Diantimony trioxide

Short-term toxicity to aquatic invertebrates

Description of key information

The lowest valid value for acute toxicity to freshwater invertebrates is 1.77 mg Sb/L for Chlorohydra viridissimus (TAI, 1990).

There are no valid acute studies with marine invertebrates.

Additional information

Three studies of acute toxicity to freshwater invertebrates that are considered valid are available (Kimball, 1978; TAI, 1990; Brooke et al., 1986). There is also an additional study by Brooke et al. (1986) which, although not considered reliable, does support the finding of Anon (1990) that the coelenterate hydra is the most sensitive invertebrate tested to date.

In the study by Kimball (1978), <1 d old daphnids were exposed in a static test design to trivalent antimony (SbCl₃) for 2 or 4 days. This study was performed with six concentrations (range: 1.65 – 44.15 mg Sb/L) and a control, using four replicates with ten neonates for each concentration. Mortality was the endpoint. In the 2-day exposure regime, the daphnids were exposed to antimony (SbCl₃) either with or without feeding. The resulting LC50 values were 12.2 and 18.8 mg Sb/L, respectively. In the 4-day exposure regime, the antimony exposure was only performed together with feeding, and the resulting LC50 was 12.1 mg Sb/L.

In the study by TAI (1990) two hydra species (Hydra oligogactis, and Chlorohydra viridissima), one snail (Physa heterostropha), one midge (Chironomus tentans) and an amphipod (Hyalella azteca) were tested in a static test design with trivalent antimony (SbCl₃) for 4 days. All tests were performed with five concentrations and a control, using duplicate replicates with five individuals for each test chamber. The measured concentrations used (nominal concentration within brackets) were 8.5 (15.6), 15.4 (31.25), 19.8 (62.6), 24.6 (125), and 21.22 (250) mg Sb/L, and a control with 0 mg Sb/L (measured concentration) for Physa heterostropha,Chironomus tentans, and Hyalella azteca. The measured concentrations used (nominal concentration within brackets) for the two hydra species Hydra oligogactis and Chlorohydra viridissima were 1.18 (0.5), 1.52 (1.0), 2.56 (2.5), 4.48 (5.0), and 5.42 (6.25), and a control with 0 mg Sb/L (measured concentration). Using mortality as the endpoint the resulting LC50s were 1.95, 1.77, 14.2, 4.1, and 21.6 mg Sb/L for Hydra oligogactis, Chlorohydra viridissima, Physa heterostropha and Chironomus tentans respectively. Therefore, the hydras were the most sensitive invertebrates.

This observation is further supported by the study by Brooke et al. (1986), in which hydra were also found to be the most sensitive of several tested invertebrates out of Hydrasp., amphipods (Gammarus pseudolimnaeus), annelids (Lumbriculus variegatus), and caddisflies (Pycnopsyche sp.). Adult hydroids (Hydra sp.) were exposed in a static test design to trivalent antimony (SbCl₃) for 4 days, and an EC50 (tentacles clubbed and/or shortened body column and tentacles) was determined. These tests were performed in quadruplicate with five concentrations (range: 0.3 - 3.3 mg Sb/L) and a control, with each replicate consisting of ten hydroids. The resulting EC50 (24 h) was 2 mg Sb/L (1.8 - 2.2 = 95% confidence interval), the EC50 (48 h) was 1.0 mg Sb/L (confidence limits not reliable, according to authors), and the EC50 (96 h) was 0.5 mg Sb/L (0.5 - 0.6 = 95% confidence interval). The tests on amphipods (Gammarus pseudolimnaeus), annelids (Lumbriculus variegatus), and caddisflies (Pycnopsyche sp.)were performed in duplicate with two concentrations (11.4 ± 3.9 and 25.7 ± 2.2 mg Sb/L) and a control, with each replicate consisting of ten individuals. The EC50 calculations for these three species all resulted in “greater than” values (> 25.7 mg Sb/L). However, the result of the hydra study is not considered reliable but only indicative, since the endpoint used (tentacles clubbed and/or shortened body column and tentacles) is subjective, and no information is provided on whether a dose-response relationship existed.

The hydra results from the study by Brooke et al. (1986) are only slightly lower then those reported by TAI (1990), i.e. an EC50 (96 h) of 0.5 mg Sb/L (Hydra sp.), compared to 1.77 mg Sb/L (Chlorohydra viridissimus) and 1.95 mg Sb/L (Hydra oligactis). This slight difference may be due to several reasons such as interlaboratory differences, different sensitivities between the different hydra species tested, or the fact that the criteria used for defining the endpoints differ in sensitivity. The criteria for hydra effects in the TAI study were the beginning of the break down of tissue integrity and an associated bacterial growth enveloping the animals, while in the study by Brooke et al. (1986) it was clubbed tentacles and/or shortened body column and tentacles.

Thus, the most sensitive of the aquatic invertebrates is the hydra, and the lowest valid EC50 (96 h) for acute toxicity is 1.77 mg Sb/L.

The reasons why the study by Doe et al.(1987) on the acute toxicity of Daphnia magna is considered unreliable, even though exposure concentrations were measured, is that there is no information presented on (i) the number of concentrations and which concentrations were used, (ii) dose-response curves (no raw data are available), (iii) the number of replicates (if any), and (iv) what statistics have been used to calculate the LC50 values.

Borgmann et al. (2005) reported a seven-day acute LC50 value of 0.687 mg Sb/L for Hyalella azteca, which is not considered to be reliable. Because of the objective of the study (large-scale screening of metal toxicity for categorization of substances on the Canadian Domestic Substances List) several modifications have been made to the standard experimental design. The main reason why the result from this study cannot be considered reliable is because antimony was clearly not the only toxicity-inducing factor in the test medium. This study used metal standards for toxicity testing. For antimony, a metal standard containing 20% HCl was used. The acid in the metal standards was neutralized by adding a solution of 1M NaHCO₃and 1M KOH in a 19:1 ratio. Along with a control treatment (containing normal test medium), an acid control was used (containing acid and neutralizing solution additions equal to the amount added in the tests with acidified metal standards). For metal standards supplied in 20% HCl, survival in the acid controls for the 1,000 µg/L treatment dropped to 32%, indicating that the organisms were adversely affected by the blank test medium. Therefore, toxicity in the metal treatments was most likely overestimated in this study. The LC50 values derived for antimony in the tests using the metal standards is not considered reliable. Toxicity tests were also conducted using sodium antimonate (NaSbO₃). However, no toxicity was observed at the highest exposure concentrations tested (1 mg Sb/L nominal, 0.197 mg Sb/L measured).

Fjallborg and Dave (2004) spiked sewage sludge with SbCl₅and after 60 days of equilibration grew radish, oats or lettuce. After 14 days’ cultivation the toxicity of the elutriate to Daphnia magna was tested. The authors report an EC50 based on the nominal concentrations spiked into the sewage sludge, for which a dose response was observed. For the elutriate after radish cultivation the EC50 was 22 mg Sb/kg, and after oat cultivation the EC50 was 5 mg Sb/kg. This study included analytical monitoring of the elutriate. These measured concentrations are only reported in graphical format. When the effects data are compared with the measured concentrations in the elutriate a dose reponse is not observed. For this reason, this non-standard study is not considered reliable for use in the assessment.

This is the only study to have tested the toxicity of an Sb(V) compound, and the study is not considered to be valid because no dose response in relation to antimony concentrations was observed. There is therefore no information on the toxicity of exposures which contain a very high proportion of Sb(V), as only relatively limited oxidation of Sb(III) to Sb(V) would be expected within the time frame of an acute toxicity test (or the renewal period of a semi-static chronic test). Sb(V) is expected to be the dominant species in oxygenated natural waters, although the rate of oxidation of Sb(III) to Sb(V) is very slow, with a half-life in laboratory solutions expected to be in the order of months (I2A 2010). PNEC values for antimony are expressed in terms of dissolved antimony concentrations.

USEPA (1988) report results from Spehar (1987) of an acute study with Ceriodaphnia dubia with measured exposure values. Although we have not been able to obtain a copy of Spehar (1987) the value reported is higher than that from TAI (1990) so would not change the assessment.