Registration Dossier

Ecotoxicological information

Endpoint summary

Administrative data

Description of key information

Additional information

For Bisphenol A, there is a rich database of both acute and chronic aquatic effect studies with vertebrates, invertebrates, and aquatic plants.

Short term toxicity to fish

There are four key studies which report of acute toxicity to fish for fresh or marine water. The lowest LC50 value in each compartment was used in the risk assessment to provide a conservative assessment. Alexander et al., 1985, (published as Alexander et al., 1988) determined an LC50 of 4.6 mg/L in an OECD 203 study with Pimephales promelas (fresh water). Springborn Bionomics, 1985a (published as Alexander et al., 1988) reported an LC50 of 9.4 mg/L in an OECD 203 study with Menidia menidia (marine water).

Long term toxicity to fish

There are four key studies (all reliability 1, reliable without restriction) available which address the long-term toxicity of Bisphenol A to fish. The most sensitive endpoint was egg hatchability in F2 of a 444 day-multigeneration study with fathead minnow (Pimephales promelas) according to EPA OPP 72-5 method which  reported a NOEC of 16 µg/L (Caunter et al., 2000, published in Staples et al., 2011, Ecotox Environ Safety 74:1548-1557). Results reported from a fish early life stage study (OECD 210) with P. promelas led to significantly higher NOECs based on hatchability, survival, and growth at 640 µg/L in all cases (Caunter et al., 1999, published as Staples et al., 2011, Ecotox Environ Safety 74:1548-1557). With the same test species there were no statistically significant effects in males or females with respect to growth, gonad weight, gonadosomatic index (GSI), or reproduction variables (e.g., number of eggs and spawns, hatchability) at any Bisphenol A concentration in the third key study by Rhodes et al.  (2008, published in Mihaich et al., 2012, Environ Toxicol Chem, 31: 2525-2535) which was conducted in accordance with EPA OPP 72-5 method. However, Rhodes et al., 2008, reported of reduced survival with male fish at 640 µg/L (NOEC = 160 µg/L). For marine species, a NOEC for reproductive success (egg/female/day) was observed in a key study investigating the life-cycle exposure to sheepshead minnow (Cyprinodon variegatus) at 66 µg /L and according to EPA OPPTS 850.1500 method (York, 2010, published in Mihaich et al., 2018, Environ Toxicol Chem, 37:398-410).

Short-term toxicity to aquatic invertebrates

There are two key studies which investigated acute toxicity of Bisphenol A towards freshwater and marine water invertebrates. For freshwater a 48 h-EC50 of 10.2 mg/L was reported (Daphnia magna; Alexander et al., 1985, published in Alexander et al., 1988). Tests on acute toxicity of Bisphenol A towards marine water invertebrates resulted in a 96-hr LC50 of 1.1 mg/L (Americamysis bahia, previous name: Mysidopsis bahia; Springborn Bionomics, 1985b; published in Alexander et al., 1988).

Long-term toxicity to aquatic invertebrates

There are four key studies available which are reliable without restriction (Klimisch 1) and which address the long-term toxicity of Bisphenol A to aquatic invertebrates. These include three studies on freshwater species: a Daphnia magna reproduction study according to OECD 211 revealing a NOEC of 3.16 mg/L (Caspers, 1998), a rotifer life-cycle study which followed the ASTM Guideline E1440-91 and report a NOEC of 1.8 mg/L (Springborn Smithers, 2006a; publishes in Mihaich et al., 2009), and a 328-d reproduction study with Marisa cornuarietis snail with a NOEC of 0.025 mg/L (Warbritton et al., 2007a; published as Forbes et al. 2008). For this latter study with M. cornuarietis there was no guideline available at this point in time. However, this study was conducted as a follow up of the Oehlmann et al., 2006, study with M. cornuarietis which had severe short-comings. The test setup of Warbritton et al., 2007a, in contrast was coordinated with the rapporteur UK of the European Risk Assessment (2003, 2008, 2010) and was performed in accordance with state of science and technology. For marine water species, there is one key study, namely a life-cycle study according to EPA OPPTS 850.1350 test method with Americamysis bahia which determined a NOEC of 0.17 mg/L (Lee, 2010; published in Mihaich et al. 2018).

Toxicity to aquatic algae and cyanobacteria

There are two key studies for exposure of algae to Bisphenol A (both rated Klimisch 1, i.e., reliable without restriction) for both compartments, freshwater and marine water. The key study by Alexander et al. (1985; published in Alexander 1988) which was conducted in accordance with EPA-600/9-78-018 test method reported the 96-h EC50 value of 2.73 mg/L for the freshwater algae Pseudokirchneriella subcapitata based on cell count. Furthermore, an EC10 of 1.36 mg/L was derived in this study. The other key study conducted by Springborn Bionomics (1985c; published as Alexander et al., 1988) was performed following test method EPA 560/6-82-002 and reported a 96-h EC50 of 1.0 mg/L with the marine water diatom Skeletonema costatum based on cell count using non-linear interpolation. These original data have been re-analysed by the UK rapporteur in the EU Risk Assessment update in 2008 using probit analysis which is in accordance with the OECD Guideline. The resulting EC50 is 1.1 mg/L (cell count), and 1.4 mg/L (chlorophyll content). In addition, the UK rapporteur calculated a 96-h EC10 of 0.4 mg/L (cell count; most sensitive endpoint).

Toxicity to aquatic plants other than algae

There is one key study on aquatic plants other than algae which was rated as Klimisch 1, reliable without restrictions (Putt, 2003; published in Mihaich et al., 2009). This study followed test method OECD 221 and exposed Lemna gibba. The EC50 was determined to be 20 mg/L (frond density) and the NOEC was reported as 7.8 mg/L (frond density, growth and biomass).

Toxicity to microorganisms

Two studies are available which were rated with Klimisch 2 (reliable, with restrictions) were used in a weight-of-evidence (WoE) approach to derive PNEC STP. Fabig (1988) reported an 18 h-EC10 of > 320 mg/L with Pseudomonas putida. This study is not the same as an activated sludge respiration test according to OECD 209 but there is indication of no toxicity towards P. putida as a microorganism. Stone and Watkinson (1983) as well is not an OECD 209 study but investigated biodegradation with an bacteria growth inhibition test with Pseudomonas fluorescens. The IC50 was reported as 54.4 mg/L. Both studies were used in a WoE approach. The REACH guidance document specifies that studies with P. fluorescens are of limited relevance for STP as this species uses glucose as substrate. However, a growth inhibition test with P. putida may be used for PNEC STP derivation. The study of Fabig (1988) was conducted in accordance with DIN 38412-8 while the recent REACH guidance document R.10 asks for ISO-10712 (1996). Despite the fact that it is not totally certain if the DIN translated directly into the ISO the study is considered sufficiently reliable to be used in a WoE approach. Thus, the EC10 > 320 mg/L was used for PNEC derivation. This approach was also taken in the European Risk Assessment report (2010).

Toxicity to other vertebrates

There is one key study for Bisphenol A which was evaluated to be Klimisch 1, reliable without restriction (Pickford et al., 2000; published as Pickford et al., 2003). In this non-guideline study which lasted 90 days, the NOEC for Xenopus laevis based on larval survival, adult growth and sex ratio was determined to be 500 µg/L, the highest dose of Bisphenol A tested. While there was no effect on any of the endpoints it shall specifically be mentioned that there was no impact on the frog sex ratio at all treatments or control, while in the positive control groupthere was a significant feminisation observed. Thus, the NOEC for all endpoints was determined to be > 500 µg/L.

Sediment toxicity

Six fully valid freshwater sediment studies exist with exposure to Bisphenol A, although in two of them the Bisphenol A concentration was measured in the water, not the sediment. In a 10-day Corophium volutator mortality and sublethal effect study with Bisphenol A-dosed sediment, the 10-day EC50 values based on total adverse effects in Bisphenol A/acetone spiked or directly spiked sediments were 31 and 36 mg/kg, respectively (Whale et al., 1999). When the endpoints of the toxicity tests are based on predicted interstitial water concentrations the 10-day LC50 and 10-day EC50 values are between 1.4 and 1.5 and 1.1 and 1.3 mg/L, respectively. In a 28-day Lumbriculus variegatus study, the No-Observed-Effect Concentration (NOEC) for both survival and biomass was determined to be 22 mg a.i./kg. The Lowest-Observed-Effect Concentration (LOEC) for both endpoints was determined to be 57 mg a.i./kg. Since no concentration tested resulted in ≥ 50% reduction in reproduction as number of surviving oligochaetes recovered or biomass, the EC50 value was empirically estimated to be > 57 mg a.i./kg, the highest mean measured concentration tested (Picard, 2010a). In a 28-day chronic Leptocheirus plumulosus study, reproduction was determined to be the most sensitive endpoint. The LOEC and NOEC for amphipod reproduction were determined to be >32 and 32 mg a.i./kg, respectively. The 28 day EC50 value for reproduction was determined to be 38 mg a.i./kg, with 95% confidence intervals of 11 to 52 mg a.i./kg (Picard, 2010b). A 28-day chronic freshwater midge study was performed exposing Chironomus riparius to measured Bisphenol A concentrations in artificial sediment of 29, 54, 110, 210 and 490 mg a.i./kg dry weight. At test termination (day 28), the mean percent emergence based on the mean measured concentrations was established to be 110 mg a.i./kg. The 28 day EC50 based on midge emergence was determined to be 150 mg a.i./kg dry weight, with 95% confidence intervals of 140 to 160 mg a.i./kg dry weight. The NOEC for emergence was established to be 210 mg a.i./kg. Since no test concentration resulted in a significant reduction of development rate compared to the control, the LOEC and the 28-day EC50 were both empirically estimated to be > 210 mg a.i./kg (Picard, 2010c). In a 42-day chronic Hyalella azteca study, the 42-day LC50 value was determined to be 0.78 mg/L and the LOEC was determined to be 1.1 mg/L. The NOEC was determined to be 0.49 mg/L (Springborn Smithers, 2006b; Mihaich et al., 2009). In a 96-hour midge (Chironomus tentans) acute study, the 96-hour LC50 value was determined to be 2.7 mg/L while the NOEC was determined to be 1.4 mg/L (Springborn Smithers, 2005; Mihaich et al., 2009).

Overall conclusion

Due to a large and robust dataset of valid and reliable studies the acute and chronic effects of Bisphenol A in the aquatic environment are well characterised.

Endocrine disruption

The registrants performed an assessment on endocrine disruption of Bisphenol A which was added in IUCLID in chapter 13. The executive summary is provided below.

To determine potential endocrine activity in the environment, a weight of evidence (WoE) evaluation was conducted for Bisphenol A. This WoE evaluation included the integration of data from high throughput screens, regulatory guideline studies, and the published literature. The approach for evaluating the information from both in vitro and in vivo studies was based on the OECD conceptual framework for testing and assessment of potential endocrine-disrupting chemicals (OECD, 2012) and consisted of a systematic evaluation of data, progressing from simple to complex across multiple levels of biological organization. In addition, hypothesis testing procedures developed by Borgert et al. (2011, 2014) were used to evaluate responses of Bisphenol A relevant for estrogen, androgen, and thyroid pathways, as well as impacts on steroidogenesis.

Bisphenol A is readily biodegradable, not bioaccumulative, and shows moderate acute and chronic toxicity to aquatic and terrestrial organisms. High throughput bioactivity screens (e.g. ToxCast™) show that while Bisphenol A has activity in estrogen receptor (ER) assays, it also has activity unrelated to estrogenic endpoints within a similar concentration range. In vitro test results show that Bisphenol A has weak estrogenic activity in vitro, but ER binding activities for Bisphenol A are several orders of magnitude below that of natural hormones. Bisphenol A is a very data-rich substance, with chronic no observed effect concentrations (NOECs) available for major taxonomic groups. Most of these studies include both mechanistic and apical endpoints, with many apical endpoints focused on population-relevant parameters of survival, growth, development, and reproduction. No observed effect concentration values for Bisphenol A that pertain to survival, growth , and reproduction among a large database of chronic and multigenerational toxicity studies in fish all occur at similar concentration ranges, indicating that the weak estrogenic activity of Bisphenol A is not necessarily the critical effect or lead mode of action associated with adverse effects. By evaluating the acute to chronic ratio (ACR) developed from this large database of studies, Bisphenol A does not appear to provoke chronic reactions in organisms that are far more sensitive than acute, toxic effects. A large ACR would be expected with compounds that target specific hormonal pathways. Based on the evidence of all available data, there is no scientifically sound basis for characterizing Bisphenol A as an endocrine disruptor for environmental species.