Risk assessment for children exposed to decabromodiphenyl (oxide) ether (deca) in the United States

Authors


Abstract

Decabromodiphenyl (oxide) ether (Deca) is a widely used brominated flame retardant in the United States predominantly in the hard-plastic housings of consumer electronics and in flame-retarded backing on textiles used in furniture. A child-specific exposure assessment of Deca was performed for the US Environmental Protection Agency's Voluntary Children's Chemical Evaluation Program (VCCEP). The VCCEP guidance for a tier 1 exposure assessment requires that a screening-level assessment be conducted using currently available data and conservative assumptions. For Deca, relevant exposure pathways considered were general environmental exposures (e.g., exposures to contaminated soil, dust, air, and food), breast milk exposures, inhalation of Deca-containing particulates in air, and mouthing Deca-containing consumer products. For each of these scenarios, a mid-range and upper estimate of age-appropriate intakes were calculated. The calculated intakes indicate that, despite the uncertainties, children appear to be exposed to Deca at levels at least 1 order of magnitude, with most being several orders of magnitude, below the National Academy of Sciences reference dose for Deca of 4 mg/kg/d. This analysis indicates that, using the available data, current levels of Deca in the United States are unlikely to represent an adverse health risk for children.

EDITOR'S NOTE:

Portions of this paper were presented by the author at the Third International Workshop on Brominated Flame Retardants (BFR2004) held 6–9 June 2004 at the University of Toronto, Toronto, Ontario, Canada.

INTRODUCTION

Decabromodiphenyl ether or oxide (Deca), the most highly brominated of the polybrominated diphenyl ethers (PBDEs), is the 2nd-largest-volume brominated flame retardant produced globally. It is used predominantly in hard-plastic housings of consumer electronics, electrical connectors, wire and cable insulation, and in flame-retarded backing on furniture textiles.

Various PBDE congeners have been detected in the environment, foods, and people. Most notable is the detection of some of these PBDEs in breast milk. Many drugs and chemicals may be detected in breast milk (LaKind et al. 2004). All humans have both natural and synthetic chemicals in their bodies, which is a result of our background exposure and living in contact with our environment. Nevertheless, the detection of such substances in human blood or breast milk can be disconcerting, and a cause of concern in those unacquainted with the principles of toxicology (i.e., the xenobiotic's hazard and dose are both critical elements in the determination of risk). Improvements in analytical technology are allowing lower and lower concentrations of chemicals in humans and the environment to be detected. There is also a concurrent increase in reporting by the media of low-level trace contaminants in various matrices, and in many reports there is an assumption that the mere detection of a chemical is biologically meaningful. Given that both toxicity and dose are important in determining risk, detection of trace levels of xenobiotics in human tissues may be unwanted and undesirable, but not necessarily an indication of a health risk.

Deca has a definite positive impact on public health—through its use as a flame retardant some 300 lives are estimated saved each year in the United States alone (Clarke 1997). It has been detected very infrequently when analyzed and at very low levels in humans when found, and a careful evaluation of its potential health risks is warranted.

Several US (NAS 2000; Babich and Thomas 2001) and international (WHO 1994; ECB 2002) organizations have formally evaluated the human health and environmental risks associated with Deca. These assessments have found that Deca does not present a health risk in its current use. This assessment evaluates the potential health risks of Deca for children in the United States and was conducted for the US Environmental Protection Agency (USEPA) Voluntary Children's Chemical Evaluation Program (VCCEP). Children's potential exposures to Deca from all sources (including electronics, upholstery, breast milk, and the general environment) were characterized using data from published literature, agency reports, and information from manufacturers. An extensive literature search showed that few data existed on the concentrations of Deca in environmental media and food in the United States. Biomonitoring data (e.g., serum levels) for Deca in humans were available, and provided an alternative way to calculate intakes. This method may have lower levels of uncertainty than calculations using limited measured data. As a result, this analysis largely relies on biomonitoring data to assess exposures, and thus risks, for children exposed to Deca in the United States.

MATERIALS AND METHODS

The child-specific risk assessment followed the VCCEP guidance for a tier I assessment (USEPA 2000a), and all applicable USEPA guidance. Conservative assumptions were made for all input parameters, and both a mid-range estimate and an upper estimate (UE) were calculated for each pathway. Since the amount of data available for an exposure assessment was limited, more precise and robust statistical evaluations (such as estimates of mean, median, upper 95th percentile, etc. exposures) were not feasible.

On the basis of the manufacture and uses of consumer products containing Deca, intakes from 6 exposure scenarios were quantified (Table 1). Total daily intakes were calculated for 3 receptor populations by aggregating the following scenario-specific intakes:

  • 1.Total aggregate intake for a nursing infant (age 0–2 y) whose mother is occupationally exposed through the formulation of Deca includes intakes from scenarios 1, 3, 4, 5, and 6.
  • 2.Total aggregate intake for a nursing infant (age 0–2 y) whose mother is occupationally exposed through the disassembly of electronics includes intakes from scenarios 2, 3, 4, 5, and 6.
  • 3.Total aggregate intake for a child (age >2–18 y) includes intakes from scenario 6.

To estimate noncancer risks associated with an estimated exposure, a hazard quotient (HQ) was calculated by dividing the estimated intake by a reference dose (RfD). The RfD for Deca used in this assessment, 4 mg/kg/d, was derived by the National Academy of Sciences (NAS) (2000) using the National Toxicology Program's (NTP's) 2-y rat bioassay results (NTP, 1986). The RfD was based on the chronic no-observed-adverse-effect level of 1,120 mg/kg/d, and a composite uncertainty factor of 300.

For noncancer health effects, quantitative risk estimates are typically provided in the form of HQs. The HQ represents the estimated exposure for a specific chemical divided by the RfD, expressed in mg/kg/d.

equation image

If an HQ value is less than 1, then it can reasonably be assumed that the chemical exposure does not pose a significant risk. As HQ values increase above 1, the potential for an adverse effect increases.

RESULTS

Exposure scenario 1 (via breast milk from mother who is involved in the formulation of Deca)

An important exposure pathway for infants may be via breast milk. Deca is expected to transfer only minimally into breast milk because of its low bioavailability (approximately 2%; El Dareer et al. 1987) and large molecular weight (959). Deca has been reported in a few breast milk samples obtained from women who were not expected to have occupational exposure. The potential concentration of Deca in a mother's milk who might be involved in the manufacturing of Deca, and thus her infant's exposure to Deca from her breast milk, were estimated indirectly. To estimate the milk concentration, a workplace air concentration was identified, an air-to-serum ratio was calculated, and then a partition coefficient value was selected to represent the amount of Deca that might transfer from the serum into the breast milk.

Workplace exposure to Deca may occur during manufacturing or formulation into the resin or liquid polymer dispersion. Deca is manufactured in a closed system by the reaction of bromine with diphenyl oxide. The highest exposure potential is likely associated with the activities of packaging Deca for shipping, or of emptying bags into a hopper for product formulation. Once formulated, Deca is encased in a polymer matrix, significantly reducing the potential for worker exposure. The American Industrial Hygiene Association has established a workplace environmental exposure level (WEEL) of 5 mg/m3 for Deca (AIHA 1996). The US Occupational Safety and Health Administration has not set a permissible exposure level for Deca. However, because it is considered a “nuisance dust,” it is subject to a 5 mg/m3 limit. Early industrial hygiene surveys identified employee 8-h time-weighted average exposures to Deca to be in the 1 to 4 mg/m3 range, with possible excursions as high as 42 mg/m3 during certain tasks (AIHA 1996). During activities that generate dust concentrations greater than 5 mg/m3, respirators are expected to be worn in the work environment (AIHA 1996).

Table Table 1.. Exposure scenarios included in the exposure assessment
ScenarioAges (y)Exposures
  1. a Deca = Decabromodiphenyl (oxide) ether.

10–2Ingesting breast milk from a mother who is involved in the formulation of Decaa (bagging operation)
20–2Ingesting breast milk from a mother who is involved in the disassembling of electronics
30–2Mouthing Deca-containing plastic electronic products
40–2Inhaling Deca particulates released from plastic electronic products
50–2Mouthing Deca-containing fabric
6All agesExposed to Deca via the general environment (e.g., soil and dust, diet, ambient air, and water)

The potential intake of Deca that might be incurred by a nursing infant because of the exposure of a working mother was considered the most extreme scenario for uptake. A woman engaged in bagging Deca during manufacture, or in emptying bags of Deca into hoppers for formulators and compounders, was assessed, because this task is believed to produce the highest possible exposure. Because there is a lack of workplace air data in the United States, a UE of 5 mg/m3 was selected on the basis of the WEEL. A mid-range estimate (ME) of 1 mg/m3 was based on a European Union study that concluded that the majority of workplace air levels were below 1 mg/m3 (ECB 2002). Workplace air concentrations and associated blood levels for Swedish workers in an electronics recycling plant were used to estimate an air:serum ratio, which was then combined with estimates of air concentrations for a US worker to derive a hypothetical serum level. The mean air level (n = 2) was 175 ng/m3 (Sjödin, Carlsson, et al. 2001) and the median serum concentration of Deca (n = 19) was 4.8 ng/g lipid (Sjödin et al. 1999), yielding a ratio of 27.4 (Deca, μg/g lipid) per (Deca, mg/m3). It is acknowledged that this is a very limited data set for calculating this ratio, but it appears reasonable on the basis of the properties of Deca.

To estimate the transfer of Deca from serum to breast milk, data regarding the partitioning of other PBDE congeners from serum to breast milk were used. Available data on the levels of other PBDEs in breast milk (Ryan and Patry 2001) and serum (Sjödin, Patterson, et al. 2001) were used to calculate a ratio of breast milk to serum (i.e., for each congener, the breast milk concentration was divided by the serum concentration). These ratios were <1.5 for the various PBDEs, and decreased with increasing molecular weight/bromination, with a value of 0.54 determined for hepta-BDPE (the highest brominated PBDE for which breast milk concentrations have been published). Thus, a UE for breast milk:serum ratio of 0.5 was selected. Because the higher PBDEs have a lower transfer rate to breast milk, using the ratio derived for hepta-BDPE may be a high-end estimate. A ratio of 0.1 was selected as an ME, because it is likely that Deca will partition into breast milk with even greater difficulty because of its greater degree of bromination and higher molecular weight and size. The fraction of breast milk that is lipid was assumed to be 4% (USEPA 2002). A UE breast milk ingestion rate of 980 mL/d was selected on the basis of the upper-percentile value for a child <1 y old (USEPA 2002). No ingestion rates for 1- to 2-y-old children were presented, but they are known to decrease after the age of 9 months (USEPA 2002). Therefore, using the 12-month average value to represent the entire 2-y period would likely overestimate actual exposures. An ME ingestion rate of 742 mL/d was used, on the basis of the mean value for children 1–6 months of age (USEPA 2002). For the ME, an infant was assumed to breast feed from the exposed mother daily from birth to 3 months of age (Collaborative Group on Hormonal Factors in Breast Cancer 2002), and for the UE, daily from birth to 2 y. An ME of 4.36 kg and a UE of 7.84 kg for body weight were derived from the 50th-percentile weights for children from birth through 3 months and 2 y, respectively (USEPA 2002). Although the gastrointestinal absorption of Deca was estimated to be <2% (El Dareer et al. 1987; Hardy 2002), it was necessary to use 100% in these intake calculations, because the reference dose used in risk calculations is based on an ingested dose rather than an absorbed dose.

Using the information, assumptions, and formula presented in Table 2, the estimated daily intake for an infant exposed via ingesting breast milk from a mother who is involved in the formulation of Deca ranged from 0.019 to 0.34 mg/kg/d for the ME and UE, respectively (Table 2). It is acknowledged that there is substantial uncertainty surrounding the methods used to calculate the serum levels, as well as the calculation of the percentage of Deca in the serum that would partition into the breast milk. However, there are enough published data on this class of chemicals and other persistent compounds to provide some level of confidence on the level of conservatism built into these calculations. These are addressed in the discussion.

Exposure scenario 2 (via breast milk from mother who disassembles electronics)

For this pathway, the mother of a breast-feeding infant was assumed to be an electronics disassembly worker. There are no US data for either workplace air concentrations or serum levels of Deca for a disassembly worker. However, Deca was detected in the serum of Swedish workers engaged in dismantling electronic equipment (Sjödin et al. 1999; Sjödin 2000) and in Swedish computer technicians (Hagmar et al. 2000). Therefore, serum concentrations of Deca for a US worker were assumed to be the same as those measured in the Swedish workers. The maximum level (9.9 ng/g serum lipid) reported by Sjödin et al. (1999) was used for the UE, and the median level (4.8 ng/g serum lipid) was used for the ME. Additionally, the Deca in the serum is assumed to partition into breast milk, as discussed above, which is fed to an infant daily from birth through 3 months (ME) and from birth through 2 y (UE).

Using the information, assumptions, and formula presented in Table 3, the estimated daily intake for a child exposed via ingesting breast milk from a mother who disassembles electronics ranged from 3.3 × 10−6 to 2.5 × 10−5 mg/kg/d for the ME and UE, respectively. There was less uncertainty regarding serum concentration than in the previous scenario because the values were obtained from published literature rather than calculated.

Exposure scenario 3 (child mouthing consumer electronics products)

Deca is used to flame-retard synthetic polymers in electrical and electronic equipment. A typical example of its use in the United States is in the cabinet backs of television sets, where Deca is used at a level of about 12% (WHO 1994). For this exposure pathway, Deca is assumed to leach from the surface of the electronic product in the saliva of the infant, and the infant is exposed by reingesting the Deca-containing saliva. Although it seems unlikely that children would mouth the types of electronic products that contain Deca, it was conservatively assumed that the possibility exists.

The available data indicate that Deca does not appear to leach very readily or to a significant degree from consumer electronic products. Norris et al. (1974) found no evidence of leaching from a pellet of acrylonitrile butadiene-styrene (ABS) terpolymer containing 4.25% Deca placed in 2 L of water for a full day at 120 °F (detection limit = 0.075 mg/L). Similarly, no leaching was observed from the ABS pellets immersed in a 3% acetic acid solution at 120 °F for 1 or 7 d. The only conditions under which any leaching was detected by these researchers were when ABS pellets containing Deca were suspended in cottonseed oil for 7 d at 135 °F, with a resulting Deca concentration of 1 mg/L (Norris et al. 1974).

Table Table 2.. Estimated intake of decabromodiphenyl (oxide) ether (Deca) by an infant ingesting breast milk from a mother who is involved in the formulation of Deca
 Mid-range estimate (birth to 3 months)Upper estimate (birth to 2 y)
Exposure parametersValueSource/commentValueSource/comment
  1. a DPE = diphenyl ether.

  2. b Although the absorption of Deca is estimated to be less than 2%, an absorption of 100% is necessary in the intake calculations, because the toxicity value is based on an ingested dose rather than an absorbed dose.

Ca = Deca concentration in workplace (mg/m3)1ECB 20025WEEL (AIHA 1996)
CFa–s = Air/serum transfer factor (μg/g lipid serum) per (mg/m3 air)27.4Sjödin et al. 1999 and Sjödin, Carlsson, et al. 200127.4Sjödin et al. 1999 and Sjödin, Carlsson, et al. 2001
Rb–m = Breast milk to serum ratio (unitless)0.1Based on the fact that higher brominated DPEa do not partition into milk as effectively as lower brominated DPE0.5Conservative assumption that Deca partitions from serum into breast milk on a lipid weight basis at the ratio that hepta-DPE does
Fl:bm = Fraction of breast milk that is lipid (g/mL)0.044% expressed as g/mL (USEPA 2002, table 2–12)0.044% expressed as g/mL (USEPA 2002, table 2–12)
CF = Conversion factor (mg/μg)0.001 0.001
IR = Ingestion rate, breast milk (mL/day)742Mean for ages 1–6 months (USEPA 2002, table 2–12)98012-month average, upper percentile (USEPA 2002, table 2–12)
ABS = Absorption (percent)100%bNecessary to use with toxicity value100%bNecessary to use with toxicity value
BW = Body weight (kg) 0–3 months (ME) and 0–2 y (UE)4.36Average of 50th percentile weights, birth through 3 months (USEPA 2002, table 2–12)7.84Average of 50th percentile weights, birth through 24 months (USEPA 2002, table 2–12)
Daily intake (mg/kg/d)0.019Calculated0.34Calculated
equation image

Hypothetically, the quantity of Deca that is leached from the plastic per day and is available to the child for uptake can be calculated by multiplying the concentration (mg/L) in the leachate by the volume (L) of leachate and dividing by the total number of days that each experiment was conducted. For the UE, using a concentration of 1 mg/L in cottonseed oil, the volume of 2 L, over a period of 7 d, this calculation yields a value of 0.29 mg/d. For the ME, using the detection limit of 0.075 mg/L, the volume of 2 L, over a period of 1 d, this calculation yields a value of 0.15 mg/d. This calculation assumes that the rate of leaching is constant over the entire period, and that a smaller volume of liquid will leach a lesser amount of Deca. Because Deca is not extracted by either water or acetic acid, and is unlikely to leach from the plastic at all when mouthed by an infant, both the UE and the ME are likely to overestimate actual exposures.

Using the information, assumptions, and formula presented in Table 4, the estimated daily intake for a child exposed via mouthing Deca-containing electronic products ranged from 4.3 × 10−6 to 2.5 × 10−4 mg/kg/d for the ME and UE, respectively. There was a significant amount of uncertainty surrounding the amount of Deca that could actually leach out of treated plastics (if at all). Some recent findings from an unpublished study of levels of Deca in household dust provide another means of calculating exposures via the likely route by which Deca might be liberated from consumer products, and these are addressed in the Discussion section.

Exposure scenario 4 (children inhaling Deca particulates in air)

For this exposure pathway, the intake of Deca by children who inhaled particulates released from plastic electronic products was evaluated. Because of its low vapor pressure and the partial encapsulation, Deca encased in hard plastic has a very low tendency to volatilize from this matrix into indoor air. It has been reported that Deca was measured in the air (in the particulate phase and not as a semivolatile) in an office (Sjödin, Carlsson, et al. 2001). A respirable air concentration of 0.087 ng/m3 was selected as a UE on the basis of the highest value reported, and 0.052 ng/m3 was selected as an ME on the basis of the mean of all samples reported, using one-half the detection limit for nondetected concentrations from this study (Sjödin, Carlsson, et al. 2001). It is acknowledged that, in fact, the results from this study may have been an artifact, but are used here since this is a screening assessment.

Table Table 3.. Estimated intake of decabromodiphenyl (oxide) ether (Deca) by an infant ingesting breast milk from a mother who disassembles electronics
 Mid-range estimate (birth to 3 months)Upper estimate (birth to 2 years)
Exposure parametersValueSource/commentValueSource/comment
  1. a DPE = diphenyl ether.

  2. b Although the absorption of Deca is estimated to be less than 2%, an absorption of 100% is necessary in the intake calculations, because the toxicity value is based on an ingested dose rather than an absorbed dose.

Cb = Deca concentration in mother's blood (ng/g lipid)4.8Median for computer disassembly workers in Sweden (Sjödin et al. 1999)9.9Highest level reported for computer disassembly workers in Sweden (Sjödin et al. 1999)
Rb–m = Breast milk to serum ratio (unitless)0.1Based on the fact that higher brominated DPEa do not partition into breast milk as effectively as lower brominated DPE0.5Conservative assumption that Deca partitions into breast milk and serum on a lipid weight basis at the ratio that hepta-DPE does (BDE-183)
Fl:bm = Fraction of breast milk that is lipid (g/mL)0.044% expressed as g/mL (USEPA 2002, table 2–12)0.044% expressed as g/mL (USEPA 2002, table 2–12)
CF = Conversion factor (mg/ng)0.000001 0.000001
IR = Ingestion rate, breast milk (mL/d)742Mean for ages 1–6 months (USEPA 2002, table 2–12)98012-month average, upper percentile (USEPA 2002, table 2–12)
ABS = Absorption (percent)100%bNecessary to use with toxicity value100%bNecessary to use with toxicity value
BW = Body weight, (kg) 0–3 months (ME) 0–2 y (UE)4.36Average of 50th percentile weights, birth through 3 months (USEPA 2002, table 2–12)7.84Average of 50th percentile weights, birth through 24 months (USEPA 2002, table 2–12)
Daily intake (mg/kg/d)0.0000033Calculated0.000025Calculated
equation image

Using the information, assumptions, and formula presented in Table 5, the estimated daily intake of Deca by a child exposed via the inhalation of particulates from plastic electronic products ranges from 3.1 × 10−8 mg/kg/d to 6.3 × 10−8 mg/kg/d for the ME and UE, respectively (Table 5). Even adopting the conservative assumptions discussed above, the calculated intakes via this pathway were orders of magnitude less than the intakes estimated via other pathways. Therefore, inhalation of Deca in the household was deemed to contribute only minimally, if at all, to the total plausible daily intake. Because the mass from this pathway was so small, and these exposures would be included in the values calculated in the general exposures pathway analysis, this value was not included in the aggregate intake calculations.

Exposure scenario 5 (child mouthing fabric)

The National Academy of Sciences (NAS) recently conducted an exposure and risk assessment for fire retardants used in upholstery textiles (NAS 2000). The analyses by the NAS suggest that exposure to Deca via contact with fabrics was minimal. They also indicated that their estimates of exposure were extremely conservative, and it was their opinion that the intake of Deca by children did not warrant concern from a human health risk perspective. The NAS (2000) concluded that no further evaluations were justified for Deca in flame-retarding upholstery textiles. For this assessment, the intake of 0.026 mg/kg/d, calculated by the NAS for children's exposure to Deca via mouthing of upholstery, was selected and included in the aggregate intake.

Exposure scenario 6 (intake by children from the general environment)

There are only limited data on Deca in the environment and food items in the United States. An exhaustive literature review revealed only a few studies reporting analyses of environmental media or food items for Deca content, and in most instances, Deca was not detectable. Because there are such limited environmental and dietary data, the use of a standard intake calculation could significantly under- or overestimate potential exposures. However, Deca has been detected in the body tissues of US residents (Cramer et al. 1990; Sjödin, Patterson, et al. 2001; Stanley et al. 1991), indicating that at least some individuals have been exposed to and have absorbed Deca.

Table Table 4.. Estimated intake of decabromodiphenyl (oxide) ether (Deca) by an infant mouthing electronics that contain Deca
 Mid-range estimateUpper estimate
Exposure parametersValueSource/commentValueSource/comment
  1. a ABS = acrylonitrile butadiene-styrene.

  2. b Although the absorption of Deca is estimated to be less than 2%, an absorption of 100% is necessary in the intake calculations, because the toxicity value is based on an ingested dose rather than an absorbed dose.

CL = Mass of Deca leached from surface into liquid per day (mg/d)0.15Norris et al. 1974. No Deca was extracted from ABSa terpolymer in water for 1 d at 120 °F. This value is the limit of detection (0.075 mg/L) multiplied by the total volume (2 L) divided by the total number of days (1 d).0.29Norris et al. 1974. Extraction of Deca from ABS terpolymer in cottonseed oil at 135 °F for 7 d. This value is the concentration of Deca leached (1 mg/L) multiplied by the total volume (2 L) divided by the total number of days (7 d).
CF = Conversion factor (d/min)0.000691 day has 1440 minutes0.000691 d has 1440 min
MT = Mouthing time (min/d)32.4Total mouthing time, average of means for ages 3–18 months (USEPA 2002, table 6–1, p. 6–12)97.2Total mouthing time, average of maximums for ages 3–18 months (USEPA 2002, table 6–1, p. 6–12)
FS = Fraction of objects with Deca (percent)1%Professional judgment (see text)10%Professional judgment (see text)
ABS = Absorption (percent)100%bNecessary to use with toxicity value100%bNecessary to use with toxiicity value
BW = Body weight, 0–2 y (kg)7.84Average of 50th percentile weights, birth through 24 months (USEPA 2002, table 11–1)7.84Average of 50th percentile weights, birth through 24 months (USEPA 2002, table 11–1)
Daily intake (mg/kg/d)0.0000043Calculated0.00025Calculated
equation image

To overcome some of the uncertainties and limitations that could be associated with calculating intake from a small data set on Deca in environmental media, intake of Deca via the general environment (i.e., for intakes not associated with infant-specific behaviors such as breast milk ingestion or mouthing) was quantified using measured serum levels. In this approach, the serum concentrations of Deca reported in the literature were assumed to represent the steady-state blood level resulting from the amount of Deca that a child might absorb via all possible general (i.e., noninfant-specific) exposure pathways, including inhalation of indoor and outdoor air; ingestion from food or drinking water; incidental ingestion; and dermal contact with soils, sediments, or dust. Using this approach, one accounts for all routes of exposure from all media, whether or not the pathway has been identified, is clearly understood, or has been adequately measured. For this reason, especially in occupational health, there has been significant emphasis placed on the importance of biological monitoring whenever it can be meaningfully applied.

The concentrations used to represent the steady-state body concentration for this calculation were taken from the 1 study that reported serum levels of Deca for individuals in the United States (Sjödin, Patterson, et al. 2001). This study provided summary results for 12 samples taken from US blood donors in 1988. Concentrations ranged from <1 to 35 pmol/g lipid, with a median concentration of <1 pmol/g lipid. Seven of the 12 samples were nondetects. After converting to the appropriate units, the median (0.96 ng/g lipid) and maximum (33.6 ng/g lipid) serum concentrations were used for the ME and UE calculations, respectively.

Calculating the absorbed dose of Deca that would produce the above-mentioned measured serum Deca in humans requires information on the half-life, fraction of body weight in which Deca partitions, and bioavailability of Deca in humans. The average half-life of Deca was reported to be 6.8 d (confidence interval of 3–12 d) in a study of workers with workplace exposure to Deca (Sjödin 2000). An average biologic half-life in humans of 6.8 d was used as the ME, and the lower-bound value of 3 d was used as the UE. Estimates of the fraction of body weight into which Deca partitions were based on the partitioning behavior of Deca in the various body tissues. In rats dosed orally with Deca, concentrations of Deca were higher in the liver than in adipose tissue, and concentrations in other tissues and muscle were much lower than in adipose tissue and liver (El Dareer et al. 1987). A value of 25% was used as the ME and a value of 50% was used as the UE. The oral bioavailability of Deca in rats was <2% (El Dareer et al. 1987). There are no data on inhalation bioavailability, and only in vitro data on dermal bioavailability (Hughes et al. 2001). Values of 2% and 1% were used as the ME and UE, respectively, for percent absorbed (by all routes). In this instance, intake is back-calculated from an absorbed dose, so a lower absorption factor provides a higher intake estimate (i.e., more conservative assumption), because the absorption factor is in the denominator of the equation. Likewise, a shorter half-life and larger volume of distribution yield higher intake estimates.

Table Table 5.. Estimated intake of decabromodiphenyl (oxide) ether (Deca) by young children inhaling particulates released from electronics
 Mid-range estimateUpper estimate
Exposure parametersValueSource/commentValueSource/comment
  1. a Only means are reported.

  2. b Although the absorption of Deca is estimated to be less than 2%, an absorption of 100% is necessary in the intake calculations, because the toxicity value is based on an ingested dose rather than an absorbed dose.

Ca = Deca concentration in air (respirable); vapor attaches to dust particulates in air (ng/m)0.052Office with computers (mean of all samples, using one-half the detection limit for nondetects) (Sjödin, Carlsson, et al. 2001)0.087Office with computers (highest value reported) (Sjödin, Carlsson, et al. 2001)
CF = Conversion factor (mg/ng)0.000001 0.000001
FI = Fraction of time spent in room with TV or computer (unitless)0.8320 h in 24-h period (Hubal et al. 2000)124 h in 24-h period; professional judgment
IhR = Inhalation rate (m3/d)5.65Average of <1 y (4.5) & 1–2 y (6.8), means(USEPA 2002, table 7–13)5.65Average of <1 y (4.5) & 1–2 y (6.8), means (USEPA 2002, table 7–13)a
ABS = Absorption (percent)100%bNecessary to use with toxicity value100%bNecessary to use with toxicity value
BW = Body weight, 0–2 y (kg)7.84Average of 50th percentile weights, birth through 24 months (USEPA 2002, table 11–1)7.84Average of 50th percentile weights, birth through 24 months (USEPA 2002, table 11–1)
Daily intake (mg/kg/d)0.00000031Calculated0.00000063Calculated
equation image

Using the information, assumptions, and formula presented in Table 6, the estimated daily intakes for persons in the community due to exposures via all media were 1.2 × 10−3 and 3.9 × 10−1 mg/kg/d for the ME and UE, respectively. Because the majority of the serum samples tested had nondetectable levels of Deca, it is likely that calculated intakes may overestimate actual exposures. Newer studies published after this assessment was conducted indicate that half-life in humans may be longer than previously predicted and bioavailability may be higher than previously predicted. Both of these changes would yield lower predicted intakes for children to yield the equivalent serum levels. Since this is a screening-level exposure assessment, the more conservative assumptions were used. Since the data for this exposure pathway were from serum samples collected in the late 1980s, it is possible that levels in US residents may have increased since then. Future biomonitoring studies being conducted by the Centers for Disease Control and Prevention will shed light on this issue.

Aggregate exposures

Aggregate exposures potentially experienced by the 3 populations evaluated in this exposure assessment—the infant of a mother who is involved in the formulation of Deca (infant, manufacturer), the infant of a mother who is involved in disassembling consumer electronics (infant, disassembler), and an older child's exposures associated with Deca in the environment—were determined by summing intakes from all applicable exposure pathways for each receptor (Table 7). The infant of a mother who is involved in the formulation of Deca could be exposed via ingesting breast milk, ingesting Deca while mouthing electronic consumer products, ingesting Deca while mouthing furniture fabric, and from general environmental exposures. The infant of a mother who disassembles consumer electronics would have the same exposures, except that the mother's breast milk would contain a different amount of Deca. Children who were no longer breast-feeding or mouthing consumer products (>2–18 y) would be exposed only via the general environment.

As presented in Table 7, there is a difference of up to an order of magnitude between the ME and UE for the 2 infant scenarios, and a difference of 2 orders of magnitude between the ME and UE for the general environment scenario. This magnitude of difference between the ME and UE is not uncommon in screening-level assessments. The highest estimated exposure (UE for the infant, manufacturer scenario) is 0.76 mg/kg/d. The lowest estimated exposure (ME for the child's exposure) is 0.0012 mg/kg/d. The large differences between the ME and UE estimates are indicative of the degree of uncertainty in these calculations. Despite the uncertainties inherent in these calculations, the UE estimates are likely to be significantly higher than the true exposures experienced by children in the United States, because high-end or maximum values were used for many, if not nearly all, of the input parameters.

Table Table 6.. Estimated intake of decabromodiphenyl (oxide) ether (Deca) by children via the general environment
 Mid-range estimateUpper estimate
Input parametersValueSource/commentValueSource/comment
  1. a Values converted from pmol/g lipid to ng/g lipid using the formula: (pmol/g lipid) × (959 g/mol) × (1 mol/1012 pmol) × (109 ng/1 g) = ng/g lipid

Css = Concentration in body, steady state (ng/g lipid)0.96aMedian (Sjödin, Patterson, et al. 2001)33.6aMaximum (Sjödin, Patterson, et al. 2001)
FL = Fraction of body weight into which Deca partitions (lipid, kg/kg BW)0.25Used by USEPA Dioxin Reassessment (USEPA 2000b)0.5Upper-end estimate
CF1 = Conversion factor 1 (lipid, g/kg lipid)1,000 1,000 
CF2 = Conversion factor 2 (mg/ng)0.000001 0.000001 
Ln(2) = Natural log of 2 (unitless)0.693 0.693 
t1/2 = Half life of chemical (d)6.8Mean (Sjödin 2000)3Lower bound on calculated confidence interval (Sjödin 2000)
ADD = Average daily dose (absorbed) (mg/kg-day)0.00046Calculated0.073Calculated
ABS = Absorption (percent)2%El Dareer et al. (1987)1%In this equation, use of a lower ABS will result in a higher intake estimate.
Daily intake (mg/kg/d)0.0012Calculated0.39Calculated
equation image

Toxicity value

The RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. The RfD for Deca used in this assessment, 4 mg/kg/d, was calculated by the NAS instead of using the RfD (0.01 mg/kg/d) from the USEPA's Integrated Risk Information System (IRIS). NAS derived this revised RfD for Deca using the NTP's 2-y bioassay results, which were not available at the time of the IRIS derivation (1984–1985). The more recent NTP study (NTP 1986) used by the NAS has several advantages over the study on which the IRIS RfD was based. The NTP study used more than 1 species (both rats and mice), a larger number of animals (50 vs 25 rats/sex/dose), a higher dose range, and also used a product of higher purity, which is more representative of the commercial formulation currently being used. The purity of the product used for toxicity testing is a particularly compelling reason to rely on the NTP study over the study used by IRIS. The primary study relied upon by IRIS (Norris et al. 1975) used a mixture containing 77% Deca, 21.8% nonabromodiphenyl oxide, and 0.8% octabromodiphenyl oxide, whereas the NTP study used material containing 94–97% Deca. The Agency for Toxic Substances and Disease Registry has derived a minimal risk level for PBDEs as a class of compounds, not specific to Deca. The RfD derived by the NAS is specific to Deca and thus is more appropriate for this assessment.

The RfD derived by the NAS is also comparable to the acceptable daily intake of 3.2 mg/kg/d calculated by the Consumer Product Safety Commission (Babich and Thomas 2001). A more detailed discussion of the animal toxicity data can be found in both the NAS (2000) assessment and the assessment presented to the VCCEP review panel (BFRIP 2002).

Risk calculation

As shown in Table 7 and Figure 1, the HQs for ME scenarios range from 0.0003 to 0.01, and the HQs for the UE scenarios range from 0.1 to 0.2, with the highest HQ associated with the UE for the infant whose mother is involved in the formulation of Deca and is employed in the bagging operation. As such, HQs indicate that the estimates of exposure shown here are unlikely to present an adverse health effect.

Table Table 7.. Decabromodiphenyl (oxide) ether (Deca) exposure estimates and hazard quotientsa
  Exposure estimate (mg/kg/d)Hazard quotient (RfD = 4 mg/kg/d)
Daily intakesExposure duration (y)Mid-rangeUpperMid-range estimateUpper estimate
  1. a RfD = reference dose; NAS = National Academy of Sciences.

  2. b The RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. It can be derived from benchmark dose, with uncertainty factors generally applied to reflect limitations of the data. The RfD for Deca, 4 mg/kg/d, was calculated by the NAS instead of using the current 1999 Integrated Risk Information System (IRIS) RfD (0.01 mg/kg/d). NAS calculated a revised RfD for Deca using the National Toxicology Program 2-y bioassay results (NTP 1986), which were not available at the time of the IRIS derivation (1984–1985).

  3. c Assumes a shorter duration for nursing (0–3 months), based on Collaborative Group on Hormonal Factors in Breast Cancer (2002).

  4. d This value incorporates the intakes for ingestion of breast milk from a mother who is involved in the formulation of Deca, plus ingestion from consumer electronic products, ingestion from mouthing fabric, and general exposures.

  5. e This value incorporates the intakes for ingestion of breast milk from a mother who is involved in disassembling Deca-containing products, plus ingestion from consumer electronic products, ingestion from mouthing fabric, and general exposures.

  6. f This value incorporates the intake from general exposures. See text for details.

Pathway specific     
  Ingestion, breast milk, manufacturer0–21.9E-02c3.4E-010.0050.09
  Ingestion, breast milk, disassembler0–20.00000330.0000250.00000080.000006
  Ingestion, consumer electronics0–20.00000430.000250.0000010.00006
  Ingestion, mouthing fabric (NAS)0–20.0260.0260.0070.007
  General exposures0–180.00120.0390.00030.1
Aggregate     
  Infant, manufacturerd0–20.046d0.76d0.010.2
  Infant, disassemblere0–20.027e0.41e0.0070.1
  Childf(>2–18)0.0012f0.39f0.00030.1

DISCUSSION

The calculations presented here indicate that the potential exposures for each scenario evaluated are quite small. It must be stressed that the ME, as well as the UE, contain a reasonable amount of uncertainty, but conservative assumptions were built into each calculation so that actual exposures should be less than is predicted in this assessment. Additional data would lower the uncertainties and overestimates in the calculations of intake. Moreover, even when using these highly conservative values, the risk calculations show that all HQs are less than 1.

Recently reported data on Deca in breast milk provide a means of checking the conservatism used in estimating exposures in this assessment. Schecter et al. (2003) reported a mean of 0.92 ng/g lipid and a maximum of 8.24 ng/g lipid of Deca in breast milk from volunteers in the United States. In contrast, the maximum estimated breast milk concentration calculated in this screening exposure assessment was 70,000 ng/g lipid. Thus, it is almost certain that the estimated levels of exposure in this exposure assessment are vast over-predictions of actual exposures experienced by infants in the United States. Since the exposures estimated in this assessment are below the RfD and are almost certainly overestimates of exposures, a large margin of safety is currently indicated for this compound.

There are currently several large-scale efforts underway to assess the levels of Deca and the other PBDEs in the blood of the US population. The results from these efforts will help shed light on the range of exposures in the US population and help reduce the uncertainty about the estimates of exposure calculated in this assessment. To help put the results from future biomonitoring studies into perspective, it is possible to calculate a blood concentration associated with an exposure equivalent to the RfD. Using the information in Table 6, a range of Deca serum lipid concentrations of approximately 350 to 3,100 ng/g would be predicted for an exposure of 4 mg/kg/d (this estimate is for both adults and children since there is no data to indicate whether there is a difference in the kinetics of Deca as a function of age). This range actually represents what might be considered a range from the lower end to mean. The upper end of the serum concentration associated with the RfD would be associated with the upper end of the range on half-life (e.g., 15 d), the upper end of bioavailability (e.g., 5%), and the lower end of volume of distribution (e.g., 20%). These assumptions would equate to a serum lipid concentration of 21,600 ng/g. All of these estimates are orders of magnitude higher than serum lipid Deca levels currently reported in the United States. Likewise, a concentration of Deca in breast milk that would be associated with the intake equivalent to the RfD can be calculated using the information in Tables 2 and 3. Assuming a lipid content of 4%, a breast milk ingestion rate of 742 mL/d, and a body weight of 4.36 kg, a Deca concentration of 588,000 ng/g lipid in breast milk would yield an exposure to Deca in a nursing infant equivalent to the RfD of 4 mg/kg/d assuming no other general environmental exposures. This value can be used to put future breast milk biomonitoring results for Deca in perspective.

Figure Figure 1..

Hazard quotients for children's exposure to decabromodiphenyl (oxide) ether (Deca) in the United States.

Recently, some additional information has been released that may give additional confidence that this assessment does not underestimate actual exposures (Rudel, personal communication). A total of 5 dust samples taken in Cape Cod, Massachusetts, USA were analyzed for decabromodiphenyl oxide; 4 samples were from homes, and the 5th was from an office. Results ranged from 917 to 1,472 ng/g. Using the maximum dust concentration of 1,472 ng/g, and dust ingestion rates of 100 to 400 mg/d (USEPA 2002), one can estimate intakes of 1.9 × 10−5 and 7.5 × 10−5 mg/kg/d. The associated HQs are 4.7 × 10−6 and 1.9 × 10−5. These values are several orders of magnitude lower than intakes estimated in our analysis for general environment exposure. Dust concentrations in other areas such as California, which require residential upholstery to meet fire safety standards, may be higher than those measured in Cape Cod, Massachusetts. However, dust concentrations would have to be greater than 7,000,000 ng/g to equal the highest intake estimated in our so-called “general exposure pathway” (3.9 × 10−1 mg/kg/d). This large difference may indicate that dust is only a minor contributor to total environmental exposures, or it may reinforce the conclusion that the calculated intakes presented herein are overestimates of actual exposures.

The factors that would change our interpretations of risk for Deca would obviously be factors that indicate that Deca is more toxic than is currently believed and that it debrominates to a significant degree under realistic environmental conditions.

Multiple national and international studies have concluded that exposures to Deca in the environment do not exceed the levels that are considered tolerable (WHO 1994; ECB 2000; NAS 2000). This study confirms these previous results and shows no apparent risks to infants and children. The lifesaving benefits provided by Deca are well recognized. Estimates suggest that brominated flame retardants (BFRs) used in Deca's applications are responsible for avoiding 280 deaths in the United States annually, at a minimum, and numerous more injuries (Clark 1997). Since BFRs, and Deca specifically, provide a real and valuable benefit to society, a careful and serious comparison of the risks and benefits of these compounds should always be evaluated when making risk management decisions.

Acknowledgements

This paper is based on work performed for the Brominated Flame Retardant Industry Panel as part of the panel's sponsorship of Deca under the USEPA's VCCEP. This paper was partially supported by the Bromine Science and Environmental Forum.

Ancillary