Concentrations of Persistent Organic Pollutants in California

Page 1 of 34

Environmental Science & Technology

(Table of Contents Art) 84x47mm (96 x 96 DPI)

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Concentrations of persistent organic pollutants in California children’s whole blood and residential dust Todd P. Whitehead*a, Sabrina Crispo Smithb,c, June-Soo Parkb, Myrto X. Petreasb, Stephen M. Rappaporta, Catherine Metayera a

School of Public Health, University of California, Berkeley, CA USA Environmental Chemistry Laboratory, California Department of Toxic Substances Control, Berkeley, CA, USA c Sequoia Foundation, La Jolla, CA, USA

b

Email addresses: [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] Corresponding author: Todd Whitehead 1995 University Ave., Suite 460, Berkeley, CA 94704 Phone: 1-510-643-2404; Fax: 1-510-643-1735; Email: [email protected] Keywords: Environmental monitoring; house dust; organochlorine pesticides; polybrominated diphenyl ethers; polychlorinated biphenyls

1 ACS Paragon Plus Environment

Page 2 of 34

Page 3 of 34


30

ABSTRACT

31

We evaluated relationships between persistent organic pollutant (POP) levels in the blood of

32

children with leukemia and POP levels in dust from their household vacuum cleaners. Blood and

33

dust were collected from participants of the California Childhood Leukemia Study at various

34

intervals from 1999-2007 and analyzed for two polybrominated diphenyl ethers (PBDEs), two

35

polychlorinated biphenyls (PCBs), and two organochlorine pesticides using gas chromatography-

36

mass spectrometry. Due to small blood sample volumes (100µL),

37

dichlorodiphenyldichloroethylene (DDE) and BDE-153 were the only analytes with detection

38

frequencies above 70%. For each analyte, depending on its detection frequency, a multivariable

39

linear or logistic regression model was used to evaluate the relationship between POP levels in

40

blood and dust, adjusting for child's age, ethnicity, and breastfeeding duration; mother’s country

41

of origin; household annual income; and blood sampling date. In linear regression,

42

concentrations of BDE-153 in blood and dust were positively associated; whereas, DDE

43

concentrations in blood were positively associated with breastfeeding, maternal birth outside the

44

U.S., and Hispanic ethnicity, but not with corresponding dust-DDE concentrations. The

45

probability of PCB-153 detection in a child’s blood was marginally associated with dust-PCB-

46

153 concentrations (p=0.08) in logistic regression and significantly associated with

47

breastfeeding. Our findings suggest that dust ingestion is a source of children’s exposure to

48

certain POPs.



49

INTRODUCTION

50

Persistent organic pollutants (POPs) are stable and widespread in the environment; they

51

accumulate in fatty tissue of biota; and they are toxic to humans and wildlife. Three important

52

classes of POPs are organochlorine (OC) pesticides, polychlorinated biphenyls (PCBs), and

53

polybrominated diphenyl ethers (PBDEs). Organochlorine pesticides, such as DDT [i.e.,

54

dichlorodiphenyltrichloroethane], were used to control insects on agricultural crops and to

55

control insect-borne diseases; PCBs were used in electrical, heat transfer, and hydraulic

56

equipment; and PBDEs were used as chemical flame retardants to treat plastics (e.g.,

57

polyurethane foam) and textiles in consumer products. DDT, PCBs, and PBDEs have been used

58

extensively worldwide, with a global production volume in excess of 1 million metric tons for

59

each.1, 2 In the U.S., DDT and PCBs have been banned since the 1970s, whereas the

60

manufacture and import of the three PBDE commercial mixtures have been phased out more

61

recently (as of January 1, 2005 for Penta-BDE and Octa-BDE; as of December 31, 2013 for

62

Deca-BDE).3

63

Humans are exposed to POPs through various routes, including consumption of

64

contaminated food, inhalation of contaminated air, and accidental ingestion of settled dust.

65

Because consumer goods that have been treated with PBDEs can still be found readily in U.S.

66

homes4, PBDE levels in settled dust remain high (e.g., median concentrations > 1 ppm for BDEs

67

47, 99, and 209 in 20105). Accordingly, it has been suggested that dust ingestion is a major route

68

of exposure to PBDEs for U.S. adults6 and positive relationships have been observed between

69

PBDE levels in matched samples of dust and serum in U.S. adults7, 8. Due to their tendency to

70

make hand-to-mouth contact and their proximity to the floor, young children are expected to

71

receive a relatively large proportion of their total PBDE intake via settled dust compared to


Page 4 of 34

Page 5 of 34


72

adults9 and a positive relationship has been observed between PBDE levels in matched samples

73

of dust and serum in one investigation of toddlers from North Carolina10.

74

In contrast, levels of DDT and PCBs are relatively low in settled dust compared to PBDE

75

levels (e.g., median concentrations < 10 ppb for PCBs 138 and 153 in 201011) and dust ingestion

76

is hypothesized to be a minor contributor to total intake of these POPs for U.S. adults compared

77

to inhalation or diet12. However, one study of 26 adults from Wisconsin reported a positive

78

relationship between PCB concentrations in dust and serum, after adjusting for fish

79

consumption.13 As with PBDEs, young children are expected to receive a relatively large

80

proportion of their total PCB intake via settled dust compared to adults.12 However, no previous

81

study has evaluated the relationship between PCB levels in the blood of young children and in

82

settled dust from their homes.

83

Exposures to DDT14-17, PCBs18, 19, and PBDEs20, 21 have been associated with poorer

84

neurodevelopment in young children. We previously demonstrated that dust concentrations of

85

certain PCBs (PCBs 118, 138, and 153)22 and PBDEs (BDES 196, 203, 206, and 207)23 were

86

positively associated with the risk of acute lymphoblastic leukemia in California children. Our

87

current objective is to evaluate the relationship between POP concentrations in blood from

88

California children and POP concentrations in settled dust from their homes, after accounting for

89

additional covariates, such as breastfeeding. By identifying determinants of POP exposures, it is

90

possible to design strategies that may limit children’s exposure to these harmful chemicals. Our

91

analysis will also assess the utility of dust-POP measurements as surrogates for POP exposure in

92

epidemiological studies of children’s health.

93 94

MATERIALS AND METHODS



95

Study population. The California Childhood Leukemia Study (CCLS) is a case–control

96

study conducted in the San Francisco Bay area and the California Central Valley that seeks to

97

identify genetic and environmental risk factors for childhood leukemia. From 1995 through

98

2008, a total of 997 cases 0–14 years of age were ascertained from clinical pediatric oncology

99

centers and pre-treatment whole blood samples leftover from diagnostic testing were available

100

for chemical analysis in approximately 85% of the case children (no blood was collected from

101

controls). In addition, vacuum-cleaner-dust samples were collected from a subset of

102

participants’ homes from 2001 to 2007. Children eligible for dust sampling were 0-7 years old at

103

diagnosis date and living in the diagnosis residence at the time of dust sampling. We obtained

104

written informed consent from participating parents and study protocols were approved by the

105

institutional review board at the University of California, Berkeley.

106

For the current analysis, we selected a group of 191 cases enrolled from 1999 through

107

2007 with available blood samples that represented a diverse combination of income level,

108

Hispanic ethnicity, and location within the study area. Of those, 64 cases had a vacuum-cleaner-

109

dust sample previously analyzed for POPs.5, 11 Participant selection was not based on anticipated

110

levels of chemical exposure.

111

POPs analysis in whole blood. After diagnostic testing, approximately 0.5 mL of

112

unused whole blood was available from each case for chemical analysis. Blood samples were

113

collected in sodium-heparin green-top vacutainer tubes, regardless of fasting state, and stored at -

114

20°C or colder prior to analysis. The blood sample preparation protocol was adapted from

115

Rogers et al.,24 as described in the Supporting Information (Figure S1) and below. Briefly, after

116

thawing, 100 µL of whole blood was spiked with internal standards (PCB-165 and BDE-139),

117

denatured with acetic acid, and extracted with 6 mL of a 1:9 mixture of methylene


Page 6 of 34

Page 7 of 34


118

chloride:hexane. Extracts were concentrated to 0.5 mL using an automated nitrogen evaporation

119

system (TurboVap LV, Biotage; Uppsala, Sweden). Concentrated extracts were purified by

120

solid-phase extraction using acidified silica gel, solvent exchanged into isooctane, and

121

concentrated to 20 µL. Finally, 13C12-labeled PCB-209 was added as an injection standard.

122

Materials used for chemical analysis were previously described.25

123

Samples were analyzed for OC pesticides, PCBs, and PBDEs using gas chromatography

124

(7890 GC, Agilent Technologies; Sunnyvale, CA)-triple quadrupole mass spectrometry (7000B

125

Series, Agilent Technologies; Sunnyvale, CA). Chromatographic conditions included pulsed

126

splitless injection (20 psi for 1 min) at 250°C, helium carrier gas at 1 mL/min, and a 30-m DB-

127

5ms column with 0.25-mm diameter and 0.25-µm film thickness (Agilent Technologies;

128

Sunnyvale, CA). The GC temperature program was initiated at 90 °C, held for 1 min, ramped at

129

50 °C/min to 150 °C, held for 1min, ramped at 8 °C/min to 225 °C, held for 6.5min, ramped at14

130

°C/min to 310 °C, and finally held for 6 min. The mass spectrometer was operated in electron

131

impact ionization mode using multiple ion detection, source temperature of 275°C, ionization

132

energy of 70 eV, and mass resolution of 1.2 amu. Retention times, precursor masses, product

133

masses, and collision cell energies for analytes, internal standards, and the injection standard

134

were previously described.25 The above analytical protocol can be used to quantify a broad array

135

of OC pesticides, PCBs, and PBDEs25; however, given the limited sample volume (~100 µL)

136

available for analysis, we report concentrations for only the two most prevalent blood

137

contaminants in each class: dichlorodiphenyldichloroethylene (p,p'-DDE, the major metabolite

138

of DDT), trans-nonachlor (a component of the insecticide chlordane), PCBs 138 and 153, and

139

BDEs 47 and 153.



140

Blood samples were analyzed in batches of 12 (100 µL each), including 9 field samples, a

141

bovine serum (HyClone bovine serum, Thermo Fisher Scientific, Inc; Waltham, MA) method

142

blank, a bovine serum blank spiked with each target analyte as a positive laboratory control, and

143

a standard reference material of human serum (National Institute of Standards and Technology

144

SRM 1958; Gaithersburg, MD). Table S1 in the Supporting Information shows that in 22 SRM

145

replicates, we generally observed suitable method accuracy (i.e., average percent errors of 2%,

146

6%, 1%, and −1% for p,p'-DDE, trans-nonachlor, PCB-138, and PCB-153, respectively) and

147

precision (i.e., coefficients of variation of 7%, 20%, 21%, 16%, 23%, and 29%, for p,p'-DDE,

148

trans-nonachlor, PCB-138, PCB-153, BDE-47, and BDE-153, respectively), although the assay

149

tended to overestimate PBDE concentrations in SRM samples (i.e., average percent errors of

150

44% and 67% for BDE-47 and BDE-153, respectively). Not surprisingly, when compared to the

151

1 mL replicates of NIST SRM 1958 that we previously characterized25, the SRM replicates of a

152

smaller sample volume (100 µL), yielded larger coefficients of variation. There was no evidence

153

of a difference in assay precision between sample matrices when comparing whole blood

154

samples (the matrix used for this study) and human serum SRMs (the matrix used for quality

155

control samples), as shown in the Supporting Information (Table S2). Table S3 in the

156

Supporting Information shows results from 23 method blanks. The average BDE-47 mass in the

157

method blanks was 26 pg per sample, which represented a large proportion of the total analyte

158

mass observed in a typical field sample. As such, we subtracted the average concentration of

159

each analyte in 23 method blanks from the concentration in each field sample. Subsequently, the

160

method detection limit (MDL) for each analyte was defined as three times the standard deviation

161

of analyte concentrations in 23 method blanks.


Page 8 of 34

Page 9 of 34


POPs analysis in dust. Dust samples collected from participants’ vacuum cleaners were

162 163

previously analyzed for PCBs11 and PBDEs5 as has been described. Briefly, dust particles

164

smaller than 150 µm were obtained with a 100-mesh sieve and 0.2-g portions were spiked with

165

13

166

column chromatography and gel permeation chromatography, concentrated to 250 µL, solvent

167

exchanged into tetradecane, and spiked with a non-overlapping set of 13C12-labeled injection

168

standards. Samples were analyzed for p,p'-DDE, 15 PCBs (including PCBs 138 and 153), and

169

22 PBDEs (including BDEs 47 and 153) by isotope-dilution/gas chromatography−high

170

resolution mass spectrometry. Concentrations of trans-nonachlor were not quantified in dust

171

samples. Quality control measures for dust samples were analogous to those described above for

172

blood samples, i.e., method blanks, positive laboratory controls, and standard reference materials

173

were employed, as previously described.5, 11 Concentrations of PBDEs5 and PCBs11 in dust

174

samples collected from the CCLS have been previously described. For the subset of dust

175

measurements used in this analysis, detection frequencies for p,p'-DDE, PCBs 138 and 153, and

176

BDEs 47 and 153 were greater than 80% and PBDEs were found at the highest concentrations

177

(see Supporting Information, Table S4 for summary statistics).

178

C12-labeled internal standards, extracted via accelerated solvent extraction, purified by silica

Questionnaire information. We used information from three questionnaires

179

administered on separate occasions (see Supporting Information, Figure S2). Shortly after

180

leukemia diagnosis (2000-2008), participating mothers completed a structured in-home interview

181

designed to ascertain information about a wide variety of topics that are potentially relevant to

182

childhood leukemia etiology. Relevant information collected at the primary interview included

183

the child’s Hispanic ethnicity and breastfeeding duration, as well as the mother’s age and place

184

of birth, and the household annual income. During the dust sampling visit (2001-2007),



185

participating mothers were interviewed a second time to ascertain information relevant to

186

chemical exposures in the home, including the construction date of the residence. Finally, during

187

a telephone interview in 2010, participating mothers answered questions designed to identify

188

possible PBDE sources in the home, including the presence of upholstered furniture with

189

crumbling or exposed foam.

190

Blood collection and the three interviews were each conducted at separate times. The

191

average interval between the date of diagnosis/blood collection and the date of primary interview

192

was 5 months (range: 1 to 28 months). The average interval between the date of diagnosis/blood

193

collection and the date of dust collection was 12 months (range: 5 to 37 months). The average

194

interval between the date of diagnosis/blood collection and the date of the third interview was 6.0

195

years (range: 3.7 to 9.8 years). The average interval between dust collection and the third

196

interview was 5.0 years (range: 2.8 to 8.5 years). The average age of the child at diagnosis/blood

197

collection was 4.6 years (range: 0.2 to 14.4 years). For the subset of participants with dust-POP

198

data (N=64), the average age of the child at diagnosis/blood collection was 4.3 years (range: 1.5

199

to 7.8 years).

200

Statistical analysis. Parallel statistical approaches were employed for two groups of

201

POPs based on detection frequency (±50%). For each POP with a detection frequency of 50% or

202

greater and for all participants (N=191), bivariate relationships between potential explanatory

203

factors and wet-weight POP concentrations (i.e., pg/mL of blood) were evaluated using non-

204

parametric Spearman rank correlation coefficients (for continuous and ordinal factors) and

205

Kruskal-Wallis tests (for categorical and dichotomous factors). Subsequently, for each POP with

206

a detection frequency of 50% or greater and for the subset of participants with dust-POP data

207

(N=64), we evaluated a multivariable linear regression model of natural-log transformed wet-


Page 10 of 34

Page 11 of 34


208

weight blood-POP concentrations. The following factors were considered for inclusion in each

209

POP model: child's age at blood collection (continuous, years), child's sex (male or female),

210

child's ethnicity (Hispanic or not), child's breastfeeding duration (continuous, weeks), household

211

annual income (ordinal,

Concentrations of Persistent Organic Pollutants in California

Recommend Documents