Registration Dossier

Ecotoxicological information

Endpoint summary

Currently viewing:

Administrative data

Description of key information

Additional information

Overview of aquatic toxicity data Short-term and long-term ecotoxicological data on the effects of hexavalent chromium compounds are available for a wide variety of organisms (freshwater and marine fish, invertebrates, algae, plants, amphibians), lifestages (juveniles, adults, fry, larvae, tadpoles, eggs, etc.), endpoints (LC50s, EC50s, NOECs, LOECs based on mortality, reproduction, hatching, etc.), and test conditions. The results are expressed as the concentrations of chromium (VI), for ease of comparison among the five hexavalent compounds. In general, the majority of ecotoxicological information is available for potassium dichromate, as it is a reference toxicant. Few results are avaiable for chromic acid and ammonium dichromate. The results indicate that the acute toxicity of chromium (VI) is dependent on a number of factors, including pH, water hardness, salinity and temperature. In general, chromium (VI) toxicity is increased with decreased pH (i.e. 8.0 to 6.0), increased temperature (i.e. 15 to 25oC) and decreased water hardness (>100 to <100 mg/l as CaCO3) or salinity (<2%). The values in parenthesis are general values for fish and aquatic invertebrates and will vary according to individual species¿ optimum environmental requirements. It should be noted that there are also studies which show little change in toxicity with changes in water properties. From the available data there does not appear to be any difference among the sensitivity of organisms to the nature of the chromium (VI) ion in short-term tests. The acute toxicity values when plotted as a sensitivity distribution of acute endpoints show that aquatic invertebrates are the most sensitive test species to chromium (VI),making up the vast majority of the lower half of the ranking. There are a large number of invertebrate results, so there are species present of a much lower sensitivity. The smaller number of algal results also appears in the upper half. A comparison between freshwater and saltwater organisms shows that the former appear to be more sensitive. Decreasing salinity appears to lead to increased toxicity; this can be seen in the results of particular studies, for example, Bryant et al. (1984) with Corophium volutator and Macoma balthica, and Persoone et al. (1989) with Branchionus plicatilis. Where saltwater organisms have been tested in water of low salinity (<2%), their sensitivity appears to become comparable with that of freshwater organisms. A more detailed breakdown of invertebrate species for freshwater organisms shows that the most sensitive group of invertebrates are cladocerans, such as Ceriodaphnia dubia and Daphnia magna. The single amphipod result lies within the range of the cladocerans, as do two of the three algal species. All of the values for fish are higher than the cladoceran and amphipod values. The summary tables of freshwater chronic toxicity values contain several values for some species, covering a range of endpoints. These have been combined into the values given in the table below. For species where more than one value was available for an end point, the geometric means of the values for survival/mortality, reproduction, and growth/development were calculated to produce one value per endpoint. Then for all species the lowest value between these endpoints was selected as the NOEC for the species. The table includes notes on data selection where there was a choice.

Major taxon

Species

NOEC (mg Cr/L)

Notes

Algae

Microcystis aeruginosa

0.35

Chlorella pyrenoidosa

0.1

Chlorella spp. (wild)

0.1

Scenedesmus pannonicus

0.11

Selenastrum capricornutum

0.033

Geometric mean of EC10

Macrophytes

Lemna gibba

0.1

Lemna minor

0.11

Spirodela polyrhiza

0.1

Spirodela punctata

0.5

Crustaceans

Ceriodaphnia dubia

0.0047

Reproduction

Daphnia carinata

0.05

Daphnia magna

0.019

Geometric mean of reproduction values

Coelenterates

Hydra littoralis

0.035

Hydra oligactis

1.1

Insect

Culex pipiens

1.1

Survival/growth NOEC

Mollusc

Lymnaea stagnalis

0.11

reproduction

Fish

Catastomus commersoni

0.29

Longer growth value

Esox lucius

0.538

Ictalurus punctatus

0.15

30-d growth NOEC

Oncorhynchus mykiss

0.07

Geometric mean of growth NOECs

Oryzias latipes

3.5

Survival NOEC

Pimephales promelas

0.68

Geometric mean of growth NOECs

Poecilia reticulate

3.5

Growth/mortality NOEC

Salvelinus fontinalis

0.01

Growth NOEC

Salvelinus namaycush

0.105

Growth NOEC

Amphibian

Xenopus laevis

0.35

Mortality NOEC

The long-term studies available do not appear to show any clear dependence of toxicity on the properties of the water. There are indications that toxicity may be higher in lower hardness waters, but there are few if any studies which allow the comparison to be made for the same species at different levels of hardness, or other properties. The range of water hardness values in the studies included in the table above, where these were reported, is 24 - 250 mg/l as CaCO3. Most of the reported values are below 50 mg/l. The pH in the tests was generally between 7.5 and 8.5. Although relationships between hardness and toxicity have been described for divalent metal cations, the fact that the chromium species here are oxoanions means that their toxicity may be less influenced by water properties. As no relationships can be established, the toxicity data will be treated together. It should be noted that the calculated concentrations do depend on the environmental properties. Assessment factor approach According to the standard assessment factor approach, the PNEC is derived from the lowest NOEC available. The lowest NOEC included in the preceding sections is 4.7 µg/l, for reproduction of the cladoceran Ceriodaphnia dubia. As there is a large amount of long-term effect data on a wide range of aquatic organisms, an assessment factor of 10 is used, giving a PNEC by this method of 0.47 µg/l. Statistical extrapolation approach According to the TGD, the effects assessment can also be supported by a statistical extrapolation method if the data base is sufficient for its application. A workshop on the use of statistical extrapolation for the derivation of PNEC values in case of data-rich substances was held in London in January 2001 in the framework of the EU Existing Substances programme. This workshop was specifically aimed at the use of statistical extrapolation for the derivation of PNEC values for the metals zinc, cadmium and hexavalent chromium, since for these metals large chronic databases are available. The workshop recommended the inclusion of statistical extrapolation in the derivation of PNEC values for these metals, provided the chronic database meets certain requirements (EU, 2001). The data set for chromium is discussed below in relation to these requirements. There is a considerable amount of ecotoxicological information available on the toxicity of the five hexavalent chromium compounds to aquatic organisms. There are 28 NOEC (or derived NOEC) values available for calculating a HC5 for chromium (VI) from a wide range of aquatic taxa including: fish, crustacea, algae, aquatic plants, insects, molluscs, amphibians, and coelenterates. These values can be matched against the criteria used by the US EPA which were adopted at the workshop, with the addition of algae and aquatic plants. This is done in the table below. Only one species is included against each criterion, but the data set contains more examples.

Criterion

Species

The family Salmonidae in the class Osteichthyes

Oncorhychus mykiss

A second family in the class Osteichthyes, preferably a commercially

or recreationally important warm water species (e.g. bluegill, channel

catfish, etc.)

Pimephales promelas

A third family in the phylum Chordata (may be in the class Osteichthyes or may be an amphibian, etc.)

Esox lucius

A planktonic crustacean (e.g. cladoceran, copepod, etc.)

Ceriodaphnia dubia

A benthic crustacean (e.g. ostracod, isopod, amphipod, crayfish)

An insect (e.g. mayfly, dragonfly, damselfly, stonefly, caddisfly,mosquito, midge, etc.)

Culex pipiens

A family in a phylum other than Arthropoda or Chordata (e.g.Rotifera, Annelida, Mollusca, etc.)

Hydra littoralis

A family in any order of insect or any phylum not already represented

Xenopus laevis

Algae

Selenastrum capricornutum

Aquatic plant

Lemna gibba

The one gap in the data set is for a benthic crustacean. One amphipod (Crangonyx pseudogracilis) is present in the selected data set for acute values, and is less sensitive than the cladocerans included. There are other non-selected values in the overall acute data set which would indicate similar or lower sensitivity to cladocerans. There are also several other representatives for some of the groups indicated in the criteria above. Hence the absence of this specific group is not considered to make the data set unrepresentative. The number of available NOEC values (28) is significantly more than the minimum requirements discussed at the workshop. The tests from which the values come cover a range of chronic endpoints, including growth, reproduction and survival, and cover sensitive life stages for longer lived-organisms (e.g. fish) and multiple life cycles for shorter-lived species (e.g. cladocerans). Multiple data values for the same species and endpoint have been combined as agreed at the workshop (see above). A further consideration for the use of the method is whether the data fit to the expected distribution. The data set has been tested against a log-normal distribution, as preferred at the workshop. based on the observed and expected frequencies and cumulative frequencies a Kolmogorov-Smirnov test does not reject the null hypothesis, that the data come from a lognormal distribution, at the 1%, 5% or 10% levels. It is clear from the plots that there is a preponderance of values towards the centre of the distribution, but with values also at some distance from it, giving relatively long tails. Overall the data set is considered suitable for use in the extrapolation method. The lower 5% value from the species distribution (HC5) has been calculated according to the following equation for a log-normal distribution (Wagner and Lokke, 1991) as preferred at the workshop. HC5 = 10 (xm-km.sm) where: HC5 = lower 5% limit of species distribution m = the number of test species (here 26) x = sample mean of log NOEC data for m species (here 2.19) k = the one-sided extrapolation constant for a normal distribution (here 1.67) s = the sample standard deviation of log NOEC values for m species (here 0.70) The resulting value for the 50% confidence level in the HC5 (HC5-50%) is 10.2 µg/l. The value for the 95% confidence level (HC5-95%) is 3.8 µg/l. Having obtained these results the application of a possible assessment factor to derive the PNEC value has to be considered. The data set used in the extrapolation covers a wide range of aquatic species and a range of chronic endpoints. It includes the types of organism indicated to be the most sensitive in acute tests, and there do not appear to be any groups of sensitive organisms which are missing from the data set. The organisms cover a range of trophic levels and feeding strategies, including primary producers, herbivores, fish which consume algae and invertebrates, fish which consume other fish, and detritivores. Against these points, there are a relatively large number of results for fish (although they cover different types) and only one each for insects or molluscs. There are also no results from mesocosm or field studies to compare to the derived values. There are two values included in the data set which lie below the HC5-50% value, one for the cladoceran Ceriodaphnia dubia and the other for the fish Salvelinus fontinalis. In the case of Ceriodaphnia dubia, the NOEC for reproduction was 4.7 µg/l; from the same report the NOEC for survival was 8.4 µg/l. These values come from a ring test and are derived from 18 individual results. In the same study the 50% effect concentration for survival and reproduction over 7 days was 14 µg/l, indicating a steep dose-response. The NOEC for Salvelinus fontinalis is 10 µg/l, which is virtually the same as the HC5 -50% value. The considerations above suggest that a small assessment factor could be applied to the extrapolated value to give a more protective PNEC. The choice of assessment factor to be used with the HC5 makes little or no difference to the overall result of the assessment, but a factor of 3 was accepted during Technical Meeting discussions as a reasonable compromise between member states that expressed a view. This gives a PNEC of 3.4 µg/l. The HC5s calculated here for chromium (VI) are similar to the HC5s calculated by Emans et al. (1993) for total chromium of 4.9 µg/l and 4.6 µg/l, based on the methods of Aldenberg and Slob (1993) and Wagner and Lokke (1991), respectively. The HC5-50% value calculated here is virtually the same as that reported by Okkerman et al. (1991) for potassium dichromate, according to the method of van Straalen and Denneman (1989) (they reported a value of 29 µg/l for potassium dichromate, the value for chromium would be 10 µg/l). Cromentuijn et al. (1997) calculated a value of 6.4 µg/l for freshwater organisms by the Aldenberg and Slob method, and 8.5 µg/l for a mixed freshwater and saltwater data set. In saltwater, chromium (VI) would be expected to be less toxic than indicated by these values, except perhaps at very low salinities. Since chromium (VI) is converted to chromium (III) under some conditions in the environment, the possible effects of chromium (III) should also be considered in the assessment. From the available data, it can be seen that chromium (III) appears to be less toxic than chromium (VI) in waters of medium hardness (>50 mg CaCO3). In lower hardness waters the acute toxicity increases; there are also indications that NOEC values decrease with decreasing hardness. There are insufficient data to carry out an HC5 calculation for chromium (III). From the freshwater data, long-term NOEC values are 0.05 mg/l for fish and 0.047 mg/l for invertebrates, and >2 mg/l for algae (although an EC50 of 0.32 mg/l is reported for another species). The fish and invertebrate values relate to hardness levels of 26 and 52 mg/l respectively. Applying an assessment factor of 10 to the lowest available NOEC gives a tentative PNEC for chromium (III) of 4.7 µg/l for soft water. This is similar to that derived for chromium (VI) above, but the two values are not directly comparable as they are based on very different data sets. However, this may indicate that in low hardness waters the two forms may not be very different in effect. The NOEC from the same invertebrate study at a hardness of 100 mg/l was 0.129 mg/l, which would give a ¿PNEC¿ of 13 µg/l. The data indicate that chromium (III) may have reduced toxicity at greater hardness levels, but as with chromium (VI) the evidence is limited (these comments relate to chronic toxicity). The PNEC is at the lower end of the range of published criteria/standards for the protection of aquatic life. For example, the UK Environmental Quality Standard for total chromium in freshwater ranges from 5 to 50 µg/l (dependent on water hardness) and in saltwater it is 15 µg/l. It should also be noted that the PNEC for chromium (III) refers to the dissolved water concentration. In laboratory tests, water soluble forms of chromium (III) have generally been used. However, in the environment, chromium (VI) is likely to be reduced to forms of chromium (III) with limited water solubility, which will be associated mainly with the particulate (sediment and suspended matter) phases of the water compartment. In summary, the PNEC values for the surface water compartment are 3.4 µg/l for chromium (VI) and 4.7 µg/l for chromium (III).