Dr Mel Cooke explains what endocrine disruptors are, what hazards they pose, where to get information about them and what to do about them.


Between 1940 and 1971, 5–10 million people in the USA were exposed to diethylstilboestrol (DES) by which time it was clear that females exposed to DES in the womb might suffer a rare type of cancer (clear cell adenocarcinoma) during puberty, and an increased rate of infertility in their adult life. Exposed males were prone to structural abnormalities of the genitals.

Tributyl tin (TBT) is a marine anti-foulant used since the mid-1960s on the hulls of boats, from where it leaches into the sea. Female dogwhelks exposed to TBT develop male physical characteristics (a condition known as imposex), and female periwinkles show a progressive replacement of the female reproductive tract by male sexual organs (a condition known as intersex).

These apparently disparate effects of chemical exposure are linked by a common toxicological mechanism: they both occur through what has become known as endocrine disruption.

Regulation 1907/2006/EC (REACH), Article 57(f) includes substances with “endocrine disrupting properties” as being of equal concern as carcinogens, mutagens and reproductive toxicants (CMRs), and can therefore also be subject to Authorisation and forced withdrawal from the market. Article 57(f) states:

“(f) substances — such as those having endocrine disrupting properties or those having persistent, bioaccumulative properties, which do not fulfil the criteria of points (d) or (e) — for which there is scientific evidence of probable serious effects to human health or the environment which give rise to an equivalent level of concern to those of other substances listed in points (a) to (e) and which are identified on a case-by-case basis in accordance with the procedures set out in Article 59.”

What are these substances? What do they do? How do we find out about them? What do we do about them? These are the questions this article will address.

Endocrine system

The endocrine system in animals is a communication network involving glands, hormones, and receptors, complementary to the nervous system. In contrast to neural messaging, the endocrine system is typically slower, as hormones use the bloodstream as the highway between the signal-producing organ and the target sites. Although several glands are involved in hormone production, there are three major signalling pathways: hypothalamus–pituitary–gonad; hypothalamus–pituitary–adrenal; and hypothalamus–pituitary–thyroid. The hypothalamus produces and controls several hormones, influencing the pituitary, which itself releases hormones to particularly affect the gonads, adrenal and thyroid organs.

These organs regulate many important aspects of human and animal life, such as changes during pregnancy, menstruation, sperm counts, timing of puberty, weight gain, aggression, and gender determination (in turtles).

In the womb, and during puberty, hormones are key factors for the proper development of many organs, particularly the reproductive tract, brain and neuro-endocrine system. This development is exquisitely sensitive to chemical exposure. If adverse effects occur, they are in many cases irreversible and stay with the affected person or organism for the rest of its life.

As we have already seen, the endocrine system is complex, and chemicals can interact with it in any number of ways. The most important mechanisms of interaction are as follows.

  • Agonistic effect: the chemical mimics the hormone, binds to the hormone receptor, and initiates a cellular response as if the natural hormone had been present. Thus the hormone “message” could be delivered at the wrong time.

  • Antagonistic effect: the chemical binds to the hormone receptor, does not initiate a cellular response, and prevents the binding of the natural hormone, preventing the hormone message from getting through.

  • Binding to transport proteins: this alters the ability of the hormone to travel in the bloodstream.

  • Interference with metabolic processes: altering the synthesis or breakdown of hormones.

Up to now, the most widely studied endocrine effects are related to only three hormones: oestrogens (female sexual development and reproduction), androgens (particularly testosterone for development and maintenance of male sexual characteristics), and thyroxine (control of reproduction and metabolic rate).

Endocrine disrupters (EDs)

An internationally accepted definition of EDs is given by the World Health Organization (WHO/IPCS) as follows: “An endocrine disrupter is an exogenous substance or mixture that alters function(s) of the endocrine system and consequently causes adverse health effects in an intact organism, or its progeny, or (sub)populations.”

This definition is important, because it brings some clarity to what continues to be a hugely complex and controversial topic.

The first feature is that of “adverse effect” in an intact organism, ie not merely a biochemical interaction such as receptor binding, or even a hormone-modulating effect. Many chemicals are capable of interacting with steroid receptors, and might be termed as having endocrine activity, but it is far from clear whether such interaction leads to any adverse effects. The endocrine system in an animal can often adapt to disturbance by chemicals by feedback mechanisms and compensation. This raises a difficult regulatory question: when is the line of adversity crossed?

Because EDs must have effects in intact organisms, substances with effects seen only in in vitro (test tube) screening tests cannot be regarded as EDs. This should prevent too many innocent substances being regarded as EDs, with the consequent regulatory burden (see below).

Finally, the definition specifies a causal link between the chemical and the adverse effect. This can be much more difficult to show than one might imagine, particularly for the following reasons.

  • There might be a long time between the exposure and the effect, so the chemical is not present when the effect is manifest. Endocrine disruption manifests only if the organism is exposed during certain sensitive stages of its lifecycle, particularly in the womb and during puberty.

  • The disruption of hormone signalling often produces a tell-tale collection of effects in organs (a fingerprint) rather than a single, measurable criterion.

  • The endocrine system is complicated, and proof of causality would require specific understanding of exactly how the substance interacts with it, and the consequences.

  • Wildlife is included in the definition, but little is known about their endocrinology. And what about invertebrates, plants, fungi and bacteria? These organisms are all important parts of the ecosystem.

  • Current REACH testing is not aimed at detecting EDs. Some higher level (Annexes IX and X) testing, particularly in the rat, may pick up some endocrine effects, but these tests are not particularly sensitive and are conducted only for a limited number of substances.


REACH uses terminology that says the ED must cause probable serious effects before it becomes a substance of very high concern (SVHC), which might be seen as requiring a lower burden of proof than the WHO definition above.

All the rage

With the difficulty in proof of association of chemical products with endocrine effects, the growing controversy has polarised the chemical industry and many non-governmental organisations (NGOs). Few chemicals have definitive proof of endocrine disruption. Those that have proof tend to be the usual heavily researched “legacy chemicals”, eg PCBs, DDT and PAHs, which have been banned or heavily regulated for many years, but remain at measurable quantities in the environment. How should the chemical industry respond without proper evidence for specific substances?

On the other hand, the NGOs cite epidemiological evidence. The past 20 years indicate increasing trends in endocrine-related disorders in humans. Evidence is strengthening that chemical exposures are involved. The concern is that populations cannot be protected when the damage is diagnosed long after the causative exposures have taken place, as was the case with PCBs.

There is a plausible role for exposure to EDs in breast, prostate, testicular and thyroid cancers.

There are reported declines in men’s reproductive health, particularly semen quality, and increased problems in boys of testicular formation and genital abnormalities. There is good, coherent evidence that suppression of androgen action in the womb can interfere with male reproductive development.

Male fish exposed to sewage treatment works effluent can develop sex organ malformation and have impaired reproductive function. Pulp and paper mill effluents can have similar effects. Fish taken from water courses contaminated with perchlorate have thyroid abnormalities.

In birds, oestrogenic EDs (such as DDT) administered in the egg can induce malformation of the female reproductive tract and alter sex hormone levels in the adult, reducing reproductive fitness. Birds can suffer dramatic changes in reproductive behaviours, including increased incidences of homosexual pairings.

Experimental animals exposed in the womb to androgen receptor antagonists (eg dicarboximide, imidazole and azole pesticides, and some phthalates) may have reduced male hormone activity, which only becomes apparent in adult life.

Oestrogenic chemicals can influence the timing of puberty in rodents.

There are many examples of vulnerable life stages in wildlife species, including lobsters, amphibians and reptiles, which are extremely sensitive to the influences of EDs.

Common chemicals implicated as EDs

The last 10 years have seen increased concern regarding EDs, both in the number of chemicals implicated and the variety of interaction with the endocrine system. This concern is magnified by the identification of several ubiquitous chemicals as potential EDs.

Perfluorooctanoic acid (PFOA) is widespread at low levels in the environment and in the blood of the general population. It is used to make fluoropolymers, and it can be produced by the breakdown in the environment of some perfluorinated surface-treatment products. PFOA inhibits thyroid function and reduces thyroid hormone levels in both humans and experimental animals, and can cause irreversible neurological changes in the womb. In humans, researchers have linked perfluorinated compounds (PFCs) in blood levels and ADHD in children. Epidemiology shows an association between PFC body burdens and increased cholesterol levels. The strength of the evidence is controversial.

Bisphenol A (BPA) is produced in large quantities for use in the production of polycarbonate plastics and epoxy resins, which have applications including food and drink packaging, baby bottles, medical devices, coatings for food cans, bottle tops, water supply pipes, and dental sealants. BPA can bind to hormone (oestrogen and progesterone) receptors and can act as a thyroid hormone antagonist. Scientists have demonstrated that exposure to BPA during organ development may cause irreversible adverse effects on the prostate gland, and changes in mammary tissue. There are emerging potential risks for cancer and neurological development.

Phthalates are the most widely used plasticizers, and have been used for about 50 years to make polyvinyl chloride (PVC) soft, flexible, and durable. Other uses are for coatings, such as nail polish, adhesives, sealants, and paints. US research provides good evidence of irreversible genital traits in young boys (eg problems with testicular descent, smaller genitals) associated with higher exposure to phthalates in the womb. Prenatal phthalate exposure is a possible risk factor in ADHD, lowered IQ, obesity and diabetes. In experimental animals, phthalates produce a series of irreversible effects in male offspring (phthalate syndrome), characterised by underdeveloped or malformed reproductive organs and difficulties with descent of the testes. These effects can be traced to interference with testosterone synthesis in foetal life.

Parabens are the most widely used preservatives in cosmetic products. They are emerging as compounds of concern. They are metabolised very quickly by the body, so their metabolites are likely to play a role in any endocrine disruption. In experimental animals, paraben administration has been shown to reduce testosterone levels in a dose-dependent manner and reduce sperm production. In vitro, parabens have been shown to have oestrogenic and anti-androgenic properties in receptor binding studies.

Plot thickeners

The chemical industry, in response to these claims and concerns, often counters that the effects are not relevant for realistic exposures. A basic toxicological principle is that the dose makes the poison, or in other words that all substances have detrimental effects if the organism is given sufficiently high doses. A key question for EDs is at what dose do the adverse effects occur in an intact animal? In general toxicological terms, this is usually a straightforward question – the assessor can determine a no-observed-adverse-effect level (NOAEL), derived from long-term animal studies. (This parameter determines the Derived-No-Effect Level, which is a crucial output of the REACH chemical safety report, and is communicated down the supply chain in the safety data sheet.)

However, some scientists and advocates of stricter control of EDs regard the traditional approach to chemical risk assessment as too simplistic, as follows.

Cocktail or mixture effects: There is good evidence that several EDs can work together. Exposure to multiple chemicals with the same endocrine mechanism can produce adverse effects where the individual chemicals would show no effect. It is possible that EDs give a synergistic effect. The risk associated with EDs may be underestimated because environmental exposures to mixtures of chemicals are not considered during risk assessment of substances. Because of the ubiquitous nature of some potential EDs, some commentators have described the environment as a “sea of oestrogens”.

Low-dose effects: non-standard dose–response relationships have been observed with EDs, so that effects have occurred at low doses that would not have been predicted from high-dose data. Most standard testing (eg for REACH) is done using comparatively high doses, so some argue that important low-dose effects have been missed. Others hotly contest the validity of this argument and the supporting data. These low-dose effects have sometimes been difficult to replicate in other laboratories. Campaigners cite the low-dose effect as a failure of the regulatory system, and therefore request the banning of potential EDs as a precautionary measure.


REACH testing

Validated test methods, such as those prescribed in REACH Annexes IX and X, are useful, but would catch only a limited range of the known endocrine disrupting effects. Standard tests may not detect relatively rare events such as reproductive tract malformation, because the low number of animals used does not give statistical power. Big gaps occur for the testing strategy for wildlife species. For a wide range of endocrine disrupting effects, validated test methods do not exist.

REACH testing does not mandate tests that are specifically designed to assess endocrine disrupting ability. However, some effects of endocrine disruption may be captured during REACH testing in the following areas: repeated-dose toxicology, carcinogenicity, mutagenicity, reproduction and development. This list shows a difficulty in the current testing battery, in that endocrine disrupting effects may be spread over many disciplines, and it might well be that no one is looking for the tell-tale “fingerprint” effect of EDs. Another way of stating this is that the definition of EDs is based on mechanism, but not on a specific adverse effect.

For example, for repeat-dose toxicity, the likely tests for REACH are the 28 day (methods: EU B.7; OECD TG 407) or 90 day (methods: EU B.26; OECD 408) oral studies in the rat. These tests do measure some parameters which are relevant to endocrine-mediated toxicity, particularly the weight and histopathology of the endocrine organs: pituitary, adrenal, ovaries and prostate. Some thyroid measurements are optional, and the test lacks important measurements related to the testes. Only adult animals are exposed. Note that the measurements that are made are not the most sensitive for revealing endocrine disruption. Finally, if these tests show low toxicity, then further tests related to potential carcinogenic or reproductive effects may not be required, which might allow potential EDs to slip through the net.

Under REACH only short-term environmental tests is required up until the 100 tonnes/year threshold. These tests measure simple endpoints and do not help in detecting EDs. REACH may require long-term testing for substances supplied at >100 tonnes/year, but European Chemicals Agency (ECHA) guidance indicates that there is no requirement to provide information on the substance’s endocrine activity, or its reproductive or specific developmental toxicity in aquatic organisms.

Specific tests for EDs

There is a wide variety of evidence for EDs, from structure–activity relationships, in vitro tests relating to receptor binding, and then in vivo testing right up to multigenerational studies. Some of these tests are not internationally validated, so their use in regulatory decision making is limited. It is not possible to cover the full extent of research into test methods for EDs.

The OECD has published a testing strategy for the detection of EDs. “Conceptual Framework for the Testing and Assessment of Endocrine Disrupting Chemicals”.

Not all of the tests would be required to determine the endocrine disrupting properties of a substance. The assessor needs to choose appropriate tests based on available information, in a weight-of-evidence approach.

Even the latest guidance for level 5 assays has considerable gaps in measurements relevant for the detection of EDs, and even well researched chemicals, such as 4-octylphenol, have an incomplete dataset. Without a comprehensive level 5 assay, the assessor can make no definitive conclusion as to whether a substance is an ED.

However REACH testing may not be capable of detecting EDs.

Key tests for EDs

The uterotrophic assay is an in vivo assay for the oestrogen activity of a test chemical. The growth phase of the uterus in the natural oestral cycle is controlled by estrogens. The test often uses ovariectomised animals to remove endogenous oestrogen, so that the growth of the uterus becomes sensitive to external sources of oestrogen or oestrogen-like substances. The primary measurement in this assay is uterine weight. Chemicals that act as oestrogen agonists would be expected to cause an increase in uterine weight, while oestrogen antagonists, when co-administered with oestrogen, would be expected to cause a decreased uterine growth.

The Hershberger assay assesses the androgenic activity of a test chemical. The test substance is given for 10 consecutive days, usually to castrated male rats. These animals are sensitive to exogenous androgens, or androgen-like substances. The assessor looks for changes in weight in androgen-dependent tissues, such as the prostate, seminal vesicle and penis.

Note that the above tests are frequently performed on castrated or ovariectomised animals to reduce natural hormone levels, but this precludes tested chemicals as fulfilling the ED definition, as they are not tested in “intact animals”.

The amphibian metamorphosis assay helps to identify substances which may interfere with the hypothalamous–pituitary–thyroid axis. The transformation from tadpole to frog is driven by thyroid hormones. In the test, tadpoles are exposed to a test chemical in water for 21 days. The assessor looks for changes to the shape of the tadpole (eg length, limb length, weight, and developmental stage), and histology of the thyroid gland.

Regulatory control of EDs

REACH aims to ensure a high level of protection of human health and the environment. An important process that REACH uses to achieve this is evaluation of registration dossiers. ECHA makes a list of substances of concern on the basis of hazard (eg CMR and EDs), exposure information, and manufacture or import tonnage.

The registration dossiers of these substances are evaluated by Member State authorities in a Community Rolling Action Plan (CoRAP). The current CoRAP list contains 90 substances, including 20 suspected EDs (dichloromethane, bisphenol A, ziram, triclosan, methylstyrenated phenol, carbon disulphide, 2-(2-butoxyethoxy)ethyl 6-propylpiperonyl ether, 4,4’-sulphonyldiphenol, diethyl phthalate, 2,4-di-tert-butylphenol, methyl 4-hydroxybenzoate, 4-hydroxybenzoic acid, p-cresol, resorcinol, triphenyl phosphate, thiram, diuron, tert-butyl methyl ether, 2-ethylhexyl 4-methoxycinnamate, 4-nonylphenol branched).

If the outcome of evaluation is unfavourable, the substance may be recommended for restrictions on its marketing and use. The two mechanisms for this are placement in REACH Annex XVII (Restrictions on the Manufacture, Placing on the Market and Use of Certain Dangerous Substances, Mixtures and Articles). This allows industry to continue to use the substance, but only under the terms stated in the Annex.

For the most severe hazards, the substance is placed on the Candidate List as an important step to inclusion in Annex XIV of REACH (List of Substances Subject to Authorisation). These severe hazards are stated in Article 57 as being carcinogens, mutagens, and reproductive toxicants (Category 1A or 1B, according to the CLP criteria), persistent, bioaccumulative substances that may accumulate through the food chain in the environment, and endocrine disrupting properties. Only one ED so far has been included in the Candidate List: 4-(1,1,3,3-tetramethylbutyl)phenol (CAS 140-66-9).

If a substance is identified as an ED in the Candidate List, there are consequences for hazard communication (REACH Article 33): companies who supply finished goods (known as “articles” in REACH) containing Candidate List substances present above 0.1% w/w must give the recipient of the article sufficient information to allow safe use of the item — as a minimum, the name of that Candidate List substance.

The supplier is also obliged to give the same information upon request to a consumer within 45 days of receipt of the request. In addition, any importer or producer of articles may be required to notify ECHA if a Candidate List substance is present in the articles above 0.1% w/w and the quantity of the substance totals above 1 tonne per year. The duty to notify began in June 2011. There is a summary of the Candidate List obligations on the ECHA website.

Title VII of REACH describes the Authorisation procedure for SVHCs, with the aim of encouraging substitution of SVHCs with safer alternatives. However, if industry applies successfully, it can get an authorisation to continue using a substance that is on REACH Annex XIV. For each application, ECHA will launch an 8-week public consultation to identify possible alternatives to the proposed Annex XIV substance for the particular uses. If the applicant can demonstrate, using a socio-economic analysis, that continuing to use the substance is better than using alternatives, then the application is accepted. Currently, Annex XIV does not contain substances specifically identified as EDs, although some phthalates (see above discussion on phthalates as EDs), DEHP (CAS 117-81-7), BBP (85-68-7), DBP (CAS 84-74-2) and DIBP (CAS 84-69-5) are included as reproductive toxicants.


Chemical suppliers are obliged to classify their products according to certain hazard criteria, as prescribed by Regulation 1272/2008/EC (CLP) EDs are classified in the same way as any other chemical.

The classification of EDs under Specific Target Organ Toxicity–Repeated Exposure (STOT RE) from repeated dose toxicity studies may be appropriate, as ED effects are often manifested in specific organs (particularly sex organs). On the other hand, STOT RE classifications typically focus on tests that do not involve exposure in the womb, STOT-RE substances do not pose concerns equivalent to CMRs, and finally STOT RE does not deal with wildlife effects.

Classifying EDs as CMRs is also plausible, as EDs frequently manifest themselves in terms of carcinogenicity and reproductive toxicity. However, some ED effects are outside this scope, eg behavioural effects. Also classification for CMRs does not include wildlife effects.

Finally, for the environment, the current reliance on acute toxicity data for classification means that unless the substances are also PBT or vPvB, EDs are unlikely to be classified as hazardous to the aquatic environment. However, it should be noted that Regulation 286/2011 (which amends the CLP Regulation) increases the reliance on test results from long-term aquatic tests for classification.

These shortcomings in the current classification system have led some to propose a separate class for EDs. This would allow flexibility as scientific knowledge increases, the development of specific testing to be included in the classification criteria, and finally human and wildlife effects could be dealt with in one class in a coherent fashion. The Danish EPA has proposed a new classification class for EDs, similar to CMRs, as follows: Category 1, confirmed EDs (based on in vivo data); Category 2a, suspected EDs (mainly based on in vivo data), Category 2b, indicated EDs (mainly based on in vitro/in silico data).

Changes to classification criteria by petitioning at UN level for the GHS, would take many years to accomplish.

Conclusions and recommendations

In order to prevent the unnecessary restriction or authorisation of chemicals, and the consequent commercial impact, the ED tag should be reserved for substances where ED effect is clearly established, ie that the endocrine disrupting property is an important feature of its hazard profile.

The WHO/IPCS definition is considered very broad, and does not discriminate between EDs of high and low concern. EDs are recognised within REACH as being SVHCs. However, a disadvantage of this is that EDs of a lower hazard might go unregulated.

Very many chemicals interact with the endocrine system, as determined by receptor binding assays and other in vitro testing, and so are potential EDs. Many other substances have been implicated as potential EDs by structure–activity relationships and computational methods. This high-throughput screening approach is prevalent in the USA.

These methods are comparatively cheap, but positive results in vitro or in silico do not predict adverse effects in animals, and the assessor must use a weight of evidence approach until definitive testing, involving resource-intensive multi-generational mammal and wildlife tests are performed. Testing for EDs also requires non-standard testing, but overall there are also many gaps in the testing strategy, particularly for assessing endocrine disruption in wildlife.

REACH testing is not aimed at finding potential EDs, and is not adequate for evaluating potential EDs, particularly at lower tonnage thresholds. There is therefore legitimate concern that REACH is not achieving its stated aim of high protection for human health and the environment. The situation could be improved by updating current REACH testing methods to include ED measurements.

It may be beneficial to form a separate regulatory classification for EDs, based on potency, specificity, severity and irreversibility.

Last reviewed 1 August 2012