- Email: [email protected]
Environmental risk assessment for the polycyclic musks, AHTN and HHCB
Toxicology Letters 111 (1999) 81 – 94 www.elsevier.com/locate/toxlet
Environmental risk assessment for the polycyclic musks, AHTN and HHCB II. Effect assessment and risk characterisation Froukje Balk a, Richard A. Ford b,1,* a b
HASKONING Consulting Engineers and Architects, P.O. Box 151, 6500 AD Nijmegen, The Netherlands Research Institute for Fragrance Materials, Suite 406, Two Uni6ersity Plaza, Hackensack, NJ 07601, USA Received 30 March 1999; received in revised form 16 August 1999; accepted 16 August 1999
Abstract Reports of the polycyclic musks AHTN and HHCB in surface water and fish, primarily in Europe, have prompted studies of their environmental effects. These effects then are used, along with the predicted environmental concentrations in a risk assessment according to the approach developed under European Union Regulation 793/93, in line with the Technical Guidance Document for risk assessment of new and existing chemicals. In 72-h studies with algae (Pseudokirchneriella subcapitata), NOECs were 0.374 mg/l (AHTN) and 0.201 mg/ l (HHCB). In 21-day reproductive tests with daphnia (Daphnia magna) NOECs were 0.196 (AHTN) and 0.111 mg/l (HHCB). In 21-day growth tests with bluegill sunfish (Lepomis macrochirus), NOECs were 0.067 (AHTN) and 0.068 mg/l (HHCB). And, finally 35-day early life stage tests with fathead minnows (Pimephales promelas) resulted in NOECs of 0.035 (AHTN) and 0.068 mg/l (HHCB). These results lead to Predicted No Effect Concentrations (PNEC) of 3.5 mg/l (AHTN) and 6.8 mg/l (HHCB) for aquatic organisms. For the soil compartment, 8-week studies with earthworms (Eisenia fetida) resulted in NOECs of 105 (AHTN) and 45 mg/kg (HHCB) and 4-week studies with springtails (Folsomia candida) resulted in a NOECs of 45 mg/kg for both substances. These values lead to a PNEC of 0.32 mg/kg dw for both materials. Using mammalian studies, PNECs for fish or worm eating predators of 10 mg/kg fw (AHTN) and 100 mg/kg fw (HHCB) can be derived. For sediment dwelling organisms, PNECs were derived by equilibrium partitioning using the aquatic PNECs. Comparing PNECs with the measured or predicted environmental exposures leads to risk characterisation ratios as follows: aquatic species: AHTN 0.086, HHCB 0.074; sediment organisms: AHTN 0.44, HHCB 0.064; soil organisms: AHTN 0.091, HHCB 0.10; fish eating predators: AHTN 0.012, HHCB 0.001; worm eating predators: AHTN 0.007, HHCB 0.001. © 1999 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: AHTN; HHCB; Musks; Ecotoxicity; Alga; Daphnia; Fish; Earthworm; Springtail; Risk assessment
* Corresponding author. Present address: 7 Alwyn Ave., London W4 4PA, UK. Tel.: +44-181-987-8626; fax: +44-181-987-8627. E-mail addresses: [email protected] (F. Balk), [email protected] (R.A. Ford) 1 Please address correspondence to RIFM. 0378-4274/99/$ - see front matter © 1999 Published by Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 4 2 7 4 ( 9 9 ) 0 0 1 7 0 - 8
82
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
1. Introduction The polycyclic musks 7-acetyl-1,1,3,4,4,6-hexamethyl-1,2,3,4-tetrahydronaphthalene (AHTN) and 1,3,4,6,7,8-Hexahydro-4,6,6,7,8,8-hexamethylcyclopenta-g-2-benzopyran (HHCB) are widely used as fragrances ingredients in detergents and other household products. This, combined with high log Kow values and lack of ready biodegradability, has resulted in their widespread occurrence in the environment (Balk and Ford, 1999). AHTN and HHCB together represent about 95% of the total market volume for the class of fragrance ingredients known as the polycyclic musks. In view of reported levels in surface water and fish primarily in Europe, the Research Institute of Fragrance Materials (RIFM) initiated an environmental risk evaluation according to the approach developed under European Union Regulation 793/93, in line with the Technical Guidance Document for risk assessment of new and existing chemicals (EU-TGD; EC, 1996). In an environmental risk assessment, the environmental exposure is compared to data on environmental effects. The exposure of the environment to AHTN and HHCB is presented in a separate paper (Balk and Ford, 1999, part I) in this issue. Because the reported occurrences of these materials are in water and fish, these were considered the most important issues to address. However, these occurrences along with measured values in sludge, allow prediction of exposures of soil and sediment-dwelling organisms and of predators feeding on these organisms or on fish. Thus, the Table 1 Characteristics of AHTN and HHCB
risks for these endpoints are also addressed. This paper reports on the environmental effects of the AHTN and HHCB on the basis of a series of ecotoxicity studies on both materials that have been carried out on behalf of RIFM, and presents the environmental risk assessment.
2. Materials and methods The structure and some relevant characteristics of AHTN and HHCB are given in Table 1. Because of the low water solubilities of AHTN and HHCB, it was necessary in the aquatic toxicity tests described below to prepare stock solutions using DMF as a solvent and Tween 80 as a dispersant or using triethylene glycol as a solvent. These stock solutions were then diluted to reach the desired concentrations in the tests. The residual level of the solvent in the test vessel was always below the maximum level allowed by the test guidelines. Solvent controls containing equivalent levels of the solvents were used in all cases along with undosed water controls. These substances have a strong tendency to sorb, thereby reducing the concentrations in solutions over time. Final concentrations in water were determined by HPLC or GC and the test concentrations are expressed as measured concentrations. Except for the highest concentration in the fish growth test, the tested concentrations did not exceed the water solubility limit. Details are summarised in Tables 2 and 3. Stock solutions for the terrestrial toxicity studies were prepared with acetone as the solvent. After thorough mixing with the soil, the acetone was then allowed to evaporate. Final concentrations in the test medium were not determined but are based on the known amounts of test substance added to the soil. All toxicity tests and chemical analyses were carried out and inspected according to the Principles of Good Laboratory Practice.
2.1. Algae The toxicity to algae was studied in a static test according to OECD Test Guideline 201 with
Table 2 Aquatic toxicity of AHTN Resultsa (mg/l)
Remarksb
P. subcapitata c 72-h static (Van Dijk, 1997a)
Test A: NOECg =0.438; NOECb =0.204; LOECbd =0.438; EgC50\0.797; EbC50 =0.468 0.434–0.508. Test B: NOEC = 0.374; LOECd =0.835; EgC50\0.835; EbC50 :0.835.
D. magna 21-day semi-static (Wu¨thrich, 1996a)
NOEC(rep) = 0.196; LOECd =0.401; ErC50-21-day = 0.244 0.239–0.249; 21-day-IC50 =0.341 (mobility); 0.243–0.433. NOECgrowth= 0.089; LOECd =0.184; 21-day-LC50 =0.314 0.226-0.448.
Carrier: 0.005% DMF and 0.005% Tween 80; n=5. Test A: HPLC identification; Start conc. 81–90% of nominal; End conc. 31–85% of nominal. Test B: start conc. 77–90% of nominal; end conc. 53–142% of nominal; geometric mean NOEC= 0.276. Carrier: 0.008% DMF and 0.002% Tween 80; n= 5; HPLC identification; conc. fresh 84–103% of nominal; conc. used 70–85% of nominal. Carrier: 0.005% DMF and 0.005% Tween 80; n= 5; HPLC identification; conc. 57–109% of nominal. Solvent triethylene glycol; GC identification; Conc. 55–108% of nominal.
Bluegill sunfish; L. macrochirus; 21-day flow-through (Wu¨thrich, 1996b)
Fathead minnow; P. promelas; 32 days post LOEChatch\0.140; NOECsurv. =0.067; hatch; 36 days overall (Croudace et al., 1997a) LOECsurv.= 0.140; 32-day-LC50 =0.100 0.097–0.100; NOECgrowth =0.035; LOECgrowthe =0.067; NOECdevelop. =0.035; LOECdevelop. = 0.067. 95% confidence limits. The number of concentrations tested (n) excludes control and solvent control. c Former name S. capricornutum. d Dunnet’s test (P= 0.05). e Wilcoxon rank sum test (P= 0.05). a
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
Test and reference
b
83
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
84 Table 3 Aquatic toxicity of HHCB Test and reference
Resultsa (mg/l)
Remarksb
P. subcapitata c; 72-h static (Van Dijk, 1997b)
NOEC= 0.201; LOECd =0.466; EgC50\0.854; EbC50 = 0.723 0.678-0.778.
D. magna; 21-day semi-static (Wu¨thrich, 1996c)
NOEC(rep)= 0.111, LOECd =0.205; ErC50-21-day= 0.282 0.260–0.312; IC50-21-day= 0.293 (mobility) 0.204–0.419e. NOEC =0.093 (clinical signs); LOECd = 0.182; NOEC(growth)= 0.182; LC50-21-day=0.452 0.316–0.911. LOEChatch\0.140; NOECsurv. =0.068; LOECsurv.= 0.140; 32-day-LC50\0.140; NOECgrowth=0.068; LOECgrowthf =0.140; NOECdevelop. = 0.068; LOECdevelop.= 0.140.
Carrier: 0.005% DMF and 0.005% Tween 80; n = 6; HPLC identification; start conc. 71–102% of nominal; end conc. 54–85% of nominal. Carrier: 0.008% DMF and 0.002% Tween 80; n =5; HPLC identification; conc. fresh 82–104% of nominal; conc. used 63–91% of nominal. Carrier: 0.005% DMF and 0.005% Tween 80; n = 5; HPLC identification; conc. 66–86% of nominal.
Bluegill sunfish; L. macrochirus; 21-day flow-through (Wu¨thrich, 1996d)
Fathead minnow; P. promelas; 32 days post hatch; 36 days overall (Croudace et al., 1997b)
Solvent triethylene glycol; GC identification; conc. 50–104% of nominal.
95% confidence limits. The number of concentrations tested (n) excludes control and solvent control. c Former name S. capricornutum. d Dunnet’s test (P = 0.05). e Estimated 95% confidence limits after data reported by Wu¨thrich (1996c). f Wilcoxon rank sum test (P= 0.05). a
b
Pseudokirchneriella subcapitata (Van Dijk, 1997a,b). Nominal concentrations ranged from 0.0625 to 1.0 mg/l (step size 2). The NOEC and EC50 are based on growth rate and biomass production after 72 h. Concentrations were measured at the start and end of the test. The test with AHTN was carried out twice (tests A and B) due to the sharp concentration decrease in the first study with the goal to confirm the observed toxicity levels in a second study.
2.2. Daphnia For Daphnia magna, a semi-static 21-day toxicity test was carried out according to OECD Test Guideline 202, part II, proposed updated version of June 1993 (Wu¨thrich, 1996a,c). The test medium was refreshed three times per week. Nominal concentrations ranged from 0.062 to 1.0 mg/l (step size 2). Concentrations were measured
at the start and end of the first and last exposure period.
2.3. Fish A 21-day prolonged toxicity test was carried out with bluegill sunfish (Lepomis macrochirus) according to OECD Test Guideline 204 under flow-through conditions (Wu¨thrich, 1996b,d). Test parameters included fish growth (weight and length). The fish weights at the start of the experiment varied between 1.4 and 1.6 g. The concentrations ranged from 0.125 to 2.0 mg/l (step size 2). Concentrations were measured at the start, halfway through and at the end of the test period. An early life stage test was carried out with fathead minnow (Pimephales promelas) according to OECD Test Guideline 210 under flow-through conditions (Croudace et al., 1997a,b). Eggs less than 24 h old were exposed to nominal concentra-
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
tions ranging from 0.0125 to 0.2 mg/l (step size 2). Concentrations were measured 13 times at regular intervals during the 36-day test period.
2.4. Earthworm The earthworm test was carried out according to ISO 11268 (Gossmann, 1997a,b). Adult worms (Eisenia fetida) were exposed to nominal concentrations in soil of 8, 19, 45, 105 and 250 mg/kg. The test medium was an artificial soil according to ISO-Standard 11268-1 and OECD Test Guideline 207, containing 10% Sphagnum peat, 20% kaolinite clay, approximately 70% fine quartz-sand (grain size 0.1–0.5 mm) and 0.5% calcium carbonate to adjust to pH 6.0 90.5. The test materials at appropriate concentrations were dissolved in equal amounts of acetone, mixed with the quartz sand and allowed to slowly evaporate. This sand was then thoroughly mixed with the other components to make the standardised soil. The control soil was treated in the same manner but with pure acetone. After preparation of the test concentrations and an equilibrium period of 1 week, the test organisms were added to the soil. The test medium was not refreshed during the test period. Weights of the adult worms ranged between 340 and 540 mg, but did not differ more than 100 mg within this range in each test container. The worms were fed weekly with finely ground cattle manure. Adult worms were removed after 4 weeks of exposure, counted and weighed. The remaining offspring remained in the test containers for another 4 weeks.
2.5. Springtails The springtail test was carried out according to the draft ISO/CD 11267 (Klepka, 1997a,b). Juvenile springtails of the species Folsomia candida 10 – 12 days of age were placed in an artificial soil and survival and reproduction after 28 days were determined. The artificial soil was the same as used in the earthworm study. Nominal test concentrations were 1, 3, 8, 19, 45 and 105 mg/kg soil. After preparation of the test concentrations and an equilibrium period of 1 week, the test organisms were added to the soil. The animals were fed with granulated dry yeast.
85
2.6. Predators Toxicity studies with AHTN and HHCB on predators present in the environment like fish-eating birds and mammals are not available. Therefore, toxicity data for laboratory mammals are used to derive a PNEC for these predators. For AHTN and HHCB, subchronic oral studies are available for rats (Hopkins and Lambert, 1996; Api and Ford, 1999). AHTN was administered in the diet of rats (15 males and 15 females per group) at daily doses of 1.5, 5, 15 and 50 mg/kg for 13 weeks. HHCB was tested by an identical protocol but at doses of 5, 15, 50 and 150 mg/kg per day.
2.7. Deri6ation of the PNEC In the EU-TGD, the PNEC (Predicted No Effect Concentration) is defined as the concentration below which unacceptable effects on organisms will most likely not occur. The environmental protection goals considered here are the aquatic ecosystem (PNECwater, PNECsediment), the terrestrial ecosystem (PNECsoil) and top predators (PNECpredator). The PNECwater was obtained using the assessment factors applied to the lowest NOEC from the aquatic toxicity studies, as described in the TGD. In this case where three or more chronic studies are available for algae, daphnia and fish, an assessment factor of 10 may be applied to the lowest NOEC to derive PNECwater. For the derivation of a PNECsoil based on two chronic studies with soil organisms, the EU-TGD proposes an assessment factor of 50 to be applied to the lowest NOEC. Before doing so, the TGD states that the results from the soil toxicity tests should be normalised using a relationship that describes the bioavailability of the test substance in the soil. The defined standard soil of the TGD contains 3.4% organic matter, whereas the tests presented here were carried out in the standard OECD soil with an organic matter content of 10%. Therefore the test results are normalised according to the formula for organic substances (EU-TGD):
86
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
NOECstandard soil =
3. Results
NOECexperiment ×Fomstandard soil/Fomexperiment
3.1. Toxicity test results for AHTN
where: NOECstandard soil is the normalised NOEC for the defined standard soil with 3.4% organic matter; NOECexperiment is the NOEC from experiment (Table 4); Fomstandard soil is the weight fraction of organic matter in soil solids of the standard soil (3.4%); and Fomexperiment is the weight fraction of organic matter in soil solids in tests (10%). This has the practical effect of dividing the observed NOEC by an additional 3, in addition to the assessment factor of 50 for a total factor of 150. PNECsoil and PNECsediment can also be derived from PNECwater on the basis of equilibrium partitioning between the solid and the aqueous phases. According to the equilibrium partitioning theory it is assumed that test organisms are exposed only through the pore water and that soil or sediment organisms and aquatic organisms are equally sensitive to the test substances. Assuming a concentration in the porewater equal to PNECwater the corresponding equilibrium concentration in the soil is estimated using the soil – water partition coefficient Ksoil – water. Ksoil – water is estimated on the basis of sorption to organic carbon in the soil and a correction factor for the soil density (EU-TGD). The general applicability of the equilibrium partitioning method has been tested less for soil than for sediment-dwelling organisms. The EU-TGD considers the approach as a screening for assessment of the risk for both organisms. This approach is presented in more detail in Table 6. For the estimation of a PNECpredator for fishand worm-eating birds and mammals, the NOEC determined for the rat is used as the starting point. For the conversion of a NOEC obtained by daily dosing (mg/kg body weight) to a NOEC level in the food (mg/kg food) a factor of 20 is used for a rat greater than 6 weeks old (EU-TGD) based on the fact that the food consumption of an adult rat is about 20 mg/kg body weight per day. According to the TGD the assessment factor to be applied to such a NOEC in the derivation of the PNEC from a 90-day toxicity test is 30.
The results for AHTN are summarised in Table 2 and Table 4.
3.1.1. Algae Mean measured concentrations were 0.035, 0.088, 0.204, 0.438 and 0.797 mg/l for test A and 0.0679, 0.140, 0.170, 0.374 and 0.835 mg/l for test B, but maintenance of the concentrations proved to be difficult and in some cases pH values rose to above 10 indicating an insufficient buffering capacity of the medium. Growth was not significantly inhibited in concentrations up to and including (A) 0.204 and (B) 0.374 mg/l. In the next higher concentration, the LOEC, (A) 0.438 and (B) 0.835 mg/l, the biomass production was inhibited by (A) 36% and (B) 54%, measured as Area Under the Curve. Growth rate (m) was the less sensitive parameter. Inhibition at 0.835 mg/l (B) was 16% (LOEC). In test (A), the NOEC for growth rate was higher, 0.438 mg/l, whereas the inhibition at 0.797 mg/l (LOEC) was 33%. The geometric mean NOEC for both tests was 0.28 mg/l. 3.1.2. Daphnia Mean measured concentrations were 0.054, 0.113, 0.196, 0.401 and 0.804 mg/l. Immobility of the parent generation was 80% at 0.401 mg/l and the 21 day-EC50 was 0.341 mg/l. The mean reproduction in test concentrations up to and including 0.196 mg/l ranged between 80 and 114% as compared to the solvent control. In the next higher test concentration of 0.401 mg/l, reproduction of the surviving adults was inhibited almost completely. Thus, 0.196 mg/l can be considered the NOEC in this study. 3.1.3. Fish Mean measured concentrations during the 21day test with L. macrochirus (Bluegill sunfish) were 0.089, 0.184, 0.392, 1.00 and 2.22 mg/l. Concentrations up to and including 0.184 mg/l did not significantly affect survival of the fish. Mortality was 70% at the next higher concentra-
Test and reference
Results for AHTN
Results for HHCB
Remarks
Earthworm; E. fetida; ISO 11268 (OECD 207) (Gossmann, 1997a,b)
8-week-NOEC =105 mg/kg; LOECa =250 mg/kg; reproduction and food consumption; 4-week-NOEC]250 mg/kg; mortality and growth. 4-week-NOEC =45 mg/kg; LOECc =105 mg/kg; mortality and reproduction.
8-week-NOEC = 45 mg/kg; LOECa = 105 mg/kg; reproduction and food consumption; 4-week-NOECgrowth = 105 mg/kg; LOEC = 250 mg/kg; 4-week-NOECsurvival]250 mg/kg. 4 week-NOEC=45 mg/kg; LOECc = 105 mg/kg; mortality and reproduction.
Initial weight adults 0.34–0.54 g; test range 8–250 mg/kg; solvent: acetone; artificial soil pH 6.1; 10% sphagnum DINb; temp. 17–23°C.
Springtail; F. candida; ISO/CD 11267 (Klepka, 1997a,b)
a
Dunnet’s test (P =0.05). Sphagnum DIN standard: organic material minimum 90%, organic carbon 52%. c Student’s t-test (P=0.05). b
10–12-day-old juveniles; test range 1–105 mg/kg; solvent: acetone; temperature 17–25°C; artificial soil; 10% sphagnum DINb.
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
Table 4 Toxicity data for AHTN and HHCB for soil organisms
87
88
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
tion of 0.392 mg/l and 100% in 1.00 mg/l after 11 days and in 2.22 mg/l after 4 days. The 21-day LC50 was 0.314 mg/l. Clinical signs such as loss of equilibrium, enhanced and/or irregular respiration and cessation of food intake were observed before the onset of death. Fish growth was significantly reduced at 0.184 mg/l. Therefore the NOEC is 0.089 mg/l. In the early life-stage test with P. promelas (fathead minnow) the mean measured concentrations were 0.008, 0.018, 0.035, 0.067 and 0.140 mg/l. Hatchability of eggs was not significantly affected in any of the test concentrations. Larval survival after 32 days was not affected in concentrations of 0.067 mg/l and below. In 0.14 mg/l larval survival was 18%. Larval growth was not affected in concentrations of 0.035 mg/l or below. At 0.067 and 0.14 mg/l, mean standard lengths were reduced by 7 and 38%, whereas weights were reduced by 7 and 75%, respectively, as compared to the solvent control. In these two higher concentrations, physical abnormalities were recorded in the surviving larvae. For the majority (84%) of the larvae surviving in 0.067 mg/l, and in all survivors in 0.14 mg/l, the caudal (tail) fin was absent. The relative tail length of the other 16% of the larvae in 0.067 mg/l was not affected at all as compared to the control. These effects were completely absent at lower concentrations. The NOEC from this early life stage study is 0.035 mg/l.
The reproduction, expressed as the number of juveniles per container, was also significantly reduced (51%) in 105 mg/kg. The NOEC from this study was, therefore, 45 mg/kg.
3.1.6. Predators In a 90-day study in rats, AHTN was added to the diet at levels of 1.5, 5, 15 or 50 mg/kg. Body weight gain was decreased in males and females at the highest dose of AHTN and there were some haematological effects. The liver weights of females at 15 mg/kg and of both sexes at 50 mg/kg were increased but there were no histopathological findings at either dose indicating that this may represent an increased demand for liver function. There were no differences in liver weights at the end of a 4-week treatment free period. Green colored lachrymal glands were seen in the females at lower doses but with no associated histopathology and pending further investigation, this is not considered an adverse effect. It was concluded, therefore, that the no-observed-adverse-effectlevel in the diet is 15 mg/kg for AHTN (Api and Ford, 1999). 3.2. Toxicity test results for HHCB The results for HHCB are summarised in Table 3 and Table 4.
3.1.4. Earthworms No mortality or growth inhibition of the adults was observed after 4 weeks in concentrations up to and including 250 mg/kg. In the range finding test 100% mortality occurred after 14 days exposure to 1000 mg/kg. Reproduction was not significantly affected up to concentrations of 105 mg/kg (14% inhibition — not statistically significant). At the level of the LOEC (250 mg/kg), the reproduction was 39% of the control.
3.2.1. Algae Mean measured concentrations were 0.042, 0.084, 0.201, 0.466 and 0.854 mg/l. As for AHTN, it was difficult to maintain the concentrations and sometimes pH values rose to above 10. Both biomass production and growth rate were not significantly inhibited up to and including 0.201 mg/l. The inhibition based on Area Under the Curve was 35% in 0.466 mg/l and 56% in 0.854 mg/l. The NOEC was 0.201 mg/l. Growth rate was inhibited by 9 and 20% (both statistically significant) in the highest two concentrations.
3.1.5. Springtails No significant mortality or effects on reproduction were observed in concentrations up to and including 45 mg/kg. Mortality was significant (18%) in the highest concentration of 105 mg/kg.
3.2.2. Daphnia Mean measured concentrations were 0.049, 0.111, 0.205, 0.419 and 0.842 mg/l. Mobility of the parent generation was not affected at concentrations up to and including 0.205 mg/l, whereas
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
89
100% were immobile in the next higher concentration — 0.419 mg/l. At the level of the NOEC (0.111 mg/l), the mean reproduction was inhibited by 21% as compared to the solvent control. At the next higher concentration, 0.205 mg/l (LOEC) the mean reproduction was inhibited by 26%. At 0.419 and 0.842 mg/l, reproduction of the surviving adults was inhibited completely.
days exposure to 1000 mg/kg. Growth was significantly inhibited (15%) in the highest concentration of 250 mg/kg. Reproduction was not significantly affected up to concentrations of 45 mg/kg (7% inhibition). At the level of the LOEC (105 mg/kg), the reproduction was 57% of the control, whereas in the highest concentration reproduction was inhibited completely
3.2.3. Fish In the 21-day growth test with L. macrochirus (Bluegill sunfish) mean measured concentrations were 0.093, 0.182, 0.393, 0.830 and 1.566 mg/l. Survival of the fish was not significantly affected up to and including 0.182 mg/l. However, at this dose (and above) clinical signs of irregular respiration, bottom and tail dominate swimming, loss of equilibrium and righting reflex were observed. Mortality was 10% at the next higher concentration of 0.393 mg/l and coincided with significantly reduced growth. In 0.830 mg/l, mortality reached 100% after 14 days and in 1.566 mg/l at day 2. The 21-day LC50 was 0.452 mg/l. The overall NOEC of the test was 0.093 mg/l as determined by the onset of clinical signs. Mean measured concentrations in the early lifestage test with P. promelas were 0.0091, 0.019, 0.037, 0.068 and 0.140 mg/l. Egg hatchability was not significantly affected in any of the test concentrations. Larval survival after 32 days was not affected in concentrations of 0.068 mg/l and below. In the highest concentration of 0.140 mg/l mean larval survival was 78%. Larval growth was not affected in concentrations of 0.068 mg/l. At 0.140 mg/l, mean length and weight were reduced by 20 and 54%, respectively, as compared to the solvent control. Larvae surviving in the highest concentration (0.140 mg/l) were recorded to be generally smaller, less well developed and appeared less active, exhibiting some erratic swimming behaviour and loss of balance. Therefore, the NOEC for HHCB from this e.l.s. fish study is 0.068 mg/l.
3.2.5. Springtails No significant mortality was observed in soils containing up to 45 mg/kg, but mortality was 72% in 105 mg/kg. Reproduction was inhibited by 23% in 45 mg/kg. However, this inhibition was not statistically significantly different from the control. The reproduction in 105 mg/kg was 16% of the control. Thus, the NOEC for HHCB is 45 mg/kg based on survival.
3.2.4. Earthworms Mortality of the adults was not affected after 4 weeks in concentrations up to 250 mg/kg. In the range finding test mortality was 100% after 14
3.2.6. Predators In a 90-day study with rats, HHCB was added to the diet of rats at levels 5, 15, 50, and 150 HHCB/kg per day. There were no significant effects at any dose level, however, a progressive dose-related decrease in body weight gain and an increase in liver weight accompanied by histopathological changes was seen at higher doses in the 14-day range finding test. Based on this test, it was concluded that the no-adverse-effect-level in the diet is 150 mg/kg (Api and Ford, 1999). 3.3. Deri6ation of PNEC The aquatic and soil toxicity data used to derive the PNEC for the various protection targets are summarised in Table 5. For AHTN and HHCB, studies of a chronic nature are available for algae, Daphnia and fish. Application of an assessment factor of 10 to the 36-day-NOEC of the fish early-life stage test results in PNECwater of 0.0035 mg/l for AHTN and 0.0068 mg/l for HHCB. For soil organisms standard long-term toxicity studies are available for two trophic levels. After normalising for the lower organic matter content in the EU-standard soil and application of an assessment factor of 50 to 4-week reproduction test with springtail, the PNECsoil is 0.32 mg/kg
90
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
(dry weight) for both AHTN and HHCB. For comparison we also derived of PNECsoil from PNECwater through equilibrium partitioning between soil pore water and the soil solid (organic) phase. This results in 4.4 and 9.8 mg/kg for AHTN and HHCB, respectively (Table 6). Comparison of both approaches, shows that PNECsoil based equilibrium partitioning is higher than the value based on experimental data from soil organisms by a factor of 14 and 31 for AHTN and HHCB, respectively. This may be partly explained by the higher assessment factor (50) used for the terrestrial data compared to the aquatic data (10). The residual deviation is less than an order of magnitude implying that the assumption of similar sensitivity of aquatic and soil organisms is not violated. This supports the derivation of PNECsediment by equilibrium partitioning, as presented in Table 6. Yet the reliability of this value is considered lower than one derived directly from toxicity data and it is an indication of PNECsediment only. Additionally, the TGD states that when the log Kow is greater than 5 such as it is for AHTN and HHCB, use of PNECsediment from equilibrium partitioning calls for an increase of the exposure parameter
PECsediment (predicted environmental concentration) by a factor of 10 for the calculation of the PEC/PNEC ratio in order to take ingestion via food into account (EU-TGD). The NOAELs derived from the 90-day oral studies with rats were 15 and 150 mg/kg body weight per day. Similar NOAELs have been observed in other studies with rats with these materials including a developmental toxicity study and a pre- and postnatal developmental study (Ford, 1998). The estimation of PNECpredator, based on the 90 day NOAEL for the rat, is calculated based on a 0.4-kg rat eating 20 g dry food/day as follows: the daily intake of the rat is 15 mg AHTN/kg bwt × 0.4 kg bwt=6 mg AHTN. This is taken up in 20g food and therefore the concentration in food is 300 mg AHTN/kg food. Dividing this by an assessment factor of 30, results in a PNECpredator of 10 mg/kg for AHTN. Similarly, the 150 mg/kg NOAEL for HHCB results in a PNECpredator of 100 mg/kg for HHCB.
3.4. Risk characterisation For the risk characterisation, the exposure of the protection targets is compared to the PNEC.
Table 5 Derivation of PNEC for the protection targets (the lowest NOEC is printed bold) Species
AHTN
HHCB
Algae (mg/l) Daphnia Fish
72-h-NOEC 0.374 21-day-NOEC 0.196 21-day-NOEC 0.089; 36-day-NOEC 0.035
72-h-NOEC 0.201 21-day-NOEC 0.111 21-day-NOEC 0.093; 32-day-NOEC 0.068
Assessment factor PNECwater
10 0.0035
Earthworm (mg/kg) Springtail (mg/kg) Lowest NOEC normalised for organic matter in standard soil (Fomexperiment/Fomstandard = 0.1/0.034)
8-week-NOEC 105 4-week-NOEC 45 16
Assessment factor PNECsoil (mg/kg dw)
50 0.32
Rat (mg/kg bw) Rat converted to daily dose (mg/kg food)
13-week-NOEC 15 13-week-NOEC 300
13-week-NOEC 150 13-week-NOEC 3000
Assessment factor for predator PNECpredator (mg/kg food)
30 10
30 100
10 0.0068 8-week-NOEC 45 4-week-NOEC 45 16 50 0.32
Symbols and basic algorithma
Soila
Sedimenta AHTN
HHCB
AHTN
HHCB
Kp: solids–water partition coefficient Kp = Foc×Koc
Foc =0.02
1260
1450
Foc = 0.05
3160
3620
Total compartment-water partitioning coefficient: Kcomp.-water = Fwatersoil×rwater+Fsolidsoil×Kp×rsolid
Fwatersoil =0.2; Fsolid =0.6; rwater =1; rsolid =2.5; rsoil =1.7
1890
2170
Fwatersed. = 0.8; Fsolid = 0.2; rwater = 1; rsolid = 2.5; rsoil = 1.3
1580
1810
PNECep =Kcomp.–water×PNECwater/rcomp. (mg/kg ww) PNECep= PNECep×CONV (mg/kg dw)
CONVsoil = 1.13
3.9
8.7
4.4
9.8
4.3 CONVsediment = 2.6
11
9.5 25
a comp., compartment (soil or sediment); F, fraction, e.g. Fwatersoil: the fraction of water in the soil compartment; Foc, the fraction of organic carbon in the soil or sediment; r, bulk density; CONV, conversion factor for wet to dry weight.
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
Table 6 Derivation of PNECsoil and PNECsediment according to equilibrium partitioning as described in the EU-TGD
91
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
92 Table 7 RCRs for AHTN and HHCB
Exposure concentrations
PNECa
RCR
AHTN Aquatic organisms (mg/l) Sediment organisms (mg/kg dw) Soil organisms (mg/kg dw) Fish-eating predators (mg/kg fw) Worm-eating predators (mg/kg fw)
0.30 0.48 0.029 0.12 0.065
(m) (m)*10b (m+p) (m) (m+p)
3.5 11 0.32 10 10
0.086 0.44 0.091 0.012 0.007
HHCB Aquatic organisms (mg/l) Sediment organisms (mg/kg dw) Soil organisms (mg/kg dw) Fish-eating predators (mg/kg fw) Worm-eating predators (mg/kg fw)
0.50 0.16 0.032 0.12 0.099
(m) (m)*10b (m+p) (m) (p)
6.8 25 0.32 100 100
0.074 0.064 0.10 0.001 0.001
a b
(m), (p), measured or predicted concentration; (m+p), predicted from measured concentration. Measured concentration multiplied by 10 as log Kow of AHTN and HHCB is above 5 (TGD).
Risk is expressed as the risk characterisation ratio (RCR) of exposure to no-effect concentrations for aquatic and sediment-dwelling organisms and their predators, and terrestrial organisms and their predators. Even though the PNECs are based on the lowest NOECs divided by an assessment factor reflecting the robustness of the database and the PECs are based on 90th-percentile measured environmental concentrations or concentrations estimated based on conservative models, it is the goal to obtain RCRs less than one. Exposure concentrations used in determining the RCRs are taken from part I of this risk assessment (Balk and Ford, 1999) while the PNEC values are taken from Tables 5 and 6. The results are presented in Table 7. For AHTN as well as for HHCB, the risk ratios are below 1 for all protection targets. All risk ratios are at or below 0.1 except for the risk ratio of AHTN for sediment organisms. The risk ratios for predators on aquatic and soil organisms are all at or below 0.01 for AHTN as well as for HHCB.
4. Discussion In the EU-TGD quantitative structure-activity relations (QSAR) are given for baseline or mini-
mum toxicity for chronic endpoints for fish (Brachydanio rerio/P. promelas), D. magna and Selenastrum capricornutum. For fish, these QSARs predict the prolonged LC50 correctly and the predicted prolonged NOEC is below the experimental results by factors of 3–10. For D. magna the predicted 16-day-NOEC is too low by a factor of 50, whereas the EC50 for algae is overestimated by a factor of 15–35. In particular the estimates of NOEC-values for AHTN and HHCB are unreliable. The tested aquatic species represent three trophic levels (primary producer, primary and secondary consumer), three taxonomic groups (green algae, crustaceans and bone fish) and three feeding strategies (phototrophic, herbivorous filter feeding and carnivorous). Fish seem to be only slightly more sensitive than the other two species. Comparison of the results of the 21-day-growth test and the 36-day-ELS test for fish shows that the growth and development of fish early life stages seems only marginally more sensitive to these substances than growth in later stages. The physical abnormalities of the fish larvae with AHTN and the underdevelopment of the larvae with HHCB occurred at concentrations where growth was affected as well as survival. In other words, where these effects are observed, the concentrations have reached lethal thresholds. The
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
mechanism for the development of the missing caudal fin with AHTN is unknown, but currently this is subject to further investigations. Preliminary results show that the effect is repeatable in another fish species at the same concentration level. The two soil organisms tested are equally sensitive for AHTN and HHCB. On the basis of equilibrium partitioning their sensitivity seems to be similar to that of aquatic organisms, implying that additional exposure by food ingestion does not play a role. This justifies the use of equilibrium partitioning for the prediction of the toxicity for sediment-dwelling organisms but at the same time shows that the additional safety factor of 10 related to exposure by sediment ingestion, as called for in the TGD for substances with log Kow values greater than 5, is too conservative. The risk assessment presented here is the result of an iterative process where initially the exposure was based on model predictions according to the EU-TGD and the effect assessment was based on aquatic toxicity data only. Triggered by risk ratios above 1, additional data were generated to replace initial predictions based of default values and high assessment and safety factors by empirical data, resulting in more reliable risk evaluations. This also explains why the RCRs presented here deviate from those in an earlier report (Van de Plassche and Balk, 1997). In the present phase, the risk assessments still include various conservative assumptions, in particular in the estimation of the exposure concentrations and toxicity levels for soil and sediment organisms. In the exposure analysis the biodegradability of the substances in soil is taken into consideration with a conservative estimate of the half-life time in soil, but the biodegradability in the sediment was neglected. In addition, the derivation of the PNECsoil includes an extra safety factor of 3 to account for the assumed difference in bioavailability related to the organic content of the soil. Due to the derivation of PNECsediment by equilibrium partitioning an extra safety factor of 10 is included in the sediment risk ratio. The fact that in spite of the conservative assumptions all risk ratios are well below 1, is reassuring.
93
From a risk-assessment point-of-view, priority should be given to refine elements in the risk assessment with the highest uncertainty, for the compartment with the highest RCR. Improved reliability of a risk assessment may be obtained along two lines: by refining the exposure assessment and by increasing the toxicity database. In particular the assessments for the soil and sediment compartments would benefit from a refinement of the input data. Measurement of actual concentrations of AHTN and HHCB in plots with a known history of repeated sludge amendments and measurements of sediment in surface water would eliminate uncertainties introduced by the use of model predictions. Extension of the soil toxicity database with a third test species, e.g. a plant, would result in a more reliable PNEC, which is reflected by a lower assessment factor of 10 instead of 50. If toxicity data for sediment dwelling organisms were available, the present extra safety factor of 10 related to the PNEC derived on the basis of equilibrium partitioning would be abandoned. However, the present low risk ratios for these substances in all compartments do not seem to justify efforts for further refinement.
5. Conclusions The polycyclic musks AHTN and HHCB are less toxic to aquatic organisms than expected on the basis of ‘minimum toxicity’ as predicted by QSAR methods. The NOEC for fish, daphnia and algae range from 35 to 280 mg/l for AHTN and from 70 to 200 mg/l for HHCB. Fish are slightly more sensitive than the other species. Fish embryo development is affected at concentrations just below and at the lethal level. The PNECs derived for aquatic organisms are 3.5 and 7 mg/l for AHTN and HHCB, respectively. The NOEC for the terrestrial worm and springtail range from 45 to 105 mg/kg. These results compare well to the NOEC for soil predicted from the aquatic toxicity on the basis of equilibrium partitioning. However, corrections for soil composition related to the assumed influence of bioavailability and the limited number of tested
94
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
species introduce additional safety factors and high assessment factors, resulting in PNECsoil of 0.32 mg/kg for both substances. Sediment toxicity data are not available and PNECsediment is derived by equilibrium partitioning from PNECwater: 11 and 25 mg/kg dw for AHTN and HHCB, respectively. Because of lack of data for wildlife species, the toxicity for predators in the aquatic and terrestrial ecosystem is extrapolated from a 90-day study for the rat. The PNECpredator values are 10 and 100 mg/kg food (fresh weight) for AHTN and HHCB, respectively. The risk expressed as a characterisation ratio (PEC/PNEC) is well below 1 for all compartments. All risk ratios are at or below 0.1 except the risk ratio of AHTN for sediment organisms. The risk ratios for predators on aquatic and soil organisms are all at or below 0.01 for AHTN as well as for HHCB. Keeping in mind the largely conservative approach taken for soil and sediment organisms, these data are reassuring that these substances will not pose an environmental risk.
Acknowledgements Results of an initial risk assessment for AHTN and HHCB have been discussed in a task force consisting of T.J. van Bergen, H.-D. Gaisser and J.D. Middleton of NEA, E.J. van de Plassche and P.T.J. van de Zandt (National Institute of Public Health and the Environment (RIVM) and the authors.
References Api, A.M., Ford, R.A., 1999. Evaluation of the oral subchronic toxicity of AHTN and HHCB in the rat, Toxicol. Lett., (submitted). Balk, F., Ford, R.A., 1999. Environmental risk assessment for the polycyclic musks AHTN and HHCB in the EU, I. fate and exposure assessment. Toxicol. Lett., 111, 57–79. Croudace, C.P., Caunter, J.E., Johnson, P.A., Wallace, S.J., 1997a. AHTN: Chronic toxicity to fathead minnow
.
(Pimephales promelas) embryos and larvae. Report to RIFM, Zeneca Project Report BL5933/B. Croudace, C.P., Caunter, J.E., Johnson, P.A., 1997b. HHCB: Chronic toxicity to fathead minnow (Pimephales promelas) embryos and larvae. Report to RIFM, Zeneca Project Report BL5934/B. EC, 1996. Technical Guidance Documents in support of Directive 96/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC catalogue numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 rue Mercier, L-2965 Luxembourg. Ford, R.A., 1998. The human safety of the polycyclic musks AHTN and HHCB in fragrances — a review. Dt. Lebensm.-Rdsch 94 (8), 245 – 286. Gossmann, A., 1997a. Effects of AHTN on reproduction and growth of earthworms Eisenia fetida in artificial soil. Report to RIFM, IBACON Project no. 1782022. Gossmann, A., 1997b. Effects of HHCB on reproduction and growth of earthworms Eisenia fetida in artificial soil. Report to RIFM, IBACON Project no. 2032022. Hopkins, M.N., Lambert, A.H., 1996. AHTN: 13-week oral (dietary) toxicity in the rat with a 4-week treatment-free period. Report RFA/5/95 to RIFM. Klepka, S., 1997a. Effects of AHTN on reproduction of the springtail Folsomia candida in artificial soil. Report to RIFM, IBACON Project no. 1781016. Klepka, S., 1997b. Effects of HHCB on reproduction of the springtail Folsomia candida in artificial soil. Report to RIFM, IBACON Project no. 2031016. Van de Plassche, E.J., Balk, F., 1997. Environmental risk assessment of the polycyclic musks AHTN and HHCB according to the EU-TGD. RIVM report 601503 008. National Institute of Public Health and the Environment, Bilthoven, NL. Van Dijk, A., 1997a. Acute toxicity of AHTN to Pseudokirchneriella subcapitata. Report to RIFM, RCC Umweltchemie AG Project 380654. Van Dijk, A., 1997b. Acute toxicity of HHCB to Pseudokirchneriella subcapitata. Report to RIFM, RCC Umweltchemie AG Project 380632. Wu¨thrich, V., 1996a. Influence of AHTN on the reproduction of Daphnia magna. Report to RIFM, RCC Umweltchemie AG Project 380665. Wu¨thrich, V., 1996b. AHTN: 21-Day prolonged toxicity study in the bluegill sunfish under flow-through conditions. Report to RIFM, RCC Umweltchemie AG Project 380698. Wu¨thrich, V., 1996c. Influence of HHCB on the reproduction of Daphnia magna. Report to RIFM, RCC Umweltchemie AG Project 380687. Wu¨thrich, V., 1996d. HHCB: 21-Day prolonged toxicity study in the bluegill sunfish under flow-through conditions. Report to RIFM, RCC Umweltchemie AG Project 380711.
Environmental risk assessment for the polycyclic musks, AHTN and HHCB II. Effect assessment and risk characterisation Froukje Balk a, Richard A. Ford b,1,* a b
HASKONING Consulting Engineers and Architects, P.O. Box 151, 6500 AD Nijmegen, The Netherlands Research Institute for Fragrance Materials, Suite 406, Two Uni6ersity Plaza, Hackensack, NJ 07601, USA Received 30 March 1999; received in revised form 16 August 1999; accepted 16 August 1999
Abstract Reports of the polycyclic musks AHTN and HHCB in surface water and fish, primarily in Europe, have prompted studies of their environmental effects. These effects then are used, along with the predicted environmental concentrations in a risk assessment according to the approach developed under European Union Regulation 793/93, in line with the Technical Guidance Document for risk assessment of new and existing chemicals. In 72-h studies with algae (Pseudokirchneriella subcapitata), NOECs were 0.374 mg/l (AHTN) and 0.201 mg/ l (HHCB). In 21-day reproductive tests with daphnia (Daphnia magna) NOECs were 0.196 (AHTN) and 0.111 mg/l (HHCB). In 21-day growth tests with bluegill sunfish (Lepomis macrochirus), NOECs were 0.067 (AHTN) and 0.068 mg/l (HHCB). And, finally 35-day early life stage tests with fathead minnows (Pimephales promelas) resulted in NOECs of 0.035 (AHTN) and 0.068 mg/l (HHCB). These results lead to Predicted No Effect Concentrations (PNEC) of 3.5 mg/l (AHTN) and 6.8 mg/l (HHCB) for aquatic organisms. For the soil compartment, 8-week studies with earthworms (Eisenia fetida) resulted in NOECs of 105 (AHTN) and 45 mg/kg (HHCB) and 4-week studies with springtails (Folsomia candida) resulted in a NOECs of 45 mg/kg for both substances. These values lead to a PNEC of 0.32 mg/kg dw for both materials. Using mammalian studies, PNECs for fish or worm eating predators of 10 mg/kg fw (AHTN) and 100 mg/kg fw (HHCB) can be derived. For sediment dwelling organisms, PNECs were derived by equilibrium partitioning using the aquatic PNECs. Comparing PNECs with the measured or predicted environmental exposures leads to risk characterisation ratios as follows: aquatic species: AHTN 0.086, HHCB 0.074; sediment organisms: AHTN 0.44, HHCB 0.064; soil organisms: AHTN 0.091, HHCB 0.10; fish eating predators: AHTN 0.012, HHCB 0.001; worm eating predators: AHTN 0.007, HHCB 0.001. © 1999 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: AHTN; HHCB; Musks; Ecotoxicity; Alga; Daphnia; Fish; Earthworm; Springtail; Risk assessment
* Corresponding author. Present address: 7 Alwyn Ave., London W4 4PA, UK. Tel.: +44-181-987-8626; fax: +44-181-987-8627. E-mail addresses: [email protected] (F. Balk), [email protected] (R.A. Ford) 1 Please address correspondence to RIFM. 0378-4274/99/$ - see front matter © 1999 Published by Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 4 2 7 4 ( 9 9 ) 0 0 1 7 0 - 8
82
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
1. Introduction The polycyclic musks 7-acetyl-1,1,3,4,4,6-hexamethyl-1,2,3,4-tetrahydronaphthalene (AHTN) and 1,3,4,6,7,8-Hexahydro-4,6,6,7,8,8-hexamethylcyclopenta-g-2-benzopyran (HHCB) are widely used as fragrances ingredients in detergents and other household products. This, combined with high log Kow values and lack of ready biodegradability, has resulted in their widespread occurrence in the environment (Balk and Ford, 1999). AHTN and HHCB together represent about 95% of the total market volume for the class of fragrance ingredients known as the polycyclic musks. In view of reported levels in surface water and fish primarily in Europe, the Research Institute of Fragrance Materials (RIFM) initiated an environmental risk evaluation according to the approach developed under European Union Regulation 793/93, in line with the Technical Guidance Document for risk assessment of new and existing chemicals (EU-TGD; EC, 1996). In an environmental risk assessment, the environmental exposure is compared to data on environmental effects. The exposure of the environment to AHTN and HHCB is presented in a separate paper (Balk and Ford, 1999, part I) in this issue. Because the reported occurrences of these materials are in water and fish, these were considered the most important issues to address. However, these occurrences along with measured values in sludge, allow prediction of exposures of soil and sediment-dwelling organisms and of predators feeding on these organisms or on fish. Thus, the Table 1 Characteristics of AHTN and HHCB
risks for these endpoints are also addressed. This paper reports on the environmental effects of the AHTN and HHCB on the basis of a series of ecotoxicity studies on both materials that have been carried out on behalf of RIFM, and presents the environmental risk assessment.
2. Materials and methods The structure and some relevant characteristics of AHTN and HHCB are given in Table 1. Because of the low water solubilities of AHTN and HHCB, it was necessary in the aquatic toxicity tests described below to prepare stock solutions using DMF as a solvent and Tween 80 as a dispersant or using triethylene glycol as a solvent. These stock solutions were then diluted to reach the desired concentrations in the tests. The residual level of the solvent in the test vessel was always below the maximum level allowed by the test guidelines. Solvent controls containing equivalent levels of the solvents were used in all cases along with undosed water controls. These substances have a strong tendency to sorb, thereby reducing the concentrations in solutions over time. Final concentrations in water were determined by HPLC or GC and the test concentrations are expressed as measured concentrations. Except for the highest concentration in the fish growth test, the tested concentrations did not exceed the water solubility limit. Details are summarised in Tables 2 and 3. Stock solutions for the terrestrial toxicity studies were prepared with acetone as the solvent. After thorough mixing with the soil, the acetone was then allowed to evaporate. Final concentrations in the test medium were not determined but are based on the known amounts of test substance added to the soil. All toxicity tests and chemical analyses were carried out and inspected according to the Principles of Good Laboratory Practice.
2.1. Algae The toxicity to algae was studied in a static test according to OECD Test Guideline 201 with
Table 2 Aquatic toxicity of AHTN Resultsa (mg/l)
Remarksb
P. subcapitata c 72-h static (Van Dijk, 1997a)
Test A: NOECg =0.438; NOECb =0.204; LOECbd =0.438; EgC50\0.797; EbC50 =0.468 0.434–0.508. Test B: NOEC = 0.374; LOECd =0.835; EgC50\0.835; EbC50 :0.835.
D. magna 21-day semi-static (Wu¨thrich, 1996a)
NOEC(rep) = 0.196; LOECd =0.401; ErC50-21-day = 0.244 0.239–0.249; 21-day-IC50 =0.341 (mobility); 0.243–0.433. NOECgrowth= 0.089; LOECd =0.184; 21-day-LC50 =0.314 0.226-0.448.
Carrier: 0.005% DMF and 0.005% Tween 80; n=5. Test A: HPLC identification; Start conc. 81–90% of nominal; End conc. 31–85% of nominal. Test B: start conc. 77–90% of nominal; end conc. 53–142% of nominal; geometric mean NOEC= 0.276. Carrier: 0.008% DMF and 0.002% Tween 80; n= 5; HPLC identification; conc. fresh 84–103% of nominal; conc. used 70–85% of nominal. Carrier: 0.005% DMF and 0.005% Tween 80; n= 5; HPLC identification; conc. 57–109% of nominal. Solvent triethylene glycol; GC identification; Conc. 55–108% of nominal.
Bluegill sunfish; L. macrochirus; 21-day flow-through (Wu¨thrich, 1996b)
Fathead minnow; P. promelas; 32 days post LOEChatch\0.140; NOECsurv. =0.067; hatch; 36 days overall (Croudace et al., 1997a) LOECsurv.= 0.140; 32-day-LC50 =0.100 0.097–0.100; NOECgrowth =0.035; LOECgrowthe =0.067; NOECdevelop. =0.035; LOECdevelop. = 0.067. 95% confidence limits. The number of concentrations tested (n) excludes control and solvent control. c Former name S. capricornutum. d Dunnet’s test (P= 0.05). e Wilcoxon rank sum test (P= 0.05). a
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
Test and reference
b
83
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
84 Table 3 Aquatic toxicity of HHCB Test and reference
Resultsa (mg/l)
Remarksb
P. subcapitata c; 72-h static (Van Dijk, 1997b)
NOEC= 0.201; LOECd =0.466; EgC50\0.854; EbC50 = 0.723 0.678-0.778.
D. magna; 21-day semi-static (Wu¨thrich, 1996c)
NOEC(rep)= 0.111, LOECd =0.205; ErC50-21-day= 0.282 0.260–0.312; IC50-21-day= 0.293 (mobility) 0.204–0.419e. NOEC =0.093 (clinical signs); LOECd = 0.182; NOEC(growth)= 0.182; LC50-21-day=0.452 0.316–0.911. LOEChatch\0.140; NOECsurv. =0.068; LOECsurv.= 0.140; 32-day-LC50\0.140; NOECgrowth=0.068; LOECgrowthf =0.140; NOECdevelop. = 0.068; LOECdevelop.= 0.140.
Carrier: 0.005% DMF and 0.005% Tween 80; n = 6; HPLC identification; start conc. 71–102% of nominal; end conc. 54–85% of nominal. Carrier: 0.008% DMF and 0.002% Tween 80; n =5; HPLC identification; conc. fresh 82–104% of nominal; conc. used 63–91% of nominal. Carrier: 0.005% DMF and 0.005% Tween 80; n = 5; HPLC identification; conc. 66–86% of nominal.
Bluegill sunfish; L. macrochirus; 21-day flow-through (Wu¨thrich, 1996d)
Fathead minnow; P. promelas; 32 days post hatch; 36 days overall (Croudace et al., 1997b)
Solvent triethylene glycol; GC identification; conc. 50–104% of nominal.
95% confidence limits. The number of concentrations tested (n) excludes control and solvent control. c Former name S. capricornutum. d Dunnet’s test (P = 0.05). e Estimated 95% confidence limits after data reported by Wu¨thrich (1996c). f Wilcoxon rank sum test (P= 0.05). a
b
Pseudokirchneriella subcapitata (Van Dijk, 1997a,b). Nominal concentrations ranged from 0.0625 to 1.0 mg/l (step size 2). The NOEC and EC50 are based on growth rate and biomass production after 72 h. Concentrations were measured at the start and end of the test. The test with AHTN was carried out twice (tests A and B) due to the sharp concentration decrease in the first study with the goal to confirm the observed toxicity levels in a second study.
2.2. Daphnia For Daphnia magna, a semi-static 21-day toxicity test was carried out according to OECD Test Guideline 202, part II, proposed updated version of June 1993 (Wu¨thrich, 1996a,c). The test medium was refreshed three times per week. Nominal concentrations ranged from 0.062 to 1.0 mg/l (step size 2). Concentrations were measured
at the start and end of the first and last exposure period.
2.3. Fish A 21-day prolonged toxicity test was carried out with bluegill sunfish (Lepomis macrochirus) according to OECD Test Guideline 204 under flow-through conditions (Wu¨thrich, 1996b,d). Test parameters included fish growth (weight and length). The fish weights at the start of the experiment varied between 1.4 and 1.6 g. The concentrations ranged from 0.125 to 2.0 mg/l (step size 2). Concentrations were measured at the start, halfway through and at the end of the test period. An early life stage test was carried out with fathead minnow (Pimephales promelas) according to OECD Test Guideline 210 under flow-through conditions (Croudace et al., 1997a,b). Eggs less than 24 h old were exposed to nominal concentra-
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
tions ranging from 0.0125 to 0.2 mg/l (step size 2). Concentrations were measured 13 times at regular intervals during the 36-day test period.
2.4. Earthworm The earthworm test was carried out according to ISO 11268 (Gossmann, 1997a,b). Adult worms (Eisenia fetida) were exposed to nominal concentrations in soil of 8, 19, 45, 105 and 250 mg/kg. The test medium was an artificial soil according to ISO-Standard 11268-1 and OECD Test Guideline 207, containing 10% Sphagnum peat, 20% kaolinite clay, approximately 70% fine quartz-sand (grain size 0.1–0.5 mm) and 0.5% calcium carbonate to adjust to pH 6.0 90.5. The test materials at appropriate concentrations were dissolved in equal amounts of acetone, mixed with the quartz sand and allowed to slowly evaporate. This sand was then thoroughly mixed with the other components to make the standardised soil. The control soil was treated in the same manner but with pure acetone. After preparation of the test concentrations and an equilibrium period of 1 week, the test organisms were added to the soil. The test medium was not refreshed during the test period. Weights of the adult worms ranged between 340 and 540 mg, but did not differ more than 100 mg within this range in each test container. The worms were fed weekly with finely ground cattle manure. Adult worms were removed after 4 weeks of exposure, counted and weighed. The remaining offspring remained in the test containers for another 4 weeks.
2.5. Springtails The springtail test was carried out according to the draft ISO/CD 11267 (Klepka, 1997a,b). Juvenile springtails of the species Folsomia candida 10 – 12 days of age were placed in an artificial soil and survival and reproduction after 28 days were determined. The artificial soil was the same as used in the earthworm study. Nominal test concentrations were 1, 3, 8, 19, 45 and 105 mg/kg soil. After preparation of the test concentrations and an equilibrium period of 1 week, the test organisms were added to the soil. The animals were fed with granulated dry yeast.
85
2.6. Predators Toxicity studies with AHTN and HHCB on predators present in the environment like fish-eating birds and mammals are not available. Therefore, toxicity data for laboratory mammals are used to derive a PNEC for these predators. For AHTN and HHCB, subchronic oral studies are available for rats (Hopkins and Lambert, 1996; Api and Ford, 1999). AHTN was administered in the diet of rats (15 males and 15 females per group) at daily doses of 1.5, 5, 15 and 50 mg/kg for 13 weeks. HHCB was tested by an identical protocol but at doses of 5, 15, 50 and 150 mg/kg per day.
2.7. Deri6ation of the PNEC In the EU-TGD, the PNEC (Predicted No Effect Concentration) is defined as the concentration below which unacceptable effects on organisms will most likely not occur. The environmental protection goals considered here are the aquatic ecosystem (PNECwater, PNECsediment), the terrestrial ecosystem (PNECsoil) and top predators (PNECpredator). The PNECwater was obtained using the assessment factors applied to the lowest NOEC from the aquatic toxicity studies, as described in the TGD. In this case where three or more chronic studies are available for algae, daphnia and fish, an assessment factor of 10 may be applied to the lowest NOEC to derive PNECwater. For the derivation of a PNECsoil based on two chronic studies with soil organisms, the EU-TGD proposes an assessment factor of 50 to be applied to the lowest NOEC. Before doing so, the TGD states that the results from the soil toxicity tests should be normalised using a relationship that describes the bioavailability of the test substance in the soil. The defined standard soil of the TGD contains 3.4% organic matter, whereas the tests presented here were carried out in the standard OECD soil with an organic matter content of 10%. Therefore the test results are normalised according to the formula for organic substances (EU-TGD):
86
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
NOECstandard soil =
3. Results
NOECexperiment ×Fomstandard soil/Fomexperiment
3.1. Toxicity test results for AHTN
where: NOECstandard soil is the normalised NOEC for the defined standard soil with 3.4% organic matter; NOECexperiment is the NOEC from experiment (Table 4); Fomstandard soil is the weight fraction of organic matter in soil solids of the standard soil (3.4%); and Fomexperiment is the weight fraction of organic matter in soil solids in tests (10%). This has the practical effect of dividing the observed NOEC by an additional 3, in addition to the assessment factor of 50 for a total factor of 150. PNECsoil and PNECsediment can also be derived from PNECwater on the basis of equilibrium partitioning between the solid and the aqueous phases. According to the equilibrium partitioning theory it is assumed that test organisms are exposed only through the pore water and that soil or sediment organisms and aquatic organisms are equally sensitive to the test substances. Assuming a concentration in the porewater equal to PNECwater the corresponding equilibrium concentration in the soil is estimated using the soil – water partition coefficient Ksoil – water. Ksoil – water is estimated on the basis of sorption to organic carbon in the soil and a correction factor for the soil density (EU-TGD). The general applicability of the equilibrium partitioning method has been tested less for soil than for sediment-dwelling organisms. The EU-TGD considers the approach as a screening for assessment of the risk for both organisms. This approach is presented in more detail in Table 6. For the estimation of a PNECpredator for fishand worm-eating birds and mammals, the NOEC determined for the rat is used as the starting point. For the conversion of a NOEC obtained by daily dosing (mg/kg body weight) to a NOEC level in the food (mg/kg food) a factor of 20 is used for a rat greater than 6 weeks old (EU-TGD) based on the fact that the food consumption of an adult rat is about 20 mg/kg body weight per day. According to the TGD the assessment factor to be applied to such a NOEC in the derivation of the PNEC from a 90-day toxicity test is 30.
The results for AHTN are summarised in Table 2 and Table 4.
3.1.1. Algae Mean measured concentrations were 0.035, 0.088, 0.204, 0.438 and 0.797 mg/l for test A and 0.0679, 0.140, 0.170, 0.374 and 0.835 mg/l for test B, but maintenance of the concentrations proved to be difficult and in some cases pH values rose to above 10 indicating an insufficient buffering capacity of the medium. Growth was not significantly inhibited in concentrations up to and including (A) 0.204 and (B) 0.374 mg/l. In the next higher concentration, the LOEC, (A) 0.438 and (B) 0.835 mg/l, the biomass production was inhibited by (A) 36% and (B) 54%, measured as Area Under the Curve. Growth rate (m) was the less sensitive parameter. Inhibition at 0.835 mg/l (B) was 16% (LOEC). In test (A), the NOEC for growth rate was higher, 0.438 mg/l, whereas the inhibition at 0.797 mg/l (LOEC) was 33%. The geometric mean NOEC for both tests was 0.28 mg/l. 3.1.2. Daphnia Mean measured concentrations were 0.054, 0.113, 0.196, 0.401 and 0.804 mg/l. Immobility of the parent generation was 80% at 0.401 mg/l and the 21 day-EC50 was 0.341 mg/l. The mean reproduction in test concentrations up to and including 0.196 mg/l ranged between 80 and 114% as compared to the solvent control. In the next higher test concentration of 0.401 mg/l, reproduction of the surviving adults was inhibited almost completely. Thus, 0.196 mg/l can be considered the NOEC in this study. 3.1.3. Fish Mean measured concentrations during the 21day test with L. macrochirus (Bluegill sunfish) were 0.089, 0.184, 0.392, 1.00 and 2.22 mg/l. Concentrations up to and including 0.184 mg/l did not significantly affect survival of the fish. Mortality was 70% at the next higher concentra-
Test and reference
Results for AHTN
Results for HHCB
Remarks
Earthworm; E. fetida; ISO 11268 (OECD 207) (Gossmann, 1997a,b)
8-week-NOEC =105 mg/kg; LOECa =250 mg/kg; reproduction and food consumption; 4-week-NOEC]250 mg/kg; mortality and growth. 4-week-NOEC =45 mg/kg; LOECc =105 mg/kg; mortality and reproduction.
8-week-NOEC = 45 mg/kg; LOECa = 105 mg/kg; reproduction and food consumption; 4-week-NOECgrowth = 105 mg/kg; LOEC = 250 mg/kg; 4-week-NOECsurvival]250 mg/kg. 4 week-NOEC=45 mg/kg; LOECc = 105 mg/kg; mortality and reproduction.
Initial weight adults 0.34–0.54 g; test range 8–250 mg/kg; solvent: acetone; artificial soil pH 6.1; 10% sphagnum DINb; temp. 17–23°C.
Springtail; F. candida; ISO/CD 11267 (Klepka, 1997a,b)
a
Dunnet’s test (P =0.05). Sphagnum DIN standard: organic material minimum 90%, organic carbon 52%. c Student’s t-test (P=0.05). b
10–12-day-old juveniles; test range 1–105 mg/kg; solvent: acetone; temperature 17–25°C; artificial soil; 10% sphagnum DINb.
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
Table 4 Toxicity data for AHTN and HHCB for soil organisms
87
88
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
tion of 0.392 mg/l and 100% in 1.00 mg/l after 11 days and in 2.22 mg/l after 4 days. The 21-day LC50 was 0.314 mg/l. Clinical signs such as loss of equilibrium, enhanced and/or irregular respiration and cessation of food intake were observed before the onset of death. Fish growth was significantly reduced at 0.184 mg/l. Therefore the NOEC is 0.089 mg/l. In the early life-stage test with P. promelas (fathead minnow) the mean measured concentrations were 0.008, 0.018, 0.035, 0.067 and 0.140 mg/l. Hatchability of eggs was not significantly affected in any of the test concentrations. Larval survival after 32 days was not affected in concentrations of 0.067 mg/l and below. In 0.14 mg/l larval survival was 18%. Larval growth was not affected in concentrations of 0.035 mg/l or below. At 0.067 and 0.14 mg/l, mean standard lengths were reduced by 7 and 38%, whereas weights were reduced by 7 and 75%, respectively, as compared to the solvent control. In these two higher concentrations, physical abnormalities were recorded in the surviving larvae. For the majority (84%) of the larvae surviving in 0.067 mg/l, and in all survivors in 0.14 mg/l, the caudal (tail) fin was absent. The relative tail length of the other 16% of the larvae in 0.067 mg/l was not affected at all as compared to the control. These effects were completely absent at lower concentrations. The NOEC from this early life stage study is 0.035 mg/l.
The reproduction, expressed as the number of juveniles per container, was also significantly reduced (51%) in 105 mg/kg. The NOEC from this study was, therefore, 45 mg/kg.
3.1.6. Predators In a 90-day study in rats, AHTN was added to the diet at levels of 1.5, 5, 15 or 50 mg/kg. Body weight gain was decreased in males and females at the highest dose of AHTN and there were some haematological effects. The liver weights of females at 15 mg/kg and of both sexes at 50 mg/kg were increased but there were no histopathological findings at either dose indicating that this may represent an increased demand for liver function. There were no differences in liver weights at the end of a 4-week treatment free period. Green colored lachrymal glands were seen in the females at lower doses but with no associated histopathology and pending further investigation, this is not considered an adverse effect. It was concluded, therefore, that the no-observed-adverse-effectlevel in the diet is 15 mg/kg for AHTN (Api and Ford, 1999). 3.2. Toxicity test results for HHCB The results for HHCB are summarised in Table 3 and Table 4.
3.1.4. Earthworms No mortality or growth inhibition of the adults was observed after 4 weeks in concentrations up to and including 250 mg/kg. In the range finding test 100% mortality occurred after 14 days exposure to 1000 mg/kg. Reproduction was not significantly affected up to concentrations of 105 mg/kg (14% inhibition — not statistically significant). At the level of the LOEC (250 mg/kg), the reproduction was 39% of the control.
3.2.1. Algae Mean measured concentrations were 0.042, 0.084, 0.201, 0.466 and 0.854 mg/l. As for AHTN, it was difficult to maintain the concentrations and sometimes pH values rose to above 10. Both biomass production and growth rate were not significantly inhibited up to and including 0.201 mg/l. The inhibition based on Area Under the Curve was 35% in 0.466 mg/l and 56% in 0.854 mg/l. The NOEC was 0.201 mg/l. Growth rate was inhibited by 9 and 20% (both statistically significant) in the highest two concentrations.
3.1.5. Springtails No significant mortality or effects on reproduction were observed in concentrations up to and including 45 mg/kg. Mortality was significant (18%) in the highest concentration of 105 mg/kg.
3.2.2. Daphnia Mean measured concentrations were 0.049, 0.111, 0.205, 0.419 and 0.842 mg/l. Mobility of the parent generation was not affected at concentrations up to and including 0.205 mg/l, whereas
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
89
100% were immobile in the next higher concentration — 0.419 mg/l. At the level of the NOEC (0.111 mg/l), the mean reproduction was inhibited by 21% as compared to the solvent control. At the next higher concentration, 0.205 mg/l (LOEC) the mean reproduction was inhibited by 26%. At 0.419 and 0.842 mg/l, reproduction of the surviving adults was inhibited completely.
days exposure to 1000 mg/kg. Growth was significantly inhibited (15%) in the highest concentration of 250 mg/kg. Reproduction was not significantly affected up to concentrations of 45 mg/kg (7% inhibition). At the level of the LOEC (105 mg/kg), the reproduction was 57% of the control, whereas in the highest concentration reproduction was inhibited completely
3.2.3. Fish In the 21-day growth test with L. macrochirus (Bluegill sunfish) mean measured concentrations were 0.093, 0.182, 0.393, 0.830 and 1.566 mg/l. Survival of the fish was not significantly affected up to and including 0.182 mg/l. However, at this dose (and above) clinical signs of irregular respiration, bottom and tail dominate swimming, loss of equilibrium and righting reflex were observed. Mortality was 10% at the next higher concentration of 0.393 mg/l and coincided with significantly reduced growth. In 0.830 mg/l, mortality reached 100% after 14 days and in 1.566 mg/l at day 2. The 21-day LC50 was 0.452 mg/l. The overall NOEC of the test was 0.093 mg/l as determined by the onset of clinical signs. Mean measured concentrations in the early lifestage test with P. promelas were 0.0091, 0.019, 0.037, 0.068 and 0.140 mg/l. Egg hatchability was not significantly affected in any of the test concentrations. Larval survival after 32 days was not affected in concentrations of 0.068 mg/l and below. In the highest concentration of 0.140 mg/l mean larval survival was 78%. Larval growth was not affected in concentrations of 0.068 mg/l. At 0.140 mg/l, mean length and weight were reduced by 20 and 54%, respectively, as compared to the solvent control. Larvae surviving in the highest concentration (0.140 mg/l) were recorded to be generally smaller, less well developed and appeared less active, exhibiting some erratic swimming behaviour and loss of balance. Therefore, the NOEC for HHCB from this e.l.s. fish study is 0.068 mg/l.
3.2.5. Springtails No significant mortality was observed in soils containing up to 45 mg/kg, but mortality was 72% in 105 mg/kg. Reproduction was inhibited by 23% in 45 mg/kg. However, this inhibition was not statistically significantly different from the control. The reproduction in 105 mg/kg was 16% of the control. Thus, the NOEC for HHCB is 45 mg/kg based on survival.
3.2.4. Earthworms Mortality of the adults was not affected after 4 weeks in concentrations up to 250 mg/kg. In the range finding test mortality was 100% after 14
3.2.6. Predators In a 90-day study with rats, HHCB was added to the diet of rats at levels 5, 15, 50, and 150 HHCB/kg per day. There were no significant effects at any dose level, however, a progressive dose-related decrease in body weight gain and an increase in liver weight accompanied by histopathological changes was seen at higher doses in the 14-day range finding test. Based on this test, it was concluded that the no-adverse-effect-level in the diet is 150 mg/kg (Api and Ford, 1999). 3.3. Deri6ation of PNEC The aquatic and soil toxicity data used to derive the PNEC for the various protection targets are summarised in Table 5. For AHTN and HHCB, studies of a chronic nature are available for algae, Daphnia and fish. Application of an assessment factor of 10 to the 36-day-NOEC of the fish early-life stage test results in PNECwater of 0.0035 mg/l for AHTN and 0.0068 mg/l for HHCB. For soil organisms standard long-term toxicity studies are available for two trophic levels. After normalising for the lower organic matter content in the EU-standard soil and application of an assessment factor of 50 to 4-week reproduction test with springtail, the PNECsoil is 0.32 mg/kg
90
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
(dry weight) for both AHTN and HHCB. For comparison we also derived of PNECsoil from PNECwater through equilibrium partitioning between soil pore water and the soil solid (organic) phase. This results in 4.4 and 9.8 mg/kg for AHTN and HHCB, respectively (Table 6). Comparison of both approaches, shows that PNECsoil based equilibrium partitioning is higher than the value based on experimental data from soil organisms by a factor of 14 and 31 for AHTN and HHCB, respectively. This may be partly explained by the higher assessment factor (50) used for the terrestrial data compared to the aquatic data (10). The residual deviation is less than an order of magnitude implying that the assumption of similar sensitivity of aquatic and soil organisms is not violated. This supports the derivation of PNECsediment by equilibrium partitioning, as presented in Table 6. Yet the reliability of this value is considered lower than one derived directly from toxicity data and it is an indication of PNECsediment only. Additionally, the TGD states that when the log Kow is greater than 5 such as it is for AHTN and HHCB, use of PNECsediment from equilibrium partitioning calls for an increase of the exposure parameter
PECsediment (predicted environmental concentration) by a factor of 10 for the calculation of the PEC/PNEC ratio in order to take ingestion via food into account (EU-TGD). The NOAELs derived from the 90-day oral studies with rats were 15 and 150 mg/kg body weight per day. Similar NOAELs have been observed in other studies with rats with these materials including a developmental toxicity study and a pre- and postnatal developmental study (Ford, 1998). The estimation of PNECpredator, based on the 90 day NOAEL for the rat, is calculated based on a 0.4-kg rat eating 20 g dry food/day as follows: the daily intake of the rat is 15 mg AHTN/kg bwt × 0.4 kg bwt=6 mg AHTN. This is taken up in 20g food and therefore the concentration in food is 300 mg AHTN/kg food. Dividing this by an assessment factor of 30, results in a PNECpredator of 10 mg/kg for AHTN. Similarly, the 150 mg/kg NOAEL for HHCB results in a PNECpredator of 100 mg/kg for HHCB.
3.4. Risk characterisation For the risk characterisation, the exposure of the protection targets is compared to the PNEC.
Table 5 Derivation of PNEC for the protection targets (the lowest NOEC is printed bold) Species
AHTN
HHCB
Algae (mg/l) Daphnia Fish
72-h-NOEC 0.374 21-day-NOEC 0.196 21-day-NOEC 0.089; 36-day-NOEC 0.035
72-h-NOEC 0.201 21-day-NOEC 0.111 21-day-NOEC 0.093; 32-day-NOEC 0.068
Assessment factor PNECwater
10 0.0035
Earthworm (mg/kg) Springtail (mg/kg) Lowest NOEC normalised for organic matter in standard soil (Fomexperiment/Fomstandard = 0.1/0.034)
8-week-NOEC 105 4-week-NOEC 45 16
Assessment factor PNECsoil (mg/kg dw)
50 0.32
Rat (mg/kg bw) Rat converted to daily dose (mg/kg food)
13-week-NOEC 15 13-week-NOEC 300
13-week-NOEC 150 13-week-NOEC 3000
Assessment factor for predator PNECpredator (mg/kg food)
30 10
30 100
10 0.0068 8-week-NOEC 45 4-week-NOEC 45 16 50 0.32
Symbols and basic algorithma
Soila
Sedimenta AHTN
HHCB
AHTN
HHCB
Kp: solids–water partition coefficient Kp = Foc×Koc
Foc =0.02
1260
1450
Foc = 0.05
3160
3620
Total compartment-water partitioning coefficient: Kcomp.-water = Fwatersoil×rwater+Fsolidsoil×Kp×rsolid
Fwatersoil =0.2; Fsolid =0.6; rwater =1; rsolid =2.5; rsoil =1.7
1890
2170
Fwatersed. = 0.8; Fsolid = 0.2; rwater = 1; rsolid = 2.5; rsoil = 1.3
1580
1810
PNECep =Kcomp.–water×PNECwater/rcomp. (mg/kg ww) PNECep= PNECep×CONV (mg/kg dw)
CONVsoil = 1.13
3.9
8.7
4.4
9.8
4.3 CONVsediment = 2.6
11
9.5 25
a comp., compartment (soil or sediment); F, fraction, e.g. Fwatersoil: the fraction of water in the soil compartment; Foc, the fraction of organic carbon in the soil or sediment; r, bulk density; CONV, conversion factor for wet to dry weight.
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
Table 6 Derivation of PNECsoil and PNECsediment according to equilibrium partitioning as described in the EU-TGD
91
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
92 Table 7 RCRs for AHTN and HHCB
Exposure concentrations
PNECa
RCR
AHTN Aquatic organisms (mg/l) Sediment organisms (mg/kg dw) Soil organisms (mg/kg dw) Fish-eating predators (mg/kg fw) Worm-eating predators (mg/kg fw)
0.30 0.48 0.029 0.12 0.065
(m) (m)*10b (m+p) (m) (m+p)
3.5 11 0.32 10 10
0.086 0.44 0.091 0.012 0.007
HHCB Aquatic organisms (mg/l) Sediment organisms (mg/kg dw) Soil organisms (mg/kg dw) Fish-eating predators (mg/kg fw) Worm-eating predators (mg/kg fw)
0.50 0.16 0.032 0.12 0.099
(m) (m)*10b (m+p) (m) (p)
6.8 25 0.32 100 100
0.074 0.064 0.10 0.001 0.001
a b
(m), (p), measured or predicted concentration; (m+p), predicted from measured concentration. Measured concentration multiplied by 10 as log Kow of AHTN and HHCB is above 5 (TGD).
Risk is expressed as the risk characterisation ratio (RCR) of exposure to no-effect concentrations for aquatic and sediment-dwelling organisms and their predators, and terrestrial organisms and their predators. Even though the PNECs are based on the lowest NOECs divided by an assessment factor reflecting the robustness of the database and the PECs are based on 90th-percentile measured environmental concentrations or concentrations estimated based on conservative models, it is the goal to obtain RCRs less than one. Exposure concentrations used in determining the RCRs are taken from part I of this risk assessment (Balk and Ford, 1999) while the PNEC values are taken from Tables 5 and 6. The results are presented in Table 7. For AHTN as well as for HHCB, the risk ratios are below 1 for all protection targets. All risk ratios are at or below 0.1 except for the risk ratio of AHTN for sediment organisms. The risk ratios for predators on aquatic and soil organisms are all at or below 0.01 for AHTN as well as for HHCB.
4. Discussion In the EU-TGD quantitative structure-activity relations (QSAR) are given for baseline or mini-
mum toxicity for chronic endpoints for fish (Brachydanio rerio/P. promelas), D. magna and Selenastrum capricornutum. For fish, these QSARs predict the prolonged LC50 correctly and the predicted prolonged NOEC is below the experimental results by factors of 3–10. For D. magna the predicted 16-day-NOEC is too low by a factor of 50, whereas the EC50 for algae is overestimated by a factor of 15–35. In particular the estimates of NOEC-values for AHTN and HHCB are unreliable. The tested aquatic species represent three trophic levels (primary producer, primary and secondary consumer), three taxonomic groups (green algae, crustaceans and bone fish) and three feeding strategies (phototrophic, herbivorous filter feeding and carnivorous). Fish seem to be only slightly more sensitive than the other two species. Comparison of the results of the 21-day-growth test and the 36-day-ELS test for fish shows that the growth and development of fish early life stages seems only marginally more sensitive to these substances than growth in later stages. The physical abnormalities of the fish larvae with AHTN and the underdevelopment of the larvae with HHCB occurred at concentrations where growth was affected as well as survival. In other words, where these effects are observed, the concentrations have reached lethal thresholds. The
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
mechanism for the development of the missing caudal fin with AHTN is unknown, but currently this is subject to further investigations. Preliminary results show that the effect is repeatable in another fish species at the same concentration level. The two soil organisms tested are equally sensitive for AHTN and HHCB. On the basis of equilibrium partitioning their sensitivity seems to be similar to that of aquatic organisms, implying that additional exposure by food ingestion does not play a role. This justifies the use of equilibrium partitioning for the prediction of the toxicity for sediment-dwelling organisms but at the same time shows that the additional safety factor of 10 related to exposure by sediment ingestion, as called for in the TGD for substances with log Kow values greater than 5, is too conservative. The risk assessment presented here is the result of an iterative process where initially the exposure was based on model predictions according to the EU-TGD and the effect assessment was based on aquatic toxicity data only. Triggered by risk ratios above 1, additional data were generated to replace initial predictions based of default values and high assessment and safety factors by empirical data, resulting in more reliable risk evaluations. This also explains why the RCRs presented here deviate from those in an earlier report (Van de Plassche and Balk, 1997). In the present phase, the risk assessments still include various conservative assumptions, in particular in the estimation of the exposure concentrations and toxicity levels for soil and sediment organisms. In the exposure analysis the biodegradability of the substances in soil is taken into consideration with a conservative estimate of the half-life time in soil, but the biodegradability in the sediment was neglected. In addition, the derivation of the PNECsoil includes an extra safety factor of 3 to account for the assumed difference in bioavailability related to the organic content of the soil. Due to the derivation of PNECsediment by equilibrium partitioning an extra safety factor of 10 is included in the sediment risk ratio. The fact that in spite of the conservative assumptions all risk ratios are well below 1, is reassuring.
93
From a risk-assessment point-of-view, priority should be given to refine elements in the risk assessment with the highest uncertainty, for the compartment with the highest RCR. Improved reliability of a risk assessment may be obtained along two lines: by refining the exposure assessment and by increasing the toxicity database. In particular the assessments for the soil and sediment compartments would benefit from a refinement of the input data. Measurement of actual concentrations of AHTN and HHCB in plots with a known history of repeated sludge amendments and measurements of sediment in surface water would eliminate uncertainties introduced by the use of model predictions. Extension of the soil toxicity database with a third test species, e.g. a plant, would result in a more reliable PNEC, which is reflected by a lower assessment factor of 10 instead of 50. If toxicity data for sediment dwelling organisms were available, the present extra safety factor of 10 related to the PNEC derived on the basis of equilibrium partitioning would be abandoned. However, the present low risk ratios for these substances in all compartments do not seem to justify efforts for further refinement.
5. Conclusions The polycyclic musks AHTN and HHCB are less toxic to aquatic organisms than expected on the basis of ‘minimum toxicity’ as predicted by QSAR methods. The NOEC for fish, daphnia and algae range from 35 to 280 mg/l for AHTN and from 70 to 200 mg/l for HHCB. Fish are slightly more sensitive than the other species. Fish embryo development is affected at concentrations just below and at the lethal level. The PNECs derived for aquatic organisms are 3.5 and 7 mg/l for AHTN and HHCB, respectively. The NOEC for the terrestrial worm and springtail range from 45 to 105 mg/kg. These results compare well to the NOEC for soil predicted from the aquatic toxicity on the basis of equilibrium partitioning. However, corrections for soil composition related to the assumed influence of bioavailability and the limited number of tested
94
F. Balk, R.A. Ford / Toxicology Letters 111 (1999) 81–94
species introduce additional safety factors and high assessment factors, resulting in PNECsoil of 0.32 mg/kg for both substances. Sediment toxicity data are not available and PNECsediment is derived by equilibrium partitioning from PNECwater: 11 and 25 mg/kg dw for AHTN and HHCB, respectively. Because of lack of data for wildlife species, the toxicity for predators in the aquatic and terrestrial ecosystem is extrapolated from a 90-day study for the rat. The PNECpredator values are 10 and 100 mg/kg food (fresh weight) for AHTN and HHCB, respectively. The risk expressed as a characterisation ratio (PEC/PNEC) is well below 1 for all compartments. All risk ratios are at or below 0.1 except the risk ratio of AHTN for sediment organisms. The risk ratios for predators on aquatic and soil organisms are all at or below 0.01 for AHTN as well as for HHCB. Keeping in mind the largely conservative approach taken for soil and sediment organisms, these data are reassuring that these substances will not pose an environmental risk.
Acknowledgements Results of an initial risk assessment for AHTN and HHCB have been discussed in a task force consisting of T.J. van Bergen, H.-D. Gaisser and J.D. Middleton of NEA, E.J. van de Plassche and P.T.J. van de Zandt (National Institute of Public Health and the Environment (RIVM) and the authors.
References Api, A.M., Ford, R.A., 1999. Evaluation of the oral subchronic toxicity of AHTN and HHCB in the rat, Toxicol. Lett., (submitted). Balk, F., Ford, R.A., 1999. Environmental risk assessment for the polycyclic musks AHTN and HHCB in the EU, I. fate and exposure assessment. Toxicol. Lett., 111, 57–79. Croudace, C.P., Caunter, J.E., Johnson, P.A., Wallace, S.J., 1997a. AHTN: Chronic toxicity to fathead minnow
.
(Pimephales promelas) embryos and larvae. Report to RIFM, Zeneca Project Report BL5933/B. Croudace, C.P., Caunter, J.E., Johnson, P.A., 1997b. HHCB: Chronic toxicity to fathead minnow (Pimephales promelas) embryos and larvae. Report to RIFM, Zeneca Project Report BL5934/B. EC, 1996. Technical Guidance Documents in support of Directive 96/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC catalogue numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 rue Mercier, L-2965 Luxembourg. Ford, R.A., 1998. The human safety of the polycyclic musks AHTN and HHCB in fragrances — a review. Dt. Lebensm.-Rdsch 94 (8), 245 – 286. Gossmann, A., 1997a. Effects of AHTN on reproduction and growth of earthworms Eisenia fetida in artificial soil. Report to RIFM, IBACON Project no. 1782022. Gossmann, A., 1997b. Effects of HHCB on reproduction and growth of earthworms Eisenia fetida in artificial soil. Report to RIFM, IBACON Project no. 2032022. Hopkins, M.N., Lambert, A.H., 1996. AHTN: 13-week oral (dietary) toxicity in the rat with a 4-week treatment-free period. Report RFA/5/95 to RIFM. Klepka, S., 1997a. Effects of AHTN on reproduction of the springtail Folsomia candida in artificial soil. Report to RIFM, IBACON Project no. 1781016. Klepka, S., 1997b. Effects of HHCB on reproduction of the springtail Folsomia candida in artificial soil. Report to RIFM, IBACON Project no. 2031016. Van de Plassche, E.J., Balk, F., 1997. Environmental risk assessment of the polycyclic musks AHTN and HHCB according to the EU-TGD. RIVM report 601503 008. National Institute of Public Health and the Environment, Bilthoven, NL. Van Dijk, A., 1997a. Acute toxicity of AHTN to Pseudokirchneriella subcapitata. Report to RIFM, RCC Umweltchemie AG Project 380654. Van Dijk, A., 1997b. Acute toxicity of HHCB to Pseudokirchneriella subcapitata. Report to RIFM, RCC Umweltchemie AG Project 380632. Wu¨thrich, V., 1996a. Influence of AHTN on the reproduction of Daphnia magna. Report to RIFM, RCC Umweltchemie AG Project 380665. Wu¨thrich, V., 1996b. AHTN: 21-Day prolonged toxicity study in the bluegill sunfish under flow-through conditions. Report to RIFM, RCC Umweltchemie AG Project 380698. Wu¨thrich, V., 1996c. Influence of HHCB on the reproduction of Daphnia magna. Report to RIFM, RCC Umweltchemie AG Project 380687. Wu¨thrich, V., 1996d. HHCB: 21-Day prolonged toxicity study in the bluegill sunfish under flow-through conditions. Report to RIFM, RCC Umweltchemie AG Project 380711.