Antimony

Short-term toxicity to fish

Description of key information

The lowest valid value for acute toxicity to freshwater fish is 14.4 mg Sb/L for Pimephales promelas (Brooke et al. 1986).

The lowest valid value for acute toxicity to marine fish is 6.9 mg Sb/L for Pargus major (Takayanagi, 2001).

Additional information

Freshwater

Three acute toxicity studies with freshwater fish species that are considered valid are available (Kimball, 1978; Brooke et al., 1986; TAI, 1990).

In the study by Kimball (1978), 8 week-old juvenile fathead minnow (Pimephales promelas) were exposed to trivalent antimony (SbCl₃) in a flow-through system, performed in duplicates with six concentrations (range: 1.0 - 27.6 mg Sb/L) and a control, with each group comprising 10 fish. The tests were performed with hard well water. The resulting LC₅₀s for the 4 and 8 d exposures were 21.9 and 20.2 mg Sb/L, respectively.

In the study by Brooke et al. (1986) juvenile rainbow trout (Oncorhynchus mykiss) were exposed in a static test design to trivalent antimony (SbCl₃) for 4 days, and the LC₅₀ was determined. The tests were performed in duplicates with two concentrations (11.4 and 25.7 mg Sb/L) and a control, with each group comprising 10 fish. The mortality in the highest of the two dose groups was 45 %, so no LC₅₀ could be determined other than ”greater-than” values. This study also contained the results of static tests performed onPimephales promelas, which resulted in LC₅₀ values of 20.8 mg Sb/L, 17.4 mg Sb/L, and 14.4 mg Sb/L for the time periods 24 h, 48 h, and 96 h, respectively. No dose-response relationship was reported, but these results are considered reliable for the following reasons: the used methodology is well described, the antimony concentrations were measured, the water characteristics remained within the tolerance limits of the test species, and the estimated LC₅₀ values were within the range of the test concentrations used. The effect values presented in this study were in line with those reported by Kimball (1978) using the same fish species.

In the study by TAI (1990) juvenile Ictalurus punctatus were exposed to trivalent antimony (SbCl₃) in a static test design for 4 days, using moderately-hard reconstituted culture water as test medium. The tests were performed in duplicates with five measured concentrations (nominal concentrations within brackets) of 8.5 (15.6), 15.4 (31.25), 19.8 (62.6), 24.6 (125), and 21.2 (250) mg Sb/L, and a control with 0 mg Sb/L (measured concentration), with each group comprising 10 fish. The resulting LC₅₀ was 24.6 ± 2.6 mg Sb/L.

The reason that Curtis and Ward (1981) is considered unreliable, even though it included analytical monitoring of the test concentrations, is that the reported effect concentration is not considered to represent a dissolved antimony concentration. This conclusion is based on the fact that (i) the reported “greater than” concentration far exceeds the water solubility of diantimony trioxide, (ii) the concentrations used are not presented, (iii) the test substances in the study were added either directly or in the form of a stock solution in deionized water and the solutions were briefly stirred with a glass rod before a water sample was removed for analysis (i. e. there was no initial pretreatment of the diantimony trioxide to ensure that it was properly dissolved before it was added to the test solution), (iv) initially water samples were not filtered before analysis; filtering through 0.45 µm filters before analysis was performed at a later stage (but it is unclear when and for which chemicals), and (v) it is specifically mentioned that the results of the analysis primarily were used for information about the physicochemical behaviour of the toxicants rather than for computing LC₅₀s (which casts some doubt on whether the nominal or measured concentrations were used).

The reasons why the results reported by Doe et al. (1987) for Oncorhyncus mykiss are considered to be unreliable, even though the exposure concentration was measured, are that there is no information presented on (i) the number of concentrations and which concentrations were used, (ii) the dose-response curves (no raw data are reported), (iii) the number of replicates (if any), and (iv) the statistics that were used to calculate the LC₅₀ values. The reported acute levels for the fish O.latipes (Nam et al, 2009) are not considered reliable for the hazard assessment of antimony and antimony compounds as the test was conducted with antimony potassium tartrate. Dissolved antimony forms a complex with tartrate, and therefore only a part of the total amount of antimony will be present as “free” antimony; the exact concentration of free antimony can only be estimated via speciation modeling. The reported LC₅₀ values therefore represent the toxicity of the dissolved Sb-tartrate complex at equilibrium, and not the toxicity of the Sb-ion. 

USEPA (1988) reports results from Spehar (1987) for an acute study with Lepomis macrochirus 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 Brooke et al. (1986); hence the data are not critical for the hazard assessment.

Saltwater

Only a single study on the acute toxicity to marine fish was considered valid (Takayanagi, 2001).

In the study by Takayanagi (2001) 3 month-old Pargus major were exposed to trivalent (SbCl₃) or pentavalent antimony (SbCl₅ or K[Sb(OH)₆]) under static conditions with a control and an unknown number of concentrations (range: SbCl₃: 7.8-25.7 mg Sb/L; SbCl₅: 0.40-1.06 mg Sb/L; K[Sb(OH)₆]: 2.8 -10.3 mg Sb/L), with each group comprising eight fish. The tests were performed using natural seawater, with a salinity of 33.7 ppt, passed through sand and activated-charcoal filters. Each aquarium was aerated. The pH was determined daily and the reported ranges were 4.9-7.8, 7.8-8.1, and 7.8-8.1, for SbCl₃, SbCl₅, and K[Sb(OH)₆], respectively. For the SbCl₃ test, a decrease in the pH of the test solution was observed. Therefore, a low pH seawater was prepared with HCl for use as control for the SbCl₃ test in order to assess pH effects. All test fish survived in the HCl-adjusted seawater, and therefore pH was considered to be a negligible factor, and the mortality found in the SbCl₃dilution waters was considered to have been caused by the SbCl₃. The concentrations of antimony were measured at the beginning and end of the experiments using the hydride-generation atomic absorption method. The resulting EC₅₀ values for 24 h, 48 h, 72 h, and 96 h exposure were 15.5, 15.5, 15.2, 12.4; 0.93, 0.93, 0.93, 0.93; and 6.9, 6.9, 6.9, 6.9, for SbCl₃, SbCl₅, and K[Sb(OH)₆], respectively (all concentrations in mg Sb/L). However, the results from using the pentavalent SbCl₅ appear to be questionable and will therefore not be used.