Occurrence of chemicals with known or suspected endocrine disrupting activity in drinking water, groundwater and surface water, Austria 2017/2018

Twenty-two groundwater sites and twenty nearby drinking water sites all over Austria were selected for sampling. The choice of sites was based on risk-assessment according to the presence of substances previously detected by AGES and the federal environment agency (Ages and Umweltbundesamt, 2015) indicating contamination by humans. Moreover, five lakes used for bathing (bathing water) and seven rivers, five of them transboundary sampling sites, were arbitrarily selected. Figure 1 shows the location of the 54 sample sites. Drinking water, groundwater and surface water were sampled by the accredited sampling personnel from August 2017 until February 2018, within the framework of the Austrian ordinance on the monitoring of the quality of water (“Gewässerzustandsüberwachungsverordnung”, “GZÜV”), Federal Law Gazette No 479/2006, following the requirements of ISO 5667-4 (Guidance on sampling from lakes, natural and man-made), ISO 5667-5 (Guidance on sampling of drinking water from treatment works and piped distribution systems), ISO 5667-6 (Guidance on sampling of rivers and streams), and ISO 5667-11 (Guidance on sampling of groundwater). Samples (one 1-litre glass bottle, and two 1-litre aluminium bottles each) were transported cooled at 4°C, and tested within 72 hours. Samples were quantitatively analyzed on estrone (E1), 17ß-estradiol (E2), estriol (E3), 17α-ethinylestradiol (EE2), bisphenol A/S (BPA/BPS), triclosan (TCS), non-ylphenol-mono- ethoxylate (NP1EO), nonylphenol-diethoxylate (NP2EO), octylphenol (OP) and its ethoxylates (OP1EO, OP2EO), perfluorooctanoic acid (PFOA), perfluorooctanesulfonic acid (PFOS), benzylbutylphthalate, di-(2-ethylhexyl)phthalate, di-isononylphthalate, di-n-butylphthalate, dimethylphthalate, di-n-octylnonylphthalate, diethylphthalate, diisodecylphthalate, and polybrominated diphenylethers BDE 28, 47, 99, 100, 153, 154.

Figure 1

The realization that hormones can reach the environment, even after treatment of sewage water, strengthened the search for the presence of pharmaceuticals and endocrine disrupting chemicals in water sources, such as sewage water effluent, surface water and groundwater (Kuch and Ballschmiter, 2001). These substances were associated with changes in microbial ecology, toxicological effects in aquatic species, and a negative effect on human health (Shore and Shemesh, 2016). Estrogen hormones are crucial for human biology and physiology. They help regulate reproduction, cardiovascular function, bone strength, cognitive behavior, successful pregnancy and gastrointestinal systems. Nevertheless, Adeel et al. (2017) illustrate in their review that they can have serious adverse effects if allowed to accumulate in the environment and to enter the human food chain. If consumed at levels above safe thresholds, they can increase the risk of cancer and induce cardiovascular diseases in humans. Supra-optimal levels of estrogens have been linked with increased incidence of breast cancer in females and prostate cancer in men, although cause-and-effect is debatable. Estrogens preferentially bind with receptor cells in breast tissues leading to cell proliferation that can ultimately form tumors (Adeel et al., 2017).

Loos et al. (2018) describe a proposal of the Joint research Centre (JRC) of the European Commission regarding the estrogens 17-alpha-ethinylestradiol (EE2), 17-beta-estradiol (E2), and estrone (E1), their Predicted no-effect concentration (PNECs), and criteria that substances can be taken out of the watch list. None of the three estrogens could be detected above the LOQ. In the case of E1 and E2 the LOQ (0.00013 and 0.00011 μg/l, respectively) was below the specific PNEC (0.0036 and 0.0004 μg/l, respectively). Even the JRC proposal was fulfilled laying down criteria that substances can be taken out of the watch list when ½ LOQ is below or equal to the PNEC. With EE2, a slightly different situation occurred because the PNEC (0.000035 μg/l) is below the LOQ (0.000084 μg/l) in our study, and therefore, no conclusion could be drawn whether the goal is reached or not. In Canada, mean influent concentrations of 17β-estradiol were 15.6 ng/l (range 2.4–26 ng/l), and of estrone 49 ng/l (range 19–78 ng/l); in final effluents, these were reduced to mean concentrations of 1.8 ng/l (range 0.2–14.7 ng/l) and 17 ng/l (range 1–96 ng/l), indicating a wide extent of variability of removal in different sewage treatment plants (Falconer, 2006). An Australian study showed the advantages of tertiary sewage treatment reducing 17β-estradiol levels of up to 19 ng/l by only traditional clarification and to < 7 ng/l by additional use of ozone or UV (Falconer, 2006). In the Paris region, all river samples showed both natural and synthetic estrogens at concentrations of 1.0 to 3.2 ng/l, with 35 to 50% of the estrogenic activity in the form of ethinylestradiol (Cargouët et al., 2003). In German rivers, the steroid hormones were found at 0.2 to 5 ng/l in surface waters and in drinking water at concentrations of 0.1 to 2 ng/l (Kuch and Ballschmiter, 2001). Our findings did not result in any detection of those hormones in Austria, with LODs ranging from 0.023 ng/l (ethinylestradiol) to 0.118 ng/l (estriole).

Brominated flame retardants (BFRs) are widely used in polymers and textiles and applied in construction materials, furniture, and electronic equipment. BFRs with the highest production volume are the polybrominated diphenyl ethers (PBDEs). Because of their persistence and low biodegradation profile, several of the PBDE congeners accumulate in biota and are widely found in the aquatic food chain (Weijs et al., 2015). Their levels in the environment and in humans have increased during the last decades, in contrast, for example, to compounds such as polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT) (Usenko et al., 2016). PBDEs exhibit a great variety of biological effects, depending on the bromine substitution pattern. They are potential endocrine disruptors, based on shared toxicity with the structurally related PCBs, polychlorinated dibenzofurans (PCDFs), and polychlorinated dibenzodioxins (PCDDs) (Vos et al., 2003).

Regarding polybrominated diphenyl ethers (PBDEs) in surface waters, no single value and also not the sum of measured parameter values reached by far the permissible maximum concentration of 0.14 μg/l referred to by Brielmann et al. (2018). They were not even found above the LOQ in rivers or bathing waters. In our measurements of PFAs, the permissible maximum concentration of PFOS in surface waters in Austria (Brielmann et al., 2018), 36 μg/l, was not reached, being several orders of magnitude below. These values also came significantly below the proposed values (in a future amendment of the EU Drinking Water Directive) of 0.1 μg/l for the single substance and of 0.5 μg/l for the sum of PFAS (Briemann et al., 2018), in analogy to the EU pesticides regulation. Lütjohann et al. (2006) found the levels of BFRs of up to more than 130 ng/l in the German surface waters, with a mean of about 30 ng/l. Christale et al. (2013) investigated PBDEs along a river affected by urban and industrial pressures in the UK and detected BDE 209 in most of the sampled sites in concentrations ranging from 17 to 295 ng/l depending on the location. In our study, PBDEs were detected in 3 (15%) drinking water samples, and in 4 (18.2%) groundwater samples, but not in surface water, with a maximum concentration of 0.62 ng/l for BDE 47 in groundwater.

Phthalates are found in a large number of articles. What is common to this group is that the objects are wholly or partly made of soft PVC. Ortho-phthalates comprise a large group of substances presenting both different and similar toxicological effects. Impairment of reproduction, especially in humans, is an effect which has been linked to a number of low molecular weight phthalates. These have been reported for diethylhexyl phthalate (DEHP), the most widely studied phthalate in the low molecular weight phthalate group, but also for dibutyl phthalate (DBP), benzyl butylphthalate (BBP), and diisobutyl phthalate (DIBP) (Kay et al., 2013). One feature these four phthalates have in common is that they produce an antiandrogenic pattern of effects and should, therefore, also be assessed together (Kay et al., 2013). Also, the high molecular weight phthalate diisononyl phthalate (DINP) is suspected of possibly being toxic for reproduction based on an antiandrogenic pattern of effects; however, DINP is not classified as toxic for reproduction (Kay et al., 2013). Diisodecyl phthalate (DIDP), another very common high molecular weight phthalate, is suspected of being toxic for reproduction, but probably via a different type of mechanism. Therefore, DIDP does not contribute to the antiandrogenic effects (KEMI, 2015).

Peijnenburg and Struijs (2006) found phthalates in Dutch fresh water in concentrations of up to 2.35 μg/I and of up to 4.96 μg/I in surface water. In water samples from the Irish Shannon river basin, DEHP was found to occur at levels of 0.77–92.84 μg/l (Jones et al., 2017). We found phthalates in 2 (10%) drinking water samples, in 11 (50%) groundwater samples, in 3 (60 %) bathing water samples, and in 4 (57.1%) river samples. The highest value was with di-(2-ethylhexyl)phthalate (0.37 μg/l, groundwater).

Per- and polyfluoroalkyl substances (PFAS) are a group of man-made chemicals that include perfluorooctanoic acid (PFOA), and perfluorooctane sulfonic acid (PFOS). PFAS have been manufactured and used in a variety of industries around the globe. PFAS often can be found—amongst other—in drinking water, typically localized and associated with specific facilities (e.g., manufacturer, landfill, wastewater treatment, firefighter training facility) (Banzhaf et al., 2017).

It is known from animal studies that short chain PFAS are almost completely absorbed orally and by inhalation, but that skin absorption may be negligible. Both short- and long-chain perfluoroalkyl acids (PFAAs) are considered metabolically inert. The strong C-F bonds exclude any normal degradation pathway. Any functional derivative (precursor) will ultimately be transformed through several steps to the acids (Health Canada, 2016). In animal experiments, the acute toxicity of short-chain PFAS is low. After repeated exposure, large doses of short-chain PFAS may damage the liver and kidneys. In general, PFAS are more toxic to males than females having a higher elimination rate (Health Canada, 2016). In various animal and in vitro studies, PFAS have shown effects on thyroid hormones and decreased their levels. The toxicokinetics and toxicity in humans for short-chain PFAS are mainly investigated for perfluorohexane sulfonic acid (PFHxS), a substance that has rather similar properties as PFOS. There are studies showing associations between PFHxS and effects on lipid metabolism, fertility, thyroid hormones, asthma, and children’s behavior (Kjølholt et al., 2015).

In their recent comprehensive paper on worldwide occurrence and levels of newly-identified perfluoroalkyl and polyfluoroalkyl substances in drinking water, Kaboré et al. (2018) screened, in total, 29 target and 104 suspect-target PFASs in drinking water samples (n = 97) from Canada and other countries (Burkina Faso, Chile, Ivory Coast, France, Japan, Mexico, Norway, and the USA) in 2015–2016. Out of the 29 PFASs quantitatively analyzed, perfluorocarboxylates (PFCAs: C4/14), perfluoroalkane sulfonates (PFSAs: C4, C6, C8), and perfluoroalkyl acid precursors (e.g., 5:3 fluorotelomer carboxylate (5:3 FTCA)) were recurrently detected in drinking water samples in concentrations up to 39 ng/l. Tap water samples from Canada showed noteworthy differences depending on their source; for instance, the sum of the 29 PFAS was significantly greater in samples from the Great Lakes/St. Lawrence river ecosystem than those produced from other sources (14 versus 5.3 ng/l). We found perfluoroalkyl substances, perfluorooctanoic acid and perfluorooctanesulfonic acid, in 5 (25%) drinking water samples, in 8 (36.4%) groundwater samples, in 2 (28.6%) river samples, and in 2 (40%) bathing water samples, with a maximum concentration of 0.0433 μg/l found for perfluorooctanesulfonic acid in drinking water.

Alkylphenols are ultimate breakdown products of alkylphenol polyethoxylate (APEs) that are used in cleaning and industrial processes. The most commonly used APEs in the market are the nonylphenol ethoxylates (NPEs) and octylphenol ethoxylates (OPEs). As a result of their widespread use and their lipophilic nature, these compounds are ubiquitous in the environment and currently of concern because of their toxicity and estrogenic properties (Kovarova et al., 2013). Alkylphenol ethoxylates, including their degradation products alkylphenols octylphenol (OP) and nonylphenol (NP), have long been known to be estrogenic (Dodds and Lawson, 1938). OP and NP have been found to be estrogenic in several in vitro and in vivo systems, where OP is the most potent of the alkylphenols. Critical effects due to OP exposure are changes in sperm morphology in rats and increased length of gestation in pigs. The critical effect of NP exposure is increased kidney weight (Jonsson, 2006). Our sole findings of substances of that group regarded 4-nonylphenol-monoethoxylate (NP1EO) in one river and in two groundwater samples, with a maximum value of 23.4 ng/l (groundwater).

Bisphenol A (BPA) is an estrogen mimic used in the manufacture of plastics, lacquers, and packaging materials. Although initially considered to be a weak environmental estrogen, more recent studies have demonstrated that BPA may be similar in potency to estradiol in stimulating some cellular responses. Moreover, emerging evidence suggests that BPA may influence multiple endocrine-related pathways (Rubin, 2011). BPA has been shown to play a role in the pathogenesis of several endocrine disorders, including female and male infertility, precocious puberty, hormone dependent tumors, such as breast and prostate cancer, and several metabolic disorders, including polycystic ovary syndrome (PCOS) (Konieczna et al., 2015).

Staniszewska et al. (2015) reported concentrations of bisphenol A (BPA), 4-tert-octylphenol (OP) and 4-nonylphenol (NP) in surface and near-bottom water of the Gulf of Gdansk, as well as in flowing rivers in samples taken in the period 2011–2012; BPA in concentrations ranging from < 5.0 to 277.9 ng/l; OP from < 1.0 to 834.5 ng/l, and NP from < 4.0 to 228.6 ng/l. The authors argue these values to be similar to those in other regions of Europe. Lv et al. (2016) detected OP and NP with concentrations ranging from below LOD to 0.49 ng/l and below LOD to 3.27 ng/l in two drinking water works, which take the source water from Taihu Lake in Jiangsu province (China). In these drinking water facilities, bisphenol A was found in concentrations ranging from below LOD up to 17.73 ng/l by Lv et al. (2016).

Besides, the already mentioned values of < 5.0 to 277.9 ng/l (Konieczna et al., 2015), in a German study, bisphenol A was found in surface waters at concentrations of 0.5 to 16 ng/ l, and in drinking water at concentrations of 0.3 to 2 ng/l (Kuch and Ballschmiter, 2001).

We detected BPA in 4 (20%) drinking water samples, in 1 (4.5%) groundwater sample, and in 1 (20%) bathing water sample, with a maximum value of 20.9 ng/l (drinking water). Recent restrictions on the use of bisphenol A (BPA), in varnishes and coatings intended to come into contact with food, have been imposed by Regulation No 2018/213 at the European level (EC, 2018c). By amending Regulation (EU) No 10/2011, this latest regulation tightens or introduces new measures on the uses of BPA as a monomer in the production of plastic materials, such as polycarbonate plastic, and in epoxy resins used in varnishes and coatings. Under REACH, in addition to identifying BPA as a substance of very high concern (SVHC) (ECHA, 2017), a new restriction for BPA present in thermal paper will apply from 2020 (EC, 2016). Stricter threshold values have also been introduced for toys and in the workplace (EC, 2017a), whilst exemptions will need to be requested in the future on its use in medical devices (EC, 2017b). The European Commission, moreover, announced plans to undertake a new strategy in the future to minimize exposure to endocrine disruptors, including from food packaging (EC, 2018b).

Triclosan has been used worldwide as a broad-spectrum antibacterial agent for over 40 years. Increasingly, reports indicate frequent detection and broad exposure to triclosan in the natural environment and the human body. Laboratory studies in various species provide strong evidence for its disrupting effects on the endocrine system, especially affecting reproductive hormones. Multiple modes of action have been suggested, including disrupting hormone metabolism, displacing hormones from hormone receptors and disrupting steroidogenic enzyme activity (Wang and Tian, 2015). Although epidemiological studies on its effects in humans are mostly negative and conflicting, the evidence suggests that triclosan is an EDC (Wang and Tian, 2015).

In the United States (1999–2012), triclosan was detected, most frequently in untreated waters (92% detection frequency; mean ± standard error, 11 270 ± 2925 ng/l; n = 237), but concentrations were significantly reduced in effluent waters (83% detection frequency; 775 ± 311 ng/l; n = 192, α = 0.05) (Perez et al., 2013). Triclosan concentration in effluent-impacted environmental waters (62% detection frequency; 130 ± 17 ng/l; n = 228) was not significantly reduced from effluent waters but was significantly greater than triclosan in environmental waters not classified as effluent impacted (11% detection frequency; 13 ± 3 ng/l; n = 1195). In finished drinking water, triclosan was largely undetected (1% detection frequency; 4 ± 2 ng/l n = 453), suggesting that for the United States, drinking water is not an appreciable source of triclosan exposure. In post treatment water, average triclosan concentrations were below part‐per‐billion levels (Perez et al., 2013). In our samples, triclosan was not found above an LOQ of 3.31 ng/l.

There is growing evidence that EDCs can disrupt thyroid homeostasis, even though the most important knowledge on this topic derives from animal studies, while clinical studies are still few and controversial (Gore et al., 2015). All the substance groups investigated are suspected of having anti-thyroid effects in some way, with specific attention to fetal and early childhood exposure, potentially affecting various brain processes such as neurogenesis, neural differentiation and migration, as well as neural connectivity, leading to morphological brain abnormalities (e.g., cortical thinning) consistent with neurodevelopmental impairment. Ghassabian and Trasande (2018) describe the following in a review: a) Experimental studies have shown that PBDEs, at environmentally relevant doses, bind to receptors, inhibit binding of triiodothyronine (T3) to thyroid hormone receptors (TRs) and suppress T3 actions; b) In vivo, bisphenol A can impair thyroid hormone action by antagonizing T3-induced TR activation (TRα1 and TRβ1) and by suppressing its transcriptional activity in a dose-dependent manner; c) Triclosan affects thyroid hormone-dependent metamorphosis in animals; d) Urinary phthalates are cross-sectionally associated with lower free thyroxine (fT4) and higher thyroid stimulating hormone (TSH); e) PFASs interfere with binding of thyroid hormone to transthyretin and upregulate deiodinase in the thyroid gland. Thyroid disruption is of particular interest because several EDCs interfere with thyroid function in a sex-specific manner, which might explain the sexual dimorphism in the brain effect of EDCs (de Cock et al., 2014).

US children’s centers are exploring associations between brain development and environmental toxicants such as, amongst others: phthalates, bisphenol A, and polybrominated diphenyl ethers (US EPA, 2017). Prenatal exposure has shown a relationship between gestational levels of BPA or phthalates to adverse cognitive and behavioral outcomes, demonstrating links to attention deficit hyperactivity disorder, reduced intelligence quotient, lessened self-regulatory capacities, anxiety, depression, lower memory function and structural changes to the brain (Palanza et al., 2016; US EPA, 2017).