ARCHIVED - Draft Screening Assessment of Hexabromocyclododecane (HBCD)

Exposure Assessment

A comparison of North American and European levels of HBCD in human breast milk, blood serum (maternal and cord blood), food, adipose tissue and dust is presented in Tables 8 - 14. According to these data, in Canada and North America, HBCD levels in human breast milk, maternal blood/cord blood, and food, as well as dietary intakes of HBCD, either fall within the rangesof, or are lower than, those found in Europe. This would be expected given the global distribution of HBCD usage in manufacturing consumer and industrial end-use products. Consequently, it is expected that Canadian exposures are less than European exposures to HBCD. Scenarios reported by the European Union include those listed in Table 15 (EU RAR 2008).

Upper-bounding estimates of the general population of Canada are presented in Appendix D.

In a study conducted by Roosens et al. (2009), serum concentrations of HBCDs were correlated with dust exposures but not with dietary exposure. Authors reported that the enrichment of the (-)α-HBCD enantiomer in humans appears to be due to in vivo enantioselective metabolism / excretion rather than dust ingestion or diet (Roosens et al. 2009).

The highest reported Canadian human breast milk concentration was 28 µg/kg lipid weight obtained from Canadian women from the Hamilton area in 2005 based on n=35 with 23 measured samples containing HBCD. This study reported low HBCD levels (ppb) found in North American human milk lipid. HBCD values were 20 to 100 times less than BDE47 (a congener of tetrabromodiphenyl ether used as a marker of exposure to this class of brominated flame retardants) in the same samples. HBCD global data suggest that human exposure is relatively uniform. This was the first report of isomeric content of HBCD α- and not ß- or γ-HBCD in human samples and also of potential chiral selectivity of HBCD in humans (Ryan et al. 2006a). As percent lipid content of human breast milk is < 6% wt/wt and often around 3%; 3% lipid content will be used to derive an estimate of intake.

Concentrations of HBCD in representative food commodities for North America were obtained from a U.S. food market basket survey (Schecter et al. 2009). In part I of this larger market basket study, total HBCD across 310 composite samples of 31 food types were measured. Total HBCD varied in and across food groups. The HBCD intake was estimated at 16 ng/day primarily from meat consumption. Limits of detection values were used for instances of non-detects. Upper-bounding intakes from food were as follows: meat, 0.86 µg/kg wet weight (ww); dairy, 0.261 µg/kg ww; eggs 0.01 µg/kg ww; fish products 1.46 µg/kg ww; fats 0.810 µg/kg ww; cereals 0.180 µg/kg ww; fruits 0.022 µg/kg ww; and vegetables 0.018 µg/kg ww.

Consumption of fish from a contaminated lake has been found to correlate with HBCD serum levels (Thomsen et al. 2008). High HBCD serum levels in Norwegians also correlated with dietary exposure to HBCD from seafood consumption. For this reason, consumption of fish with HBCD concentrations of 4.6 µg/kg wet weight (lake trout consumption in Lake Ontario, Canada) was incorporated into the derivation of upper-bounding estimates of exposure for the general population of Canada (Alaee et al. 2004). Additional data on concentrations in food are presented in Table 7.

The Canadian and Russian arctic outdoor ambient air concentration of HBCD that was selected was 1.8 pg/m³ or 1.8 x 10^-6 µg/m³ for Alert, Tagish and Dunai (PWGSC-INAC-NCP 2003). As no Canadian indoor air concentrations of HBCD were identified, measured HBCD in indoor air from residences of the United Kingdom was used as a surrogate (median HBCD concentration of 180 pg/m³ or 0.0002 µg/m³) (Abdallah et al. 2008a).

The highest dust concentration data for indoor air in Canadian homes reported by Abdallah et al. (2008b) of 1300 µg/kg dry weight were used to derive the upper-bounding estimates of exposure for the general population of Canada .

As concentrations of HBCD in Canadian drinking water were not found, the concentration of HBCD in lakes in the U.K. of 270 pg/L or 2.7 x 10^-4 µg/L was used (Harrad et al 2009).

In Canada , the highest exposures were for breast-fed infants 0-6 months of age, with estimated exposures of 1.1 x 10^-1 µg HBCD/kg-bw per day (Health Canada 2008). Concentrations in human breast milk are presented in Table 8. As HBCD is bioaccumulative, it was considerd appropriate to also derive estimates of daily intake based on measured levels of HBCD in human blood of the general population of Canada . Based on first-order kinetics the estimate of daily intake derived from the highest mean Canadian maternal/cord blood level of 2.4 µg/kg lipid weight and a half-life of 64 days with 100% absorption from the oral route was 0.01 µg/kg-bw per day. This value is very similar to the deterministic exposure estimate derived for food sources, confirming the use of the daily dietary intakes as an appropriate measure of exposure.

Consumer Products

HBCD is a brominated flame retardant that may be released from the matrix of consumer products as it is not covalently bound; it is only mixed or dissolved in the material. For this reason, HBCD may migrate from the product over time due to abrasion and usage. As HBCD has a low vapour pressure, it will not volatilize or off-gas from the product.

Upper-bounding estimates of potential oral exposure to HBCD from the mouthing of cushion or upholstered furniture were derived based on the model scenario (Environ 2003a, 2003b). These estimates are presented in Appendix E. Estimates of exposure for infants aged 0-6 months was 5.6 x 10^-5 µg/kg-bw per day, while the oral exposure estimate for toddlers aged 6 months to 4 years of age was 2.7 x 10^-5 µg/kg-bw per day. These estimates for infants and toddlers were derived using an established exposure algorirthm, one formally applied for the chlorinated organo-phosphate flame retardant Tris-2-chloroethyl phosphate (TCEP) (Canada 2009; Environ 2003a, 2003b). The algorithm uses a release rate of 84 mg HBCD/m² of fabric surface area to model wear of unaged or UV-aged fabrics, and was considered an appropriate surrogate to estimate exposure to children from mouthing cushions or upholstery. In comparison, the European Union used a release rate of 2000 mg/m² ( a rate used for general textile release of HBCD) to model a similar exposure scenario. The approach taken in the current screening assessment is also consistent with the approach used by the U.S. Environmental Protection Agency’s Voluntary Children’s Chemical Evaluation Program (Environ 2003a, 2003b). 

A preliminary health risk assessment for HBCD emitted into indoor air by drawing the curtain was carried out with an exposure calculation tool (US EPA MCCEM; Miyake et al. 2009). The lifetime average daily dose was calculated to be 2.67×10^-4µg/kg-bw per day. Parameters used by Miyake et al (2009) included an average indoor air peak concentration of 8.6 ng/m³ for HBCD, with input parameters for room size, room volume, and air exchange rate set to be 5.25 m × 3.80 m × 2.70 m, 53.9 m³, and 0.45 h^-1, respectively. Miyake et al. (2009) derived a margin of exposure of 2.1×10⁵, which they reported indicated low concern for this exposure scenario.

Exposures by the dermal route are considered to be negligible based on the findings of Roper et al. (2007) that the stratum corneum is an efficient barrier to radiolabelled ¹⁴C-HBCD penetration via the dermal route. Exposure estimates for all routes of exposure to consumer products except for mouthing (i.e., dermal and inhalation) were considered negligible and thus were not carried forward in the risk characterization of the European risk assessment (EU RAR 2008). 

Health Effects Assessment

One carcinogenicity bioassay was identified (Kurokawa et al. 1984). B6C3F1 mice, 50 per sex per group, were exposed via the diet for 18 months, resulting in intakes of approximately 0, 13, 130 or 1300 mg/kg-bw per day. There were no overt signs of toxicity. The exposed animals had hepatic changes (hepatocyte swelling, degeneration, necrosis, vacuole formation, fatty infiltration), but this was not well correlated to dose. Incidences of total liver tumours were within the normal range for this strain of mouse.  

The European Union reported that consistently negative results had been observed for HBCD in a range of mutagenicity assays with Salmonella typhimurium (Simmon et al. 1976; Baskin and Phillips 1977; GSRI 1979; Zeiger et al. 1987; Ogaswara and Hanafusa 1993; Hossack et al.1978; US EPA 1990a), in an in vitrocytogenetic test for chromosomal aberrations with human peripheral blood lymphocytes (Guid and Schadly 1996) and in an in vivo assay for clastogenicity in the mouse micronucleus test (Engelhardt and Hoffman 2000). In a non-standard assay with two Chinese hamster cell lines containing duplication mutations in the hypoxanthine-guanine phosphoribosyltransferase (HPRT) gene, a small but significant increase of somatic recombinations was observed (Helleday et al. 1999). The European Union concluded that HBCD lacks significant genotoxic potential both in vitro and in vivo, and suggested that “there is no reason to explore this endpoint further” (EU RAR 2008). Accordingly, HPCD is not considered to have genotoxic potential.  

Zeller and Kirsch (1969) exposed male and female Sprague-Dawley rats for 28 days to dietary concentrations equivalent to 0, 940, 2400 or 4700 mg/kg-bw per day. This study was considered insufficient to assign effect levels, but the data did identify the liver and thyroid as target organs for HBCD toxicity (EU RAR 2008).

Chengelis (1997) exposed male and female Sprague-Dawley rats for 28 days by gavage, at doses of 0, 125, 350 or 1000 mg/kg-bw per day. No significant histopathological lesions were observed. The protocol did not include measurement of thyroid gland weight or serum concentrations of thyroid-stimulating hormone (TSH), T3 or T4. Relative liver weight was significantly increased at the two highest doses in males. The lowest-observed-adverse-effect level (LOAEL) was 125 mg/kg-bw per day, based upon significantly increased relative liver weight in all groups of exposed females. The European Union noted a potential issue of contamination of controls in a 90-day study carried out at the same laboratory (Chengelis 2001 as cited in EU RAR 2008).

Van der Ven et al. (2006) exposed five Wistar rats of each sex by gavage for 28 days to 0, 0.3, 1, 3, 10, 30, 100 or 200 mg/kg-bw per day. The protocol focused upon immune and endocrine effects, including the thyroid hormone axis, hematology, bone size and mineralization and retinoid parameters.  Such endpoints are not typically examined in OECD guideline repeated-dose studies, which could explain why those effects were undetected in other studies. The “most remarkable” findings were dose-related decreased total thyroxin, increased pituitary weight, increased immunostaining of TSH in the pituitary, increased thyroid weight and thyroid follicle cell activation. These effects were restricted to females. In females, liver weight increases were noted at a dose of 29.9 mg/kg-bw per day (BMDL, 22.9 mg/kg-bw per day), while pituitary weight increases were noted at a dose of 50.6 mg/kg-bw per day (BMDL, 29.9 mg/kg-bw per day). The thyroid weight increase occurred at 3.4 mg/kg-bw per day (BMDL, 1.6 mg/kg-bw per day). In a follow-up report, Germer et al. (2006) studied hepatic cytochrome P450 levels and CYP 450 activity. Induction of CYP 3A4 was observed in females while induction of CYP 2B was reported for males, suggesting that sex-specific metabolism could explain the thyroid toxicity noted in females only.

Chengelis (2001) exposed Sprague-Dawley rats (15/sex/group) by gavage (in corn oil) for 90 days, at dose levels of 0, 100, 300 or 1000 mg/kg-bw per day. Five animals per sex per group were maintained for a 28-day recovery period. Increases in liver (all dose groups), thyroid (mid- and high-dose groups, females only) and prostate (dose-dependant increase with statistical significance in the high-dose group) weights were noted. Minimal hepatocellular vacuolization was observed in all exposed animals. The LOAEL was 100 mg/kg-bw per day, based upon increased relative liver weight in both sexes. The European Union reported that control animals may also have been inadvertently exposed (EU RAR 2008).

Zeller and Kirsch (1970) exposed rats via the diet for 90 days, at concentrations that were equivalent to doses of 0, 120, 240, 470 or 950 mg/kg-bw per day. The European Union had noted that the study identified the liver as a target organ, but that effect levels could not be deduced (EU RAR 2008).

Murai et al. (1985) fed pregnant Wistar rats (20 per group) diets that delivered approximate doses of 0, 7.5, 75 or 750 mg/kg-bw per day from days 0-20 of gestation. Six animals per group were allowed to deliver and the pups were maintained until 7 weeks. The absolute and relative maternal liver weight was increased significantly at the highest dose (750 mg/kg-bw per day). There were no significant changes in number of implants, resorptions, live or dead fetuses, or external, visceral or skeletal anomalies observed in the pups (fetal NOAEL, 750 mg/kg-bw per day).

Stump (1999) dosed 25 Charles River rats by gavage on days 6-19 of gestation, at dose levels of 0, 500 or 1000 mg/kg-bw per day. There were no indications of maternal or fetal toxicity reported in this study.

Ema et al. (2008) conducted a two-generation reproductive assay with Crl:CD(SD) rats. The F0 animals consisted of 24 rats per sex per group. Dietary administration resulted in dose levels of 0, 10, 101 and 1008 mg/kg-bw per day for males and 0, 14, 141 and 1363 mg/kg-bw per day for females. Diet preparations were formulated by mixing HBCD particles into an appropriate amount of powdered diet for each dose group. Administration was initiated 10 weeks prior to mating to capture the full spermatigenic cycle, throughout mating, gestation and lactation. The mid dose was the LOAEL (101 mg/kg-bw per day), based upon a dose-related decrease in fertility index in the F0 generation, a significant decrease in the number of primordial follicles in the ovary and a significant increased incidence of animals with decreased size of thyroid follicles in the two highest dose groups in both sexes in the F0 generation and the highest dose group of females in the F1 generation. Neurotoxicity parameters were measured. The only significant effect was a lower completion rate of mid-air righting reflex in F2 female pups at the highest dose (1363 mg/kg-bw per day). The NOAEL for this study was 10 mg/kg-bw per day. The European Union had noted that this study was carried out according to OECD guideline 416 and was in accordance with the principles for good laboratory practice (EU RAR 2008). 

Subsequent to the European Union’s assessment, van der Ven et al. (2009; see also Lilienthall et al. 2009a) conducted a one-generation dietary study with Wistar rats, with targeted exposures of 0, 0 (corn oil solvent control), 0.1, 0.3, 1, 3, 10, 30 or 100 mg/kg-bw per day. Exposure was throughout premating (10 weeks for males, 2 weeks for females), mating, gestation and lactation. Each F0 group consisted of 10 males and 10 females. All F1 litters were maintained. Offspring were further exposed from weaning until 11 weeks of age. The authors considered “the most sensitive effects” to be the decreased trabecular bone mineral density and decreased concentration of apolar retinoids in the liver of F1 females and an increased immune response in F1 males. They noted that the immunological effects appeared to be induced during development and therefore were probably persistent. Similarly, retinoids regulate the transcription of numerous genes and can affect developmental programming, skeletal morphogenesis, embryonic growth, sex differentiation, vascularisation and reproduction. Modulation of the retinoid concentrations was proposed to be related to the immune response. Retinoid signalling is also implicated in the development of the testis and bone tissue, both of which were affected in F1 animals. The lowest critical effective doses were 0.18 mg/kg-bw per day (BMDL, 0.056 mg/kg-bw per day) for decreased tibia trabecular bone mineral density in F1 females, 1.45 mg/kg-bw per day (BMDL, 0.46 mg/kg-bw per day) for increased immune response (immunoglobulin G, sheep red blood cells) in F1 males and 5.1 mg/kg-bw per day (BMDL, 1.3 mg/kg-bw per day) for decreased sum of apolar retinoids in liver of F1 females. Concurrently, the offspring were assessed for dopamine-dependent behaviour and hearing function, by haloperidol-induced catalepsy and brainstem auditory evoked potentials (BAEPs). Reduced latencies to movement onset were observed mainly in females. The overall pattern of BAEP alterations (increased thresholds and prolonged latencies of early waves) suggested a predominant cochlear effect. Although the authors (Lilienthal et al. 2009a) reported that the lower bounds of benchmark doses were between =1 and 10 mg/kg-bw per day for catalepsy and BAEP thresholds, no supplementary data were available, as were for the previous endpoints described. 

Eriksson et al. (2006) exposed neonatal (day 10) male NMRI mouse pups to HBCD by gavage once, at a dose of 0, 0.9 or 13.5 mg/kg-bw. At the age of three months, the mice were assessed for spontaneous behaviour and learning and memory capability.   

Ten male mice per group were tested for spontaneous behaviour by measuring locomotion (horizontal movement, detected by infrared beams), rearing and total activity (all movements, e.g., grooming). The activities were measured for three 20-minute periods. Quantitative data were not presented. For all variables, the control animals became habituated, i.e., activity in response to the novelty of the test chamber diminished over time. The animals exposed to HBCD were hypoactive during the first part of the 60-minute period, while toward the end of the test period they became hyperactive.   

Associative learning and memory were assessed by a Morris swim maze. Groups of 12-17 male mice were tested for the ability to locate a submerged platform in a pool for four consecutive days, and on the fifth day, were tested to find the platform in a changed location in the pool. Five trials were carried out each day. During the acquisition period (days 1-4), both exposed and control mice improved their ability to locate the platform. On the fourth day, the mean latencies of the mice exposed to 13.5 mg/kg-bw were significantly longer than controls (p < 0.01) and the group exposed to 0.9 mg/kg-bw (p < 0.05). The mice in the lower dose group did not differ significantly from controls. On the fifth day, the mice exposed to 13.5 mg/kg-bw took significantly longer (p < 0.05) to find the new position of the platform. The EU RAR (2008) considered that the study was well performed and that the LOAEL (based upon significantly altered spontaneous behaviour including hyperactive condition and reduced hatituation) was 0.9 mg/kg-bw, the lowest dose tested in the study.

A developmental assay with Sprague-Dawley rats was published subsequent to the European Union assessment. Saegusa et al. (2009) exposed pregnant Sprague-Dawley rats to 0, 100, 1000 or 10 000 ppm HBCD via the diet, from gestational day 10 until day 20 after delivery (the day of weaning). On day 20 post-delivery, dosing was terminated and all dams sacrificed. Histopathological assessment was carried out on 10 male and 10 female offspring from each group. The remaining offspring were maintained on regular diet until 11 weeks of age and then sacrificed for histological assessment. The authors reported that maternal exposure resulted in a weak hypothyroidism effect, with weight and histopathological changes of the thyroid and serum T3 and TSH concentrations in offspring receiving 10 000 ppm until weaning. An increase of thyroid weight and decrease of serum T3 concentration continued until the adult stage in groups receiving at least 1000 ppm. With regard to the effect on brain development, HBCD showed evidence of affecting oligodendroglial development at a dose of 10 000 ppm, probably as a result of developmental hypothyroisism. The authors concluded that, based on the developmental brain effect, 100 ppm was the NOAEL for HCBD from changes in thyroid parameters (8.1 - 21.3 mg/kg-bw per day by maternal exposure level). The LOAEL would therefore be 1000 ppm, or 80.7 - 212.9 mg/kg-bw per day, based upon decreased triiodothyronine and increased relative thyroid weight in male offspring at week 11.