Registration Dossier

Ecotoxicological information

Toxicity to terrestrial arthropods

Currently viewing:

Administrative data

Link to relevant study record(s)

Description of key information

The chronic NOEC for toxicity to soil invertebrates is 999 mg Sb/kg dw (Moser, 2007).

Key value for chemical safety assessment

Additional information

There are no acute toxicity studies on soil invertebrates available. There are six studies available on antimony toxicity to invertebrates which included analytical confirmation of the exposure concentrations but, as stated above, only the study by Moser (2007), which results in a bounded NOEC of 999 mg Sb/kg dw, will be used when deriving a PNECsoil.

Study 1

Moser (2007) studied the toxicity of Sb to the springtail Folsomia candida (reproduction and mortality) in approximately 32-week aged Sb2O3amended soil. This study was carried out according to ISO 11267 (1999). Ten (10-12 d old) synchronised springtails were put onto 30 g moist soil in a glass vessel during an exposure period of four weeks. Five replicates were used for each test concentration and the measured test concentrations used were 90, 322, 999, 2930, 10119 mg Sb/ kg soil dw. Additionally, the springtails were also tested in an uncontaminated field soil control and an untreated artificial soil as a further control, which was based on OECD 207. The positive control used on springtail reproduction was the herbicide Betosip (a.i. 159 g/L phenmedipham). Granulated dry yeast served as food for the springtails and was added onto the soil surface at the beginning of the test and after 14 days. At the end of the exposure period, the number of juveniles in each test vessel was counted after floating. The mortality of the springtails was also recorded. At day 28, the pH and the moisture of the artificial soil for each concentration in additional vessels without springtails were determined. It was shown that the pH and moisture content did not diverge from the guideline recommendations.

For both mortality and reproduction, the highest effect observed was below 50% in any of the treatments. Consequently, LC50 and EC50 values could not be calculated but were estimated as being >10119 mg Sb/kg soil dw for both endpoints. The NOEC for reproduction was determined as 999 mg Sb/kg dw. The LOEC for reproduction was derived as 2930 mg Sb/kg dw. All values refer to measured concentrations.

The results from this study are considered reliable. During the ageing period a substantial part of the Sb2O3was transformed into soluble Sb. Further transformation of Sb2O3after this ageing period would have been very slow and so a sufficiently constant toxic pressure is considered to have been obtained during these bioassays. Using Sb2O3amended soils avoids the confounding effects of counter-ions or lowered pH, and therefore the observed toxic effects can be attributed to the increasing Sb dose alone.

However, since not all Sb2O3had dissolved during the aging period, the NOEC of 999 mg total Sb/kg dw underestimates the toxicity. Therefore, the NOEC is based on the pore water Sb concentration multiplied by the equilibrium solid:liquid distribution coefficient (Kd) for Sb in this soil. The Kd value for the soil used in the present study is 38 L/kg, which is the value observed for the Sb2O3amended soil aged for five years and for the soluble SbCl3added to soil (Oorts et al., 2005). The resulting NOEC after this correction is 370 mg Sb/kg dw (9.7 mg Sb/l * 38 L/kg).

Study 2

Simini and co-workers (2002) exposed the earthworm Eisenia fetida, measuring adult survival and cocoon production according to ISO 11268-2 (International Organization for Standardization, 1998b), to various concentrations of Sb2(SO4)3in a natural sandy loam. The original ISO method was designed for use with artificial soil (USEPA Standard Artificial Soil). However, research by the authors showed that the test could also successfully be conducted using natural soils (Kuperman et al. 2004). The method was modified for use with natural soils having physical and chemical characteristics that support a relatively high level of metal bioavailability. After performing a range-finding test, a definitive test was performed with an exposure period of three weeks and seven concentration groups, besides the control. The nominal added concentrations used to assess effects on reproduction were 60, 86, 104, 124, 149, 179, and 215 mg Sb/kg. Four replicates were used per concentration, with five earthworms per replicate. Toxicity tests using the salt carrier control CaSO4.2H2O were performed in order to evaluate the effect of sulphate. The positive control was 4-nitrophenol. All soil treatment concentrations and controls were subjected to simulated aging/weathering procedures, which included alternating wetting/air-drying cycles for three weeks prior to commencement of definite tests. Hydration and moisture equilibration followed the weathering and aging process before exposing organisms in definitive studies.

Sulphate control treatments showed no statistically significant effect on reproduction when compared with negative controls. Juvenile or cocoon production in positive controls (4-nitrophenol) were within the baseline established for the laboratory cultures of E. fetida. The decrease in pH in the highest Sb treatment was below 1.0 pH unit, compared with untreated soil (i.e. negative control). In the sulphate control, soil pH decreased by less than 1.0 pH unit in the 7000 and 35000 mg SO4treatments, when compared with negative controls.

Only about 58% of nominal Sb concentrations were recovered. The nominal concentrations were used when determining the ecotoxicological parameters of Sb with the results of definitive tests of Sb toxicity resulting in NOEC/LOEC values for adult survival of 617/697 mg Sb/kg and 60/86 for juvenile production. The authors recommended that the nominal values should be adjusted to 58% of the nominal values, as that was the average recovery. This may not necessarily be an exposure problem, but could instead be an analytical problem and therefore the nominal values are used. Since the Sb soil concentration was not measured during the course of the experiment it is not possible to know if a steady-state was reached after the three weeks of aging/weathering, but it is not considered likely. As a consequence, the toxicity pressure has probably not been constant during the exposure period.

The results from this study can only be used as supportive evidence. According to the ISO protocol on which the study was based, the earthworm reproduction test is considered valid only if the coefficient of variation for the mean number of cocoons from the control is ≤30% at the end of the test. The coefficient of variation for the reproductive performance of the control in the study by Simini et al. (2002) could be calculated from the raw data in an annex and was 45%.

To conclude, the results from this study are considered unreliable.

Study 3

Oorts et al. (2005) studied the toxicity of Sb to the springtail Folsomia candida (reproduction and mortality) in freshly spiked (Sb2O3or SbCl3) and in five-year aged Sb2O3amended soil. The study was carried out according to ISO 11267 (1999). Ten (10-12 d old) synchronised springtails were exposed to 30 g moist soil in a glass vessel during an exposure period of four weeks. Granulated dry yeast was added onto the soil surface every week and served as food for the springtails. At the end of the exposure period the number of juveniles in each test vessel was counted after floating. The mortality of the springtails was also recorded.

A large variation in the toxicity data in several of the exposed groups was noticeable and reduces the possibility of detecting deviations from the control. The authors, who used ANOVA and the post hoc Duncan test, reported an unbounded NOEC of 2000 mg Sb/kg dw nominal concentration (measured concentration of 1864 mg Sb/kg dw) for number of juveniles and adult mortality. The use of the step-down approach with the Jonckheere-Terpstra test results in a NOEC of 1000 mg Sb/kg dw nominal concentration (= measured concentration of 897 mg Sb/kg dw) and a LOEC of 2000 mg Sb/kg dw nominal concentration (measured concentration of 1864 mg Sb/kg dw).

An unbounded NOEC of 1804 mg Sb/kg dw (measured concentration) results from a freshly Sb2O3spiked soil. Using statistical methods considered more appropriate for dose-response studies (the step-down approach), rather than the method used by Oorts et al. (2005), which was ANOVA and Duncan´s method, results in a bounded NOEC of 897 mg Sb/kg dw. However, due to the increasing porewater concentration, which indicates that the toxicity could still increase during the exposure period, a derived NOEC (not unbounded) using freshly spiked Sb2O3soils may be too high. Having the same soil solution concentration in an exposure regime with constant toxic pressure (i.e. aged soil) may have resulted in lower NOEC values.

The data from freshly SbCl3spiked soils results in a NOEC < 20 mg Sb/kg dw (measured concentration of 10 mg Sb/kg dw). This "less than” value was considered unreliable by Oorts et al. (2005) since they considered that the weight of evidence showed no toxic effect at doses up to 500 mg Sb/kg dw nominal concentration (measured concentration of 384 mg Sb/kg dw). The difference in response between the control and the first five SbCl3doses, up to 500 mg Sb/kg dw, nominal concentration was considered to be within the non-significant variation in response that was observed for Sb2O3amended soils.

However, it is difficult to draw any firm conclusions about the absence of toxicity for the freshly spiked Sb2O3soils up to 500 mg Sb/kg dw nominal concentration, given the variation in response, within and between the different dose groups, especially for the number of juveniles. The NOEC < 20 mg Sb/kg nominal dose resulting from the freshly spiked SbCl3, which was considered unreliable by Oorts et al. (2005), has a measured Sb soil solution concentration of 0.62 mg/L. This is almost identical to the Sb soil solution concentration of 0.63 mg Sb/L, which was measured in the freshly spiked Sb2O3nominal concentration 200 mg Sb/kg dw group, a dose group with three out of four values lower than the lowest of the control values. In addition, the toxic pressure between the two concentrations most likely differed because Sb2O3is expected to dissolve slowly over the time frame of the experiment. The Sb soil solution concentration measured in the freshly spiked Sb2O3500 mg nominal dose group, was 9 mg Sb/L, in which three out of four values were again lower than the lowest of the control values.

The effect observed in the aged Sb2O3spiked soil study at the two highest doses may be explained by the presence of a herbicide, which was a mixture including substances with toxicity towards earthworms. The Sb soil solution concentration in the highest of the remaining doses, i.e. 24 mg Sb/kg dw., was 0.745 mg Sb/L. The four data points at that concentration indicate a response, but the difference is not significant.

Study 4

In Heijerick and Vangheluwe (2003) the microdrile oligochaete Enchytraeus albidus was exposed to various concentrations of trivalent antimony (SbCl3) in an artificial soil and the endpoints mortality and reproduction were studied. The standard soil used is described and recommended by OECD (1984). This study was performed with five replicates per concentration, five concentrations (measured concentration range: 59.1 – 6980 mg Sb/kg dw) and a control (measured concentration 0.37 mg Sb/kg dw), with each replicate consisting of 10 enchytraeids with fully developed clitellum. Four replicates were used for the toxicity test and the remaining replicate was used for the chemical analyses. Adult enchytraeids were exposed for 21 days (which is the period in which cocoons are produced). After this period, the adult enchytraeids were removed from the soil and survival was examined. The cocoons were incubated for another 21 days (i.e. a total of 42 days) after which reproduction was measured (number of young enchytraeids). The resulting NOEC and LOEC were 760 mg Sb/mg kg dw and 2012 mg Sb/mg kg dw, respectively, for both endpoints. One of the three performance criteria specified in the OECD test guideline 220 (Enchytraeid reproduction test) that should be met in the control in order for the test to be valid was not fulfilled. This is that the coefficient of variation for number of juveniles should not be higher than 50% at the end of the reproduction phase; in this study it was 81%. The results from this study are therefore considered to be unreliable.

Study 5

Kuperman et al. (2002) exposed the enchytraeid Enchytraeus crypticus, measuring adult survival and juvenile production according to ISO 16387 (International Organization for Standardization, 2003) to various concentrations of Sb2(SO4)3in a natural sandy loam. The original ISO methods were designed for use with artificial soil (USEPA Standard Artificial Soil). These methods were modified for use with natural soils having physical and chemical characteristics that support a relatively high level of metal bioavailability. The modifications included different soil hydration levels that have lower water holding capacity compared with the artificial soil, and a shorter duration test forE. crypticus(28 vs. 42 d) because of the shorter generation time of this species compared with Enchytraeus albidus, for which the ISO 16387 test conditions were optimized.

In the enchytraeid Enchytraeus crypticus reproduction test ten enchytraeid adults with eggs in the clitellum were placed on top of prepared soil in each container and exposed to Sb2(SO4)3for four weeks. The nominal concentrations of Sb used were 0, 100, 140, 196, 274, 384, 538, 753, and 1054 mg Sb/kg. Four replicates were used per concentration. After two weeks, soil in each test container was carefully searched and adult worms were removed and counted. The remaining test substrate, including any cocoons laid during the first two weeks of the test, was incubated for an additional two weeks. After four weeks from the start of the test, soil in the test containers was fixed with ethanol and Rose Bengal biological stain was added. Staining continued for a minimum of 24 hours. The content of each test container was wet-sieved and transferred to a counting tray and worms were counted. Measurement endpoints included number of surviving adults after 14 days and number of juveniles produced after 28 days.

The authors were not able to measure the Sb concentrations in the soils adequately. Despite a change of methods and improved efficiency of Sb extraction, on average only 58% of the added antimony could be measured. For this reason, nominal concentrations were used to determine the ecotoxicological parameters for Sb. Using methods developed by the authors, the amended soils were weathered and aged for three weeks before definitive testing. Hydration and moisture equilibration followed the weathering and aging process before exposing organisms in definitive studies. The decrease in pH in the highest Sb treatment was 1.2 pH units when compared with untreated soil (i.e. negative control). In the sulphate control, soil pH decreased by less than 1.0 pH units in both the 7000 and 35000 mg SO4-treatments when compared with the negative control. Sulphate control treatments showed no statistically significant (p>0.05) effect on adult survival or reproduction measurement endpoints when compared to the negative controls. Juvenile or cocoon production in positive controls (4-nitrophenol) were within the baseline established for the laboratory cultures of E. crypticus. Results of definitive tests of Sb toxicity resulted in NOEC/LOEC values for adult survival of 384/538 mg Sb/kg. For juvenile production the resulting NOEC/LOEC values were 100/140 mg Sb/kg. The authors recommended that the nominal values should be adjusted to 58% of the nominal values, as that was the average recovery. This may not necessarily be an exposure problem, but could instead be an analytical problem. However, since it is not known which of the two alternatives is best it was decided to use the conversion factor of 0.58 which results in NOECs of 58 mg Sb/kg and 223 mg Sb/kg for juvenile production and adult survival respectively. Since the Sb soil concentration was not measured during the course of the experiment it is not possible to know if a steady-state was reached after the three weeks of aging/weathering, but this is not considered to be likely. As a consequence, the toxicity pressure was probably not constant during the exposure period.

The results indicate a NOEC for the reproduction of Enchytraeus crypticus of 58 mg Sb/kg dw. There are, however, some concerns about using this value since i) the result was obtained using Sb2(SO4)3which means that counter ions (i.e. sulphate) and protons were also added to the soil, which may have influenced the results, ii) it is not known how the toxic pressure may have varied over the exposure period, and iii) the authors had difficulties in measuring the Sb concentration in soil.

Study 6

Phillips et al. (2002) exposed the collembolan Folsomia candida, measuring adult survival and juvenile production according to ISO 11267 (International Organization for Standardization, 1998a), to various concentrations of Sb2(SO4)3in a natural sandy loam.

In the Folsomia candida reproduction test, 10-12 day-old juveniles were exposed to Sb2(SO4)3for four weeks. The nominal concentrations of Sb used were 0, 100, 140, 196, 274, 384, 538, 753, and 1054 mg Sb/kg. Five replicates were used per concentration. After the exposure period, purified water was added to each test chamber to bring the level up to half its volume. After gentle mixing with a spatula, the chamber was examined under a dissecting microscope for the presence of juveniles and adults. The juveniles that floated to the surface were counted and removed. This procedure was repeated until no other springtails floated to the surface. The chamber was given a final mixing and examined once more to ensure that all individuals were counted. Adult survival and juvenile production were used as endpoints.

The authors were not able to measure the Sb concentrations in the soils adequately. Despite a change of methods and improved efficiency of Sb extraction, on average only 58% of the added antimony could be measured. For this reason, nominal concentrations were used to determine the ecotoxicological parameters for Sb. Using methods developed by the authors, the amended soils were weathered and aged for three weeks before definitive testing. Hydration and moisture equilibration followed the weathering and aging process before exposing organisms in definitive studies. The decrease in pH in the highest Sb treatment (nominal concentration = 1054 mg Sb/kg dw.) was 1.2 pH unit when compared with untreated soil (i.e. negative control). In the sulphate control, soil pH decreased by less than 1.0 pH unit in both the 7000 and 35000 mg SO4-treatment when compared with the negative control. Sulphate control treatments showed no statistically significant (p>0.05) effect on adult survival or reproduction measurement endpoints when compared to the negative controls. Juvenile or cocoon production in positive controls (carbamate) were within the baseline established for the laboratory cultures of F. candida. Results of definitive tests of Sb toxicity resulted in NOEC/LOEC values of 100/126 mg Sb/kg for adult survival. Juvenile production was significantly (p=0.045) decreased in the lowest positive Sb treatment compared with the negative control, producing an unbound LOEC of 100 mg Sb/kg (based on Fisher's least-significant-difference pairwise comparison test). Bounded NOEC and LOEC values of 100 and 126 mg/kg were determined using the more conservative Bonferroni mean comparison test. The authors recommended that the nominal values should be adjusted to 58% of the nominal values, as that was the average recovery. This may not necessarily be an exposure problem, but could instead be an analytical problem. However, since it is not known which of the two alternatives is best it was decided to use a conversion factor of 0.58 which results in a NOEC of 58 mg Sb/kg for both juvenile production and adult survival.

The results indicate that a NOEC for reproduction may be around 58 mg Sb/kg, and the authors suggested that it could be even lower. The post hoc test used by the authors (Fisher's LSD) does not correct for multiple comparisons, which means that it is easier to find statistical significance with the Fisher's LSD test (it has more power) than with other multiple comparison tests that correct for multiple comparisons, but that also means that false positive results occur in more than 5% of analyses. In order to correct for this the authors used the Bonferroni correction on the p-values resulting from the Folsomia candida reproduction test, which resulted in a NOEC of 100 mg Sb/kg. The Bonferroni correction is generally overly conservative, especially for large k. However, the use of the step-down approach also results in a NOEC of 100 mg Sb/kg. There are some concerns about using this value since i) the result was obtained using Sb2(SO4)3which means that counter ions (i.e. sulphate) and protons were also added to soil, which may have influenced the results, ii) it is not known how the toxic pressure may have varied over the exposure period, and iii) the authors had difficulties when measuring the Sb concentration in soil.