ cell hyperplasia throughout the testes were present in two of the exposed fish at higher concentrations (74 and 230 mg/l, respectively) (Gronen et al. 1999).

In order to evaluate the efficacy of treatment for reclaimed water from sewage this biomarker was used in China with experiments developed for the fish crucian carp (Carassius carassius) (An L et al. 2007). Another research evaluates binary mixtures measuring Vtg induction with Japanese medaka (Oryzias latipes) fish (Sun et al. 2009).

This method has been used to assess the effects of mixtures by a fathead minnow fish exposure test to prove the concentration addition principle, meaning that the joint risk of the components can be significant even though each one could be at a concentration below the active threshold (Brian et al. 2007).

b) CYP1A

Another useful biomarker is based on the induction of CYP1A activity. This is an enzyme composing the cytochrome P450 system, which participate in the synthesis of lipids and steroids. The CYP1A is measured in fish by several methods. One of the most common is the catalytic activity of ethoxyresorufin-O-deethylase (EROD).The binding to the arylhidrocarbon receptor (AhR) induces transcription of CYP1A, among other target genes. It can be used to assess exposure to aromatic hydrocarbons such as PCBs, dioxins and furans (Goksøyr 2006).

It is also a good biomarker of carcinogenicity, not only in fish but in humans. The extrapolation of results from animal studies of P450 induction to humans is a complex process that requires, ultimately, the direct proof from human studies (Ma and Lu 2007).

c) Retinoids

Retinoids are molecules coming from dietary sources of vitamin A which act as signaling and participate in many biological processes, including some related to the reproductive function. Lower than normal hepatic levels of retinol may be caused by pollutants coming from pulp mill effluents. However, estradiol causes an increase of plasma retinol levels.

Carotenoids are precursors of vitamin A important in the feeding of embryo. Reduced ovary pigmentation due to carotenoid depletion and diminished biosynthetic capacity of sex steroids was observed in the fish Gobiomorphus cotidianus in a river in New Zealand (Landman et al. 2008). This effect has been seen in other animal species, like amphibians. Frog limb deformities are found to be linked to the retinoid system disruption (Novák et al. 2008).

3.4.4. Human epidemiology tools to detect endocrine disruption

Environmental epidemiology data can be very valuable to establish the health effects of a community exposed to environmental agents (Paddle and Harrington 2000). There are epidemiological studies to assess the effects of pesticides and phthalates expressed on abnormal anogenital distance in newborn to two year old boys and girls. This measurement which is performed after placing the infant on the dorsal decubitus position with both hips flexed and the distance from the perineum to the penis is done using a caliper. It has been found that this measurement can be used as a biomarker of androgen and anti-androgen effects in population studies as it was used before in rodent reproductive toxicological studies (Thankamony et al. 2009). Cryptorchidism could be produced by an in utero exposure to contraceptive pills. However, Martin et al. (2008) after conducting a meta-analysis of 50 epidemiological studies could not find evidence of a causative estrogenic mode of action.

4. Risk characterization

This final step of the risk assessment process integrates data of the dose-response relationship of an agent with estimates of the degree of exposure in a population to characterize the likelihood and severity of health risk. It serves as an interface between risk assessment and risk management (Williams and Paustenbach, 2002). This step has been defined by IPCS as: “The qualitative and, wherever possible, quantitative determination, including attendant uncertainties, of the probability of occurrence of known and potential adverse effects of an agent in a given organism, system, or (sub)population, under defined exposure conditions (IPCS 2004).

4.1. Uncertainty and confounding factors

Uncertainty refers to a lack of enough knowledge such as those arising from the analytical measurement, sampling and models and also from data gaps in the assessment, such as poor information on human exposures or the toxicity of a chemical (Williams and Paustenbach, 2002).

4.1.1. Toxicokinetics and special dose-response patterns of EDCs

Not every EDC follows the same dose-response curve, as monotonic and non-monotonic functions can be seen. The NOEL (no-observed-effect level (NOEL) is defined by IUPAC as the “greatest concentration or amount of a substance, found by experiment or observation, that causes no alterations of morphology, functional capacity, growth, development, or life span of target organisms distinguishable from those observed in normal (control) organisms of the same species and strain under the same defined conditions of exposure”. As hormesis is an effect that appears linked to endocrine disruption in many research papers, a special reference to this topic is developed below.

4.1.2. Hormesis and its connection to endocrine disruption

Hormesis has been defined as a biphasic dose-response phenomenon characterized by a low-dose stimulation and a high-dose inhibition (Calabrese and Baldwin 2002, Calabrese 2008). The hypothesis that a paradoxal toxicity effect may appear at low-doses, showing an inverted U-shaped dose-response curve which is sometimes referred to as non-monotonic response, is still controversial according to some authors due to the uncertainties of the experimental design and of the testing methods (Mushak 2007). In 2, the pattern of J-shaped and inverted-U curves are shown for illustrative purposes.

2. The J-shaped curve shows the response as a biological dysfunction, while the inverted-U curve depicts the response as a normal biological function, such as growth (re-drawn from Hoffman 2009)

Even though the phenomenon of hormesis has sometimes been linked to growth stimulation of the organism or other apparent beneficial effects, in general other toxicity mechanisms different to those of higher concentrations appear. This may be connected to reproductive effects. Several chronic toxicity tests carried out below the NOEL threshold level with various aquatic species have shown an increased number of offspring.

Several hypotheses on the genesis of hormesis consider the influence of nutrients, or detoxification processes that make the response lower at higher concentrations. Hormesis may also be linked to chemical interactions, such as synergy or potentiation and reported specially in chronic experimental studies with many species, from yeasts to mammals. Endocrine disruptors could exert stimuli at low-doses and may induce an inverted U-shaped dose response. According to some researchers (Calabrese 2008) these estrogenic effects are clear examples of hormesis. Researchers of the International Dose-Responses Society have drawn the hypothesis that the mechanism of homodimerization of steroid hormone receptors is responsible for the non-monotonic dose responses observed with certain EDCs (Zhang et al. 2008). According to this theory, a possible explanation to the U-shaped dose-response curve relies on the inherently non-linear binding process of the steroid hormone receptors to the ligand from initially monomeric, to dimmers, built by homodimerization, which finally produce the gene expression of the hormone (Li et al. 2007).

The combined effect of chemicals is a common matter in nature can give rise to hormetic or non-hormetic effects, depending on the complexity of the interaction of the chemicals and the concentrations and characteristics of the organisms. Agreeing with the point of view of Kefford, who recommends considering the stimulatory effects when evaluating ecotoxicity tests at the individual level, there still seems not to be a convincing way to interpret the meaning of this effect at the community levels (Kefford et al. 2007).

Hormesis has been extensively seen in bioassays using reference invertebrate, such as cladocer crustaceans as shown in the following table:

Table 6.Hormetic effects in Cladocer crustacea



Endocrine disruptor or type of effluent

Experimental range/hormetic concentration

Hormetic effect


Daphnia carinata

21 days three generation test


0.005 to 0.500 μg l-1

Shortened time to the first brood in 2nd generation

Zalizniak et al., 2006

Daphnia magna

30-day exposure


1 to 100 μg l-1

36 μg/l

Fecundity increase

Flaherty et al., 2005

Daphnia magna

25-day LC50


0.5, 1.0 and 5.0 μg l-1

Increase in number of neonates, decrease size

Bodar et al., 1988

Daphnia magna

Chronic toxicity tolerance test (4 days pre-exposure, 4 days depurating, 28 days exposure) 2 brood


Pre-exposure to 2.5 and 25 nM

1.5 to 15 nM

Higher final wet weights and reproductive rates than the control groups

Tsui et al., 2005

Ceriodaphnia dubia

7 days chronic toxicity test

Pulp mill effluent

At concentration < 40%

Reproduction stimulation

Middaugh et al. 1997

It is well known that some metals act as micronutrients in enzymatic, hormone systems and metalloprotein, wich are essential for life (zinc, copper, iron, selenium). It is not altogether unreasonable to suppose that some of the stimulatory effects denoted are not more than the growth curve to an optimal dose for the micronutrient interaction with the active protein, thus causing growth over the normal values. Growth hormesis is thought to be linked to biological adaptive responses that occur at low level doses of toxicants, beyond the saturation of the counter-inhibition capacity. This theory was anticipated by Herbert Spencer as early as in 1862 who described the antagonistic forces that enact to bring back equilibrium when disturbing forces work an excess of change in some direction.

Sex determination is the genetic or environmental process by which the sex of an individual is established in a simple binary fate decision whereas sex differentiation is the process by which an undifferentiated gonad is transformed into an ovary or a testis. These authors also mention that the normal sex ratio tends to be 1:1 (50% males and 50% females) and that the gonads of all fish initially develop as ovaries, then a proportion of fish become male (Penman and Piferrer 2008). Fish are one of the species that have been studied in more detail with relation to the influence of environmental abiotic conditions such as temperature and sex determination and differentiation.

4.1.3. Influence of temperature on fish sex differentiation and on ecotoxicity

A rise in temperature may affect fish if produced during the most sensitive period of early development, producing a higher incidence of males than normal or activating the brain-pituitary-gonadal axis to start the reproductive cycle, increasing estradiol secretion and altering the expression of aromatase (cyp19a), through a yet not fully elucidated molecular mechanism (Penman and Piferrer 1998).

This effect has been observed in several laboratory experiments, as for example the one with the reference fish Medaka, Oryzias latipe which demonstrates its sensitivity to the effects of high temperature (Selim et al. 2009).

The environmental changes may also affect the sensitivity of organisms to chemicals, and toxicity in general increases with rising temperature (Heugens et al. 2002).

4.2. Biosynthesis and mechanism of action of estrogenic EDCs

In humans, the enzyme aromatase is an enzyme that converts testosterone into estradiol and it is present in the ovaries of premenopausal women, in the placenta of pregnant women, and in the peripheral adipose tissues of women and of men (USEPA 2007). Estradiol is one of the feminine steroidal hormones. As a reference, the normal estradiol blood concentration, in ovulating women is 400 pg/ml (Medline Plus, Medical Encyclopedia, NIH).

One of the biosynthetical pathways of estradiol is via methyl group oxidation of testosterone mediated by the enzyme aromatase to form a benzene ring, as depicted on 3:

In fish, induction of aromatase in the ovary also converts testosterone to estrogen, stimulating the liver to secrete vitellogenin, which is then transported in the bloodstream to the oocytes. The female secondary sex characteristics and negative feedback linkages to the hypothalamus and hypophysis are also induced by aromatase (Cyp19a1a). It also participates in testicular differentiation in both gonochoristic and hermaphrodite fish species. The masculinization of female fish may be mediated by the inhibition of Cyp19a1a enzymatic activity (Guiguen et al. 2009).

There are several theories about how endocrine-disrupting compounds (EDCs) interact with the receptors to trigger the responses on the endocrine system causing either agonism or antagonism. In the case of the estrogen receptors, the receptor bound to the membrane has an influence on the proteins, while the cytosolic one binds to the genes of the chromosomes in the nucleus after binding with estrogen and provokes RNA transcription. Some of the macro characteristics of the mechanisms of EDCs are that the most vulnerable period is considered to be the fetal period to the early postnatal developmental period when organs continue to undergo substantial development. The effect in general shows a latency before manifesting and it is likely for the organism being exposed organism to a mixture of EDCs. As seen later on under the hormesis section (4.1.3), another characteristic is the low dose effect that means could be more potent than at higher doses and the nontraditional dose-response curves. Effects may be transmitted to subsequent generations through modifications of gene expression such as DNA methylation and histone acetylation.

There are two estrogen receptors, ERα and ERβ, structurally belonging to the estrogen/thyroid hormone superfamily of nuclear receptors. ERβ is abundant in male urinary tract. Their functional domains are three: one with an amine group (A/B), the second with DNA binding groups (called C) and the D/E/F with ligand binding domains.

After binding to the receptor, conformational changes occur and the gene transcription includes processes that are still not fully understood but produce the receptor dimerization, its interaction with DNA, the participation of coactivators and the formation of a preinitiation complex (Nilson et al. 2001).

5. Risk estimation

The integration of the results of the assessment is done at this stage, linking exposure to EDCs with the effects and to develop an estimation of the risk and probability of adverse effects that could be useful for risk managers. When evaluating populations, it is necessary to use epidemiology methods to estimate the exposure to multiple stressors and incorporate non-chemical stressors into cumulative risk assessments. An example of such approach is found in the article by Zartarian and Schultz (2009) which guides to the new tools developed by the USEPA such as HEDS (Human Exposure Database System) for dose modeling of aggregate and cumulative risks related to human exposure (U.S. EPA 2009a), and probabilistic models used to calculate percentiles and their uncertainties for a target population applying tools like Monte Carlo simulations to ponder population exposure (U.S. EPA 2009b). An estimation of lognormal distributions of the tap water intake of children and adults to the overall population was done by this method, which is useful for public health assessments (Roseberry and Burmaster 1991).

Risk analysis based on evidence has contributed to policy decisions, because stakeholders expect to be convinced on the supporting science upon which decisions are based. Weight of evidence (WOE) approaches developed in the fields of medical evidence, forensic science and radioactive waste management can be applied to environmental risk (Pollard et al. 2008). The weight-of-evidence (WOE) approaches are methodologies that can be used to perform integration of evidence of a toxic effect. Professional judgement is a qualitative WOE approach, whilst quantitative methods, such as direct scoring can be used to draw out conclusions from multiple lines of evidence (Linkov 2009).

5.1. Computational tools

The reference dose (RfD) for cancer and non-cancer effects is studied through the derivation of LOAEL and NOEL. As defined on the Integrated Risk Information System (IRIS) glossary, USEPA (http://www.epa.gov/IRIS/help_gloss.htm), these are defined as:

No-Observed-Adverse-Effect Level (NOAEL): An highest exposure level at which there are no statistically or biologically significant increases in the frequency or severity of adverse effect between the exposed population and its appropriate control; some effects may be produced at this level, but they are not considered adverse, nor precursors to adverse effects.

Lowest-Observed-Adverse-Effect Level (LOAEL): The lowest exposure level at which there are statistically or biologically significant increases in frequency or severity of adverse effects between the exposed population and its appropriate control group; also referred to as lowest-effect level (LEL).

Table 7.Chronic toxicity data for human health risk assessment by drinking water and developmental and cancer end-points

Chemical Name

Oral RfD (mg/ kg bw-day)

(cancer and developmental end-points)


Unit Risk (µg/l)




Bisphenol A





Polychlorinated biphenyls (PCBs)









Source: NCBI database

NOAEL: No Observed Adverse Effect Level; NOEL: No Obsevered Effect Level; LOAEL: Lowest Observed Adverse Effect Level

An example of the use of these data is the estimation of the human exposure to bisphenol A. An uncertainty factor of 500 on the NOAEL was chosen (10 for interspecies differences, 10 for inter-individual differences and 5 for the uncertainties of the database) therefore, the overall NOAEL of 5 mg/kg bw/day gives a temporary TDI of 0.01 mg/kg bw. Realistic worst-case estimates of consumer exposure via foodstuffs, ranging from 0.00048 mg/kg bw/day for adults to 0.0016 mg/kg bw/day for infants, are below this t-TDI of 0.01 mg/kg bw (SCF/CS/PM/3936 European Commision Health & Consumer Protection Directorate, 2002).

Estimates of exposure concentration posing minimal risk to humans (Minimal Risk Levels, MRLs) include adjustments to reflect human variability and extrapolation of data from laboratory animals to humans. Table 7 shows some examples of these estimations for some EDCs.

Table 8. Minimal Risk Levels (MRLs)


MRL (mg/kg/day)














Source: Agency for Toxic Substances and Disease Registry (ATSDR) December 2008

The toxicological profile for chlorinated dibenzo-p-dioxins shows that the most toxic congener of dioxins is 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Its carcinogenicity has been recognized but after a reassessment it has been suggested that the noncancer hazard may be even more important. The US EPA, the World Health Organization (WHO) and U.K. Food Standards Agency (FAO) committee (JECFA) on dioxins have reached similar conclusions. A no-observed-adverse-effects level (NOAEL) of 13 ng kg-1 (maternal body burden) was identified as the most pertinent for deriving a reference dose (RfD) for humans. RfD has been establish between 1 and 10 pg kg-d-1 (TCDD TEQ) as a safety criterion (Greene et al. 2003).

5.2. Models and statistical tools

In order to estimate exposure with greater certainty it is necessary to use reliable and sensitive analytical methods able to detect traces of chemicals. Even when the methods are improving in this regard, sometimes data gaps in monitoring arise, so, several mathematical models may be used to estimate exposure. Field measurements are used to confirm model results. The statistical distributions of environmental data are often lognormal, implying that the geometric average is a better estimate for the central tendency of data than the arithmetic mean. However, the latter one is generally used to identify environmental concentrations for use in exposure assessment. The analytical limit of detection (LOD) are either taken as zero or as a fraction of the value, typically one-half, impacting on the risk assessment and the decisions taken after it (Paustenbach 2000).

Probabilistic and possibilistic logic approaches are found on the literature as statistical models for risk estimation. For example, fuzzy logic possibilistic approach was used in a South African study to assess the risk of co-occurring stressors in an aquatic ecosystem. The stressor-response is expressed as a conditional possibility and the possibility distribution of the stressors yield an estimate of the ecological risk. Stochastic events are dealt with by the probability theory, but it is not always easy to estimate the frequencies to calculate the probability distributions. The Bayesian approach is considered to be better suited, but the boundaries of ecosystems are vague and difficult its application (Jooste 2001).

6. Integrative risk assessment of endocrine disruptor mixtures

Authors like Suter et al (2003) consider that integration is possible in all risk assessment steps. Within the planning step, the assessment questions and end-points can derive unified conceptual models, including sources, pathways and effects considering humans as a potential receptor. Exposure analysis may integrate the sources, most of the pathways, fate, transport, bioaccumulation and biodegradation of substances. The hazard analysis stage includes common reported effects and mechanisms and the use of biomarkers and indicators. The observed and estimated direct and indirect effects are extrapolated including uncertainty factors and models. At the risk characterization step, an evaluation of the combined effects is done, through an exposure-response model. The increasing knowledge of biological systems, allow using an integrated approach based on common mechanisms (Vermeer et al. 2007).

The degree of integration of a risk assessment is chosen according to the case, and even though human health risk assessment has evolved separated from ecological health evaluations, but the group of endocrine disruptors is fit for integration as they may potentially affect the environment and humans. The criteria to meet is that bioaccumulation of the agent and/or it metabolites is expected in species used for human use. Integrated risk assessment can help understand impacts on human health and on ecosystems due to the extrapolation from animal experiments after defining mechanisms of action. So, in the risk characterization phase the differences among mechanisms of action between species should be considered as a basis for inter-species extrapolation. (Bridges and Bridges 2004). Finding out the linkage among environmental exposures and human health using biomarkers as quantitative estimates of human exposures to chemicals is mentioned under the USEPA document on this subject (U.S. EPA 2009c). As shown by several researchers, endocrine disruptors are good example of this, as endocrine activity mediated by steroid hormone receptors is difficult to predict based solely on the actions of the individual chemicals found in a mixture.

It is a recommended practice taking into account confounding factors such as interactions among compounds specially when dealing with the risk assessment of the exposure of humans or wildlife to complex mixtures.

6.1. Confounding factors arising from antagonistic responses and interactions

Such is the case of the evaluation of dioxins toxicity that affects human health as some dietary factors influence the risk characterization. Flavonoids coming from diet sources may possess AhR-antagonistic action (Safe et al. 2002,) Moreover, as shown in a research by Satoh et al. (2001) some EDCs can bind not only to ER but to the androgen receptor (ADR). In these cases, the question is whether the expressed action will finally be estrogenic at the organism or population level. In this same experiment, one particular EDCs showed both estrogenic and androgenic activities when the ADR and ER Competitive Binding Assay was applied. For instance, 4-nonylphenol has an ADR binding IC50 of 1.3 × 10-5 M, where IC50 in this case is defined as the concentration of chemical required to reduce the specific testosterone binding by 50%. The same compound has an ER binding IC50 of 7.2 × 10-7 M, defined as the ratio of diethylstilbestrol (DES) to that of the EDC.

Studies conducted by Gaido et al. (2003) show that chemicals in a mixture of EDCs may interact in additive or in antagonistic manner and recommend to develop mechanisms-based hazard risk assessment of receptor agonists/antagonists to account for these multiple variables. To make this question even more intricate, as found on the latest resource document of USEPA on cumulative risk assessment (U.S. EPA 2007a), metabolites can differ in action from the parent compound, as the example of a PCB congener (3,4',5-trichlorobiphenyl) with antiestrogenic in vitro activity, while its hydroxylated metabolite is strongly estrogenic. Therefore, this has to be taken into account when grouping the substances according to their expected joint action, which makes laboratory experiments a very necessary tool. It is useful to elucidate mechanisms of endocrine disruption through molecular biology studies. In the case of estrogenicity, this action is defined by the affinity of a substance to bind to ER-α or ER-β receptors. Methods like the yeast estrogen screen and ER-α gene reporter system based on MCF7 cells are being used to try and find additivity of mixtures. The final effect is not easy to predict when antiestrogenic, estrogenic and androgenic substances coexist (Kortenkamp 2007).

6.2. Non-monotonic dose-response curves

Another consideration on the special case of endocrine disruptors is discussed by Beronius et al. (2009). Non-monotonic dose-response curves are frequently found with regards to EDCs toxicity. As argued by this author, no threshold can be assumed as endogenous hormones already exceed the physiological threshold, making extrapolation from high to low dose difficult.

6.3. How to assess the risk of endocrine disruption in mixtures

6.3.1. The phased approach

The phased approach described by Menzie et al. (2007) can be used to combine effects in multiple stressors. After developing a conceptual model including all relevant stressors and their interactions, a screening is done to focus on those presumed with highest effects. The evaluation of the effects of each individual one is followed by the consideration of combined effects, through qualitative and/or quantitative methods. This kind of procedure is also recommended by the US EPA in order to get the information for a hazard and risk assessment of multiple contaminants.

In an effects-based assessment, Bayesian networks are good choices for the graphical representation of the interactions among stressors. The stressors and their effects are then geographically located (for example through the use of GIS software). A mechanistic knowledge is useful to understand how stressors cause effects individually or combined.

In stressor-based assessments, the model depicts how stressors affect receptors and is commonly applied when evaluating the risks of environmental pollution. Screening is done when magnitudes of stressors are below benchmarks thresholds or overall target data. The additivity of doses is calculated through toxic equivalence or relative potency methods for chemicals considered toxicologically similar.

6.3.2. Toxic Equivalency Factor approach

According to Safe (1998), the TEF approach for assessing the hazard of toxic chemical mixtures combines the contribution of each component and their relative potencies and can be represented through the following equation:

TEQ = Sum (Ci) x TEFi


TEQ: toxic equivalent

Ci: concentration of each individual compound in the mixture

TEFi: relative potencies

The toxic equivalency factor (TEF) approach is useful for persistent halogenated aromatics AhR agonists.

6.3.3. Selective receptor modulator approach

As mentioned before, mixture components can be additive for one response or antagonist for another one. For EDCs mixtures structurally diverse, with possess agonistic action in some tissues and antagonist in others, the “selective receptor modulator” (SERM) approach can be applied. The binding of the ligand to estrogen changes the conformation of the receptor and recruits nuclear factors required for gene expression. The antiestrogenic drug tamoxifen is an example of a selective estrogen receptor (Gaido et al. 2003).

6.3.4. Factorial analysis

An example of this approach is the combination of the effects in mixtures of estrogenic chemicals employing the Japanese medaka (Oryzias latipes) fish exposure test, with a test duration of 14 days. The binary mixtures of estrogenic chemicals were 17β-estradiol (E2), 4-tert-nonylphenol (NP) and bisphenol A (BPA). In this case, there were no significant interactions between E2 and NP, E2 and BPA, and NP and BPA suggesting that these three model estrogenic chemicals share a common mechanism for inducing Vtg synthesis, and that no interactions occur when they act in combination. A factorial design, which allows the effects of individual chemicals and their interactions to be statistically tested, was employed using 27 combinations crossed for the three compounds. The dose levels were chosen based on the concentration-responses of the individual chemicals determined in study (Sun et al. 2009).

7. Synthesis and discussion

The traditional framework for risk assessment can be applied but there are some particularities that have been developed for the case of endocrine disruptors. In the Guidelines for Reproductive Toxicity Risk Assessment of the US Environmental Protection Agency, EPA/630/R-96/009, there is a distinction to the traditional framework as the hazard identification step is combined with the dose-response relationships, route, timing, and duration of exposure and it is denominated hazard characterization step. All available experimental animal and human data, including observed effects, associated doses, routes, timing, and duration of exposure, are examined to determine if an agent causes reproductive toxicity in that species and, if so, under what conditions. The same approach is followed in the Guideline for Developmental Toxicity Risk Assessment, EPA/600/FR-91/001.

As seen in this review, the methodology for human and environmental risk assessment of endocrine disruptors is still not fully harmonized. The factors that make uncertainty increase are confounding factors such as the particular toxicokinetics characteristics of EDCs, their different mechanisms of action and receptors, their multiple pathways and sources and ubiquity in nature. Some of the remarkable issues are the preferential vulnerability to these xenobiotic compounds at certain developmental stages and the fact that responses can be sometimes higher at low-doses. The mixture effects of EDCs are not easy to assess as there may be interactions, synergy, agonism or antagonism among them, depending on the case, but some tools have appeared in general relying on the use of a tiered approach by measurement of the substances concentrations whenever possible, screening the endocrine disruptive action and confirming the action using a suite of in vivo tests.

Some of the research gaps considered by Williams and Paustenbach (2002) relate to the somewhat limited availability of health risk estimates with a lack of specific hazard information for developmental, reproductive, and neurological effects. Agreeing with this point of view, in the authors´ opinion, there is a need for improving the methodologies used to assess the combined effects of molecules sometimes with antagonistic or synergistic activities, developing accurate and reliable tests and screens to characterize exposure and elucidating the mechanisms of action in the case of mixtures. How to predict actions in the case of complex mixtures is a question that remains still not full answered.

In our opinion, an effects-based assessment can be applied to the case of EDCs using field observations of developmental and reproductive anomalies in wildlife and epidemiology data of relevant effects on human health, as for example the rate of hypospadia in neonates. The evaluation of results of ad hoc experiments, designed to demonstrate the possible effects and to derive dose-response data of individual and combined stressors should pay attention to the existence of synergistic, additive or antagonistic effects and the probability of non monotonic dose-response curves. As seen before, stressor concentrations below its individual threshold can still combine with others giving rise to an observable effect.

In summary, a well designed experimental work of effect-directed bioassays and chemical analysis up-joined to thorough epidemiological studies in vulnerable sub-populations are the best ways to develop reliable risk assessments of endocrine disruptors for human and environmental health. We recommend the use of a tiered approach to efficiently determine the effects in wildlife consisting of in vitro screens and in vivo tests, including laboratory exposure tests and fish surveys. In order to assess human health effects the application of biochemical screens with human cell lines are suggested to be adopted in conjunction to epidemiology studies in the vulnerable sub-populations. The risk estimation is done after determining the effects and the dose through testing, based on the conceptual model showing the sources, stressors, pathways and receptors. Based on the conceptual model the measurement of the concentrations of the most probable EDCs in the matrixes of influence, such as drinking water, river water, sediment and fish, are used to enter into the databases and calculate the dose-response using models and computational tools. Finally, risk assessment should provide a categorization of the risk based on the severity and likelihood of the effects. A well constructed report should be done in order to effectively inform risk managers and decision makers concerning the risks of endocrine disruption for the considered case.


An L, Hu J, and Yang M (2008) "Evaluation of estrogenicity of sewage effluent and reclaimed water using vitellogenin as a biomarker" Environmental Toxicology and Chemistry, vol. 27, No. 1, pp. 154-158

An L, Hu J, Yang M (2008) “Evaluation of estrogenicity of sewage effluent and reclaimed water using vitellogenin as a biomarker” Environmental Toxicology and Chemistry, vol. 27, No. 1, pp. 154-158

Ankley GT, Miller DH, Jensen KM, Villeneuve DL, Martinovi´c D (2008) "Relationship of plasma sex steroid concentrations in female fathead minnows to reproductive success and population status"Aquatic Toxicology, vol. 88 pp. 69-74

Arnold SF, Robinson MK, Notides AC, Guillette LJ, Jr, and McLachlan JA "A Yeast Estrogen Screen for Examining the Relative Exposure of Cells to Natural and Xenoestrogens" Environmental Health Perspectives, vol. 104, No. 544-48

Assmuth T and Hildén M (2008) "The significance of information frameworks in integrated risk assessment and management", Environmental Science & Policy, vol. 11, pp.71-86

Belknap AM, Solomon KR, MacLatchy DL, Dubé MG, and Hewitt LM (2006) "Identification of compounds associated with testosterone depressions in fish exposed to bleached Kraft pulp and paper mill chemical recovery condensates" Environmental Toxicology and Chemistry, vol. 25, No. 9, pp. 2322-2333

Beronius, A., et al. (2009) "Health risk assessment procedures for endocrine disrupting compounds within different regulatory frameworks in the European Union" Regulatory Toxicological and Pharmacology doi:10.1016/j.yrtph.2009.05.019 Article in press

Blair RM, Fang H, Branham WS, Hass BS, Dial SL, Moland CL, Tong W, Shi L, Perkins R, and Sheehan DM (2000)"The estrogen receptor relative binding affinities of 188 natural xenochemicals: structural diversity of ligands" Toxicological Sciences, vol. 54, pp.138-153

Bodar CWM, Van Leeuwen CJ, Voogt PA, Zandee DI (1988) "Effect of cadmium on the reproduction strategy of Daphnia magna" Aquatic Toxicology, vol. 12, issue 4, pp. 301-310

Boever PD, Demaré W, Vanderperren E, Cooreman K, Bossier P, and Verstraete W (2001)"Optimization of a Yeast Estrogen Screen and Its applicability to study the release of estrogenic isoflavones from a soygerm powder" Environmental Health Perspectives, vol. 109, No. 7, pp. 691-697

Bogers R, De Vries-Buitenweg S, Geuijen I, van de Waart B, Kuiper R, Van Der Linden S, Puijker L, Murk A (2007) "An in vitro/in vivo screening assay as a sensitive tool to assess endocrine disruptive activity in surface water" Environment International, vol. 33, pp. 292-301

Bretveld R, Kik S, Hooiveld M, van Rooij I, Zielhuis G, Roeleveld N (2008) "Time-to-pregnancy among male greenhouse workers" Occupational and Environmental Medicine, vol. 65, No. 3, pp.185-90

Bretveld R, Kik S, Hooiveld M, van Rooij I, Zielhuis G, Roeleveld N.Dubé M and Munkittrick K (2001)"Integration of effects-based and stressor-based approaches into a holistic framework for cumulative effects assessment in aquatic ecosystems" Human and Ecological Risk Assessment, vol. 7, No. 2, pp. 247-258

Brian JV, Harris CA, Scholze M, Kortenkamp A, Booy P, Lamoree M, Pojana G, Jonkers N, Marcomini A, Sumpter JP (2007) "Evidence of estrogenic mixture effects on the reproductive performance of fish" Environmental Science and Technology, vol. 41, No.1, pp. 337-344

Bridges J W and Bridges O (2004) "Integrated risk assessment and endocrine disrupters" Toxicology, vol. 205, pp. 11-15

Briggs D (2008) "A framework for integrated environmental health impact assessment of systemic risks" Environmental Health, vol.7, issue 61

Brooks BW, Foran CM, Peterson BN, Weston J, La Point TW, Huggett DB (2003)"Linkages between population demographics and municipal effluent characteristics" Bulletin of Environmental Contamination and Toxicology, vol. 71, pp. 504-511

Calabrese EJ (2008) "Hormesis: why is it important to toxicology and toxicologists". Environmental Toxicology and Chemistry, vol. 27, No. 7, pp. 1451-1474

Calabrese, E.J and Baldwin L.A. 2002, "Defining hormesis" Human & Experimental Toxicology, vol. 21, pp. 91-97

Callahan M A and Sexton K (2007)"If cumulative risk assessment if the answer, what is the question? "Environmental Health Perspectives, vol.115, No. 5, pp. 799-806

Cameron E and Peloso GF (2005) "Risk management and the precautionary principle: a fuzzy logic model" Risk Analysis, vol. 25, No. 4

Christianson-Heiska IL, Haavisto T, Paranko J, Bergelin E, Isomaa B (2008) "Effects of the wood extractives dehydroabietic acid and betulinol on reproductive physiology of zebrafish (Danio rerio)-A two-generation study" Aquatic Toxicology, vol. 86 (3), pp. 388-396

Cong L, Qin Z-F, Jing, X-N, Yang L, Zhou J-M, Xu X-B (2006) "Xenopus laevis is a potential alternative model animal species to study reproductive toxicity of phytoestrogens"Aquatic Toxicology, vol. 77, No. 3, pp. 250-256

European Commission. "Communication from the commission to the Council, the European Parliament and the European Economic and Social Commitee" "A European Environment and Health Strategy"Executive summary. http://ec.europa.eu/environment/health/strategy.htm [accessed 14 January 2010]

Feijtel T, Kloepper-Sams P, den Haan K, van Egmond R, Comber M, Heusel R, Wierich P, Ten Berge W, Gard A, de Wolf W, and Niessen H (1997) "Integration of bioaccumulation in an environmental risk assessment" Chemosphere, vol. 34, No. 11, pp. 2337-2350

Fenner-Crisp P (2000) "Endocrine modulators: risk characterization and assessment" Toxicologic Pathology, vol. 28, No. 3, pp. 438-440

Fernández MF, Olmos B, Olea N (2007) "Exposición a disruptores endocrinos y alteraciones del tracto urogenital masculino (criptorquidia e hipospadia)" Gaceta Sanitaria, vol. 21, No. 6, pp.500-14

Flaherty CM, Dodson, SI (2005) "Effects of pharmaceuticals on Daphnia survival, growth, and reproduction" Chemosphere, vol. 61, issue 2, pp. 200-207

Forbes VE, Aufderheide J,Warbritton R,van der Hoeven N,Caspers N (2007) "Does bisphenol A induce superfeminization in Marisa cornuarietis? Part II: toxicity test results and requirements for statistical power analyses" Ecotoxicology and Environmental Safety, vol. 66, No. 3, pp. 319-25

Gaido K, You L and Safe S (2003) "Topic 3.8 Modification of endocrine active potential by mixtures" Pure and Applied Chemistry, IUPAC. vol. 75, Nos. 11-12, pp. 2069-2079

Galloway T "Biomarkers in environmental and human health risk assessment" (2006) Marine Pollution Bulletin, vol. 53, pp. 606-613

Gardiner DM, Ndayibagira A, Grun F, and Blumberg B (2003) "Deformed frogs and environmental retinoids". Pure and Applied Chemistry, vol. 75, pp. 2263-2274

Gasnier C, Dumont C, Benachour N, Clair E, Chagnon MC, Séralini GE (2009) "Glyphosate-based herbicides are toxic and endocrine disruptors in human cell lines" Toxicology, vol. 262, pp. 184-191

Gentile J H and Harwell M A (2001) "Strategies for Assessing Cumulative Ecological Risks" Human and Ecological Risk Assessment: An International Journal, vol. 7, issue 2, pp. 239-246

Goksøyr A "Endocrine disruptors in the marine environment: mechanisms of toxicity and their influence on reproductive processes in fish" (2006) Journal of Toxicology and Environmental Health, Part A, vol. 69, pp.175-184

Goodman M and Laverda N (2002) "Children´s health and environmental exposure to chemicals: implications for risk assessment and public health policy" Human and Ecological Risk Assessment: Theory and Practice, Edited by Dennis J. Paustenbach ISBN 0-471-14747-8, John Wiley & Sons, Inc.

González MM, Martín J, Santos JL, Aparicio I, Alonso E (2010) "Occurrence and risk assessment of nonylphenol and nonylphenol ethoxylates in sewage sludge from different conventional treatment processes" Science of the Total Environment, vol. 408, pp. 563-570

Greene J, Hays S, Paustenbach D (2003) "Basis for a proposed reference dose (RfD) for dioxin of 1-10 pg/kg-day: a weight of evidence evaluation of the human and animal Studies" Journal of Toxicology and Environmental Health, Part B, vol.6, Issue 2, pp. 115-159

Gronen S, Denslow N, Manning S, Barnes S, Barnes D, and Brouwer M (1999) "Serum vitellogenin levels and reproductive impairment of male Japanese medaka (Oryzias latipes) exposed to 4-tert-octylphenol" Environmental Health Perspectives Volume 107, Number 5, pp. 385-390

Heugens, E H W, Hendriks, A J, Dekker, T, van Straalen, N M and Admiraal, W (2002) "A Review of the effects of multiple stressors on aquatic organisms and analysis of uncertainty factors for use in risk assessment" Critical Reviews in Toxicology, vol. 31, issue 3, pp. 247-284

Hewitt LM, Kovacs TG, Dubé M, MacLatchy D, Martel P, MacMaster ME, Paice MG, Parrott JL, Van den Heuvel MR and Van der Kraak GJ (2008)"Altered reproduction in fish exposed to pulp and paper mill effluents: roles of individual compounds and mill operating conditions" Environmental Toxicology and Chemistry, vol. 27, No. 3, pp. 682-697

Hoffmann GR (2009) "A perspective on the scientific, philosophical, and dimensions of hormesis" Dose-Response, vol. 7, pp.1-51

Holbech H, Kinnberg K, Petersen GI, Jackson P, Hylland K, Norrgren L & Bjerregaard P (2006) "Detection of endocrine disrupters: Evaluation of a Fish Sexual Development Test (FSDT) " Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology, vol. 144, No. 1, pp. 57-66

Hood E (2005) "Are EDCs blurring issues of gender ?" Environmental Health Perspectives, vol. 113, No. 10, pp.

International Union of Pure and Applied Chemistry, IUPAC Technical Report (2003) "Endocrine disruptors in the environment" Pure and Applied Chemistry, vol. 75, No. 5, pp. 631-681

IPCS risk assessment terminology (2004) Inter-Organization Programme for the Sound Management of Chemicals. World Health Organization

Isidori M, Lavorgna M, Nardelli A, Parrella A (2006) "Toxicity on crustaceans and endocrine disrupting activity on Saccharomyces cerevisiae of eight alkylphenols"Chemosphere, vol. 64, No. 1, pp. 135-143

IUPAC Glossary of Terms Used in Toxicology, 2nd ed. IUPAC Recommendations 2007. http://sis.nlm.nih.gov/enviro/iupacglossary/glossaryn.html [accessed 08 September 2009]

Jenkins RL, Wilson EM, Angus RA., Howell WM, and Kirk M (2003) "Androstenedione and progesterone in the sediment of a river receiving paper mill effluent" Toxicological Sciences, vol. 73, pp. 53-59

Jobling S, Beresford N, Nolan M, Rodgers-Gray T, Brighty GC, Sumpter JP and Tyler (2002) "Altered sexual maturation and gamete production in wild roach (Rutilus rutilus) Living in Rivers That Receive Treated Sewage Effluents" Biology of Reproduction, vol. 66, pp. 272-281

Jobling S, Williams R,Johnson A, Taylor A, Gross-Sorokin M, Nolan M, Tyler CR,van Aerle R,Santos E and Brighty G (2006) "Predicted exposures to steroid estrogens in U.K. rivers correlate with widespread sexual disruption in wild fish populations" Environmental Health Perspectives, vol. 114(S-1), pp. 32-39. doi: 10.1289/ehp.8050

Jooste S (2001) "A possibilistic approach to diverse-stressor aquatic ecological risk estimation" Available on website http://www.wrc.org.za

Kefford BJ, Zalizniak L, Warne M, Nugegoda D (2007) "Is the integration of hormesis and essentiality into ecotoxicology now opening Pandora's Box?" Environmental Pollution, vol. 151, pp. 516-523

Koh YKK, Chiu TY, Boobis A, Cartmell E, Lester JN, Scrimshaw MD (2007) "Determination of steroid estrogens in wastewater by high performance liquid chromatography-tandem mass spectrometry" Journal of Chromatography A, vol.1173, pp. 81-87

Kortenkamp A (2007) "Ten years of mixing cocktails: a review of combination effects of Endocrine-Disrupting Chemicals" Environmental Health Perspectives, vol. 115, supplement 1, pp. 98-105

Kuiper GGJM, Lemmen JG, Carlsson B, Corton JC, Safe SH, Van der Saag PT, Van der Burg B, and Gustafsson J-A (1998) "Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor β"Endocrinology, vol. 139, No. 10

Landman MJ, Taylor S, van den Heuvel MR (2008) "Carotenoid depletion and ovarian dysfunction in common bully exposed to a New Zealand pulp mill effluent"Aquatic Toxicology, vol. 87, pp. 303-309

Lange IG, Daxenberger A, Schiffer B, Witters H, Ibarreta D, Meyer HHD (2002) "Sex hormones originating from different livestock production systems: fate and potential disrupting activity in the environment" Analytica Chimica Acta, vol. 473, pp. 27-37

Larsson L and Förlin L (2002) "Male-biased sex ratios of fish embryos near a pulp mill: temporary recovery after a short-term shutdown" Environmental Health Perspectives, vol. 110, No. 8, pp. 739-742

Li L., Andersen M., Heber S. and Zhang Q (2007) "Non-monotonic dose-response relationship in steroid hormone receptor-mediated gene expression" Journal of Molecular Endocrinology, vol. 38, pp. 569-85

Linkoy I, Loney D, Cormier S, Satterstrom FK, Bridges T (2009) "Weight-of-evidence evaluation in environmental assessment: review of qualitative and quantitative approaches" Science of the Total Environment 407 (2009) pp. 5199-5205

Liu HL, Liu XH, Wang XY, Wang XR,Yu HX "Toxicity of BPA and TBBPA to Daphnia magna and zebrafish Brachydanio Rerio" (2007) Huan Jing Ke Xue, vol. 28, No. 8, pp.1784-7

Martin OV, Shialis T, Lester JN, Scrimshaw MD, Boobis AR and Voulvoulis N (2008)" Testicular Dysgenesis Syndrome and the estrogen hypothesis: a quantitative meta-analysis" Environmental Health Perspectives, vol. 116, No. 2

Matthiessen P, Arnold D, Johnson AC, Pepper TJ, Pottinger TG, Pulman KGT(2006) "Contamination of headwater streams in the United Kingdom by oestrogenic hormones from livestock farms" Science of the Total Environment, vol. 367, pp. 616-630

Mackay D, Paterson S, Cheung B, Neely WB (1985) "Evaluating the environmental behaviour of chemicals with a level III fugacity model" Chemosphere, vol.14, No.3/4, pp 335-374

McMaster ME., Hewitt LM and Parrott J (2006) "A decade of research on the environmental impacts of pulp and paper mill effluents in Canada: field studies and mechanistic research" Journal of Toxicology and Environmental Health, Part B, vol.9, No.4, pp. 319-339

Menzie CA, MacDonell M M and Mumtaz M (2007) "A phased approach for assessing combined effects from multiple stressors" Environmental Health Perspectives, vol. 115, No. 5, pp. 807-816

Meriläinen P, Oikari A (2008) "Exposure assessment of fishes to a modern pulp and paper mill effluents after a black liquor spill" Environmental Monitoring Assessment, vol. 144, pp. 419-435

Middaugh DP, Beckham N, Fournie JW, Deardorff TL(1997)"Evaluation of bleached Kraft mill process water using Microtox®, Ceriodaphnia dubia, and Menidia beryllina toxicity tests" Archives of Environmental Contamination and Toxicology, vol. 32, pp. 367-375

Mills LJ, Chichester C (2005) "Review of evidence: Are endocrine-disrupting chemicals in the aquatic environment impacting fish populations?" Science of the Total Environment, vol. 343, pp. 1- 34

Mills LJ, Gutjahr-Gobell RE, Borsay Horowitz D, Denslow ND, Chow MC, and Zaroogian GE (2003)"Relationship between reproductive success and male plasma vitellogenin concentrations in cunner, Tautogolabrus adspersus" Environmental Health Perspectives, vol.111, No.1

Munkittrick, KR, Van Der Kraak, GJ., McMaster, ME, Portt, CB, van den Heuvel, MR & Servos MR (1994) "Survey of receiving-water environmental impacts associated with discharges from pulp mills.2.Gonad size, liver size, hepatic EROD activity and plasma sex steroid levels in white sucker". Environmental Toxicology and Chemistry, vol. 13, No. 7, pp.1089-1101

Mushak, P (2007) "Hormesis and its place in nonmonotonic dose-response relationships: some scientific reality checks". Enviromental Health Perspectives, vol. 115, number 4, pp. 500-506

Nakari T, Erkomaa K (2003) "Effects of phytosterols on zebrafish reproduction in multigeneration test" Environmental Pollution, vol. 123, No. 2, pp. 267-273

NRC (National Research Council) (1994)"Science and Judgment in Risk Assessment" Washington, DC: National Academy Press.

NRC (National Research Council) (1983) "Risk Assessment in the Federal Government: Managing the Process". Washington, DC: National Academy Press.

Nilsson S, Mäkelä S, Treuter E, Tujague M, Thomsen J, Andersson G, Enmark E, Pettersson K, Warner M, Gustafsson JA. "Mechanisms of estrogen action"(2001) Physiology Reviews, vol. 81, issue 4, pp. 1535-65

Novák J, Beníšek M, Hilscherová K(2008)"Disruption of retinoid transport, metabolism and signaling by environmental pollutants" Environment International, vol. 34 pp. 898-913

OECD- Organisation for economic Co-operation and development (2003) " Emerging risks in the 21st century: an agenda for action.

Owens JW, Swanson SM, and Birkholz DA (1994) "Environmental monitoring of bleached Kraft pulp mill chlorophenolic compounds in a northern Canadian river system" Chemosphere, vol. 29, No. 1, pp. 89-109

Paddle GM and Harrington JM (2000) "Environmental epidemiology- strengths and weaknesses" International Archives of Occupational Environmental Health, vol. 73, pp. 7-14

Palacea V P, Klaverkamp JF, Baron CL and Brown SB (1997) "Metabolism of 3H-retinol by lake trout (Salvelinus namaycush) pre-exposed to 3,3′,4,4′,5-pentachlorobiphenyl (PCB 126) " Aquatic Toxicology, vol. 39, issues 3-4, pp. 321-332

Panter GH, Thompson RS, Sumpter JP (1998) "Adverse reproductive effects in male fathead minnows (Pimephales promelas) exposed to environmentally relevant concentrations of the natural oestrogens, oestradiol and oestrone" Aquatic Toxicology, vol. 42, pp. 243-53

Parks LG, Lambright CS, Orlando EF, Guillette LJ, Jr., Ankley GT, and Gray LE, Jr. (2001) "Masculinization of female mosquitofish in Kraft mill effluent - contaminated Fenholloway River water is associated with androgen receptor agonist activity". Toxicological Sciences, vol. 62, pp. 257-267

Paustenbach DJ (2000) "The practice of exposure assessment: a state-of-the-art review". Journal of Toxicology and Environmental Health, Part B, No. 3, pp. 179-291, 2000

Penman DJ and Piferrer F (2008) "Fish Gonadogenesis Par t I: Genetic and environmental mechanisms of sex determination". Reviews in Fisheries Science, vol. 16 issue S1, pp. 14-32

Peruzzo PJ, Porta AA, Ronco AE (2008) "Levels of glyphosate in surface waters, sediments and soils associated with direct sowing soybean cultivation in north pampasic region of Argentina" Environmental Pollution, vol. 156, pp. 61-66

Pollard SJT, Davies GJ, Coley F, Lemon M (2008) "Better environmental decision making - recent progress and future trends" Science of the total environment, vol. 400, pp. 20-31

Regulation (EC) No 1907/2006 of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH). Official Journal of the European Union 29.5.2007

Rhodes S, Farwell A, Hewitt LM, MacKinnon M, George Dixon D (2005) "The effects of dimethylated and alkylated polycyclic aromatic hydrocarbons on the embryonic development of the Japanese medaka" Ecotoxicology and Environmental Safety, vol. 60, No. 3, pp. 247-258

Richard S, Moslemi S, Sipahutar H, Benachour N, and Séralini GE (2005)"Differential effects of glyphosate and Roundup on human placental cells and aromatase" Environmental Health Perspectives, vol. 113, No. 6, pp. 716-720

Roseberry AM and Burmaster DE (1991) "Lognormal distributions for water intake by children and adults” Risk Analysis, vol. 12, No. 1, pp.99-104

Routledge EJ and Sumpter JP (1996) "Estrogenic activity of surfactants and some of their degradation products using a recombinant yeast screen" Environmental Toxicology and Chemistry, vol. 15, No. 3, pp. 241-248

Safe SH (1998) "Hazard and risk asessment of chemical mixtures using the Toxic Equivalency Factor approach" Environmental Health Perspectives, vol. 106, Supplement 4, pp. 1051-1058

Safe SH, Pallaroni L,Yoon K,Gaido K, Ross S, and McDonnell D (2002) "Problems for risk assessment of endocrine-active estrogenic compounds" Environmental Health Perspectives, vol. 110 (supplement 6), pp. 925-929

Sánchez-Avila J, Bonet J, Velasco G, Lacorte S (2009) "Determination and occurrence of phthalates, alkylphenols, bisphenol A, PBDEs, PCBs and PAHs in an industrial sewage grid discharging to a municipal wastewater treatment plant" Science of the Total Environment, vol. 407, pp. 4157-4167

Satoh K, Nagai F, and Aoki N (2002) "Several environmental pollutants have binding affinities for both androgen receptor and estrogen receptor". Journal of Health Science, vol. 47, issue 5, pp. 495-501

Schwaiger J, Mallow U, Ferling H, Knoerr S, Braunbeck T, Kalbfus W, Negele RD (2002) "How estrogenic is nonylphenol? A transgenerational study using rainbow trout (Oncorhynchus mykiss) as a test organism" Aquatic Toxicology, vol. 59, Nos. 3-4, pp. 177-189

Selim K M, Shinomiya A, Otake H, Hamaguchi S, Sakaizumi (2009) "Effects of high temperature on sex differentiation and germ cell population in medaka, Oryzias latipes". Aquaculture vol. 289, pp. 340-349

Sişman T, Geyikoğlu F, Atamanalp M (2007) "Early life-stage toxicity in zebrafish (Danio rerio) following embryonal exposure to selected polychlorinated biphenyls" Toxicology Indian Health, vol. 23, No.9, pp-529-36

Sohoni P,Tyler CR, Hurd K,Caunter J, Hetheridge M, Williams T,Woods C, Evans M, Toy R, Gargas M, Sumpter JP (2001) "Reproductive effects of long-term exposure to Bisphenol A in the fathead minnow (Pimephales promelas) " Environmental Science & Technology, vol. 35, No. 14, pp.: 2917-25

Solomon GM and Schettler T (2000) "Environment and health: 6. Endocrine disruption and potential human health implications", Canadian Medical Association Journal, vol.163 , No.11, pp. 1471-76

Stair P (2008) "Male reproduction health declines" Worldwatch Institute.Vision for a sustainable world, http://www.worldwatch.org/node/5484

Stebbing ARD (1997) "A theory for growth hormesis" BELLE Newsletter Vol 6, No. 1

Sun L, Zha J, Wang Z (2009) "Interactions between estrogenic chemicals in binary mixtures investigated using vitellogenin induction and factorial analysis". Chemosphere, vol. 75, pp. 410-415

Suter G, Vermeire T, Munns W, Sekizawa J (2003) "Framework for the integration of Health and Ecological Risk Assessment". Human and Ecological Risk Assessment: vol. 9, No.1, pp. 281-301

Tanaka JN, Grizzle JM (2002) "Effects of nonylphenol on the gonadal differentiation of the hermaphroditic fish, Rivulus marmoratus" Aquatic Toxicology, vol. 57, No. 3, pp. 117-125

Tarasova EN, Mamontov AA, Mamontova EA (1997) "Polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) in Baikal Seal" Chemosphere, vol. 34, No. 11, pp. 2419-2427

Tate TM, Spurlock JO, Christian FA (1997) "Effect of glyphosate on the development of Pseudosuccinea columella snails" Archives of Environmental Contamination and Toxicology, vol. 33, No. 3, pp. 286-89

Thankamony A, Ong KK, Dunger DV, Acerini CL, Hughes IA (2009) "Anogenital distance from birth to two years: a population study" Environmental Health Perspectives. doi: 10.1289/ehp.0900881

Travis CC and Arms AD (1988) "Bioconcentration of organics in beef, milk, and vegetation" Environmental Science and Technology, vol. 22, No. 3

Tsui, M.T.K., Wang, W.X. 2005 "Influences of maternal exposure on the tolerance and physiological performance of Daphnia magna under mercury stress" Environmental Toxicology and Chemistry, vol. 24, issue 5, pp. 1228-1234

United States Environmental Protection Agency, U.S. EPA 2009a "Aggregate and cumulative risk" http://www.epa.gov/heasd/risk/index.htm [accessed 19 August 2009]

United States Environmental Protection Agency, U.S. EPA 2009b "Model development- exposure models" http://www.epa.gov/heasd/risk/projects/c1b_exposure_models_development.htm [accessed 19 August 2009]

United States Environmental Protection Agency, U.S. EPA 2009c "Interpretation and use of biomarkers" http://www.epa.gov/heasd/risk/research/rc2_biomarkers_interpretation.htm [accessed 19 August 2009]

United States Environmental Protection Agency, EPA 2007a "Concepts, methods and data sources for Cumulative Health Risk Assessment of multiple chemicals, exposures and effects: a resource document" EPA/600/R-06/013F, 2007

United States Environmental Protection Agency, EPA 2007b "Aromatase Integrated Summary Report ii"

United States Environmental Protection Agency, U.S. EPA 2003 "Framework for Cumulative Risk Assessment" U.S. Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment, Washington Office, Washington, DC, EPA/630/P-02/001F, 2003

United States Environmental Protection Agency, EPA (1996) "Guidelines for Reproductive Toxicity Risk Assessment of the US Environmental Protection Agency" EPA/630/R-96/009

U.S. Geological Survey "Glyphosate herbicide found in many midwestern streams, antibiotics not common"

Vermeire T, Munns W, Sekizawa J, Suter G,Van der Kraak G ( 2007) "An assessment of integrated risk assessment". Human and Ecological Risk Assessment, vol. 13, pp. 339-354

Vizziano D, Baron D, Randuineau G, Mahe`S, Cauty Cand Guigue Y (2008) "Rainbow trout gonadal masculinization induced by inhibition of estrogen synthesis is more physiological than masculinization induced by androgen supplementation". Biology of reproduction, vol. 78, pp. 939-946

Wartman CA, Hogan NS, Hewitt LM, McMaster ME, Landman MJ, Taylor S, Kovac TG, van den Heuvel MR (2009) "Androgenic effects of a Canadian bleached Kraft pulp and paper effluent as assessed using threespine stickleback (Gasterosteus aculeatus)"Aquatic Toxicology, vol. 92, pp. 131-139

WHO-IPCS 2002 "Global assessment of the state-of-the-science of endocrine disruptors" International Programme on Chemical Safety, World Health Organization WHO/PCS/EDC/02.2

Williams PRD, Paustenbach D (2002) "Risk characterization: principles and practice" Journal of Toxicology and Environmental Health, Part B, No. 5, pp. 337-406

Wilson VS, Bobseine K and Gray LE (2004) "Development and characterization of a cell line that stably expresses an estrogen-responsive luciferase reporter for the detection of estrogen receptor agonist and antagonists" Toxicological Sciences, vol. 81, pp. 69-77

Zalizniak L, Nugegoda D (2006) "Effect of sublethal concentrations of chlorpyrifos on three successive generations of Daphnia carinata" Ecotoxicology and Environmental Safety, vol. 64, issue 2, pp. 207-214

Zartarian V and Schultz B (2009) "The EPA's human exposure research program for assessing cumulative risk in communities" Journal of Exposure Science and Environmental Epidemiology, pp.1-8

Zhang Q, Pi J, Woods C, Jarabek AM, Clewell HJ III and Andersen ME (2008) "Hormesis and adaptive cellular control systems" Dose-Response, vol. 6, pp.196-208Hormesis and Adaptive Cellular Control Systems.