Journal of Exposure Science and Environmental Epidemiology (2012) 22, 617 -- 624
& 2012 Nature America, Inc. All rights reserved 1559-0631/12
www.nature.com/jes
ORIGINAL ARTICLE
Adiposity, body composition, and weight change in relation to
organochlorine pollutant plasma concentrations
Anneclaire J. De Roos1,2, Cornelia M. Ulrich3,4, Andreas Sjodin5 and Anne McTiernan1,2,6
We investigated cross-sectional associations of body composition and weight change with polychlorinated biphenyls (PCB) and
organochlorine pesticides/pesticide metabolites measured in blood collected at the baseline of the Physical Activity for Total
Health study of postmenopausal, overweight women living in the Seattle, Washington metropolitan area. Indicators of greater
adiposity were associated with lower plasma concentrations of most PCBs with six or more chlorine atoms. This pattern
was observed for current weight, body mass index, fat mass percent, subcutaneous abdominal fat, intra-abdominal fat, waist
circumference, hip circumference, waist-to-hip ratio, and maximum adult weight. Conversely, PCB 105, PCB 118, and p,p0 -DDE
were generally increased or showed no association with these variables. Weight gain since age 35 was associated with lower
concentrations of almost every organochlorine we studied, and past weight loss episodes of at least 20 pounds (Z9.1 kg) were
associated with higher concentrations. Our results have implications for epidemiologic studies of organochlorines in terms of
covariates that may be important to consider in statistical analyses, particularly as such considerations may differ importantly
by specific analyte. Our finding of increased organochlorine concentrations with past weight loss episodes may have public
health significance; however, this association requires confirmation in longitudinal studies.
Journal of Exposure Science and Environmental Epidemiology (2012) 22, 617--624; doi:10.1038/jes.2012.43; published online 16 May 2012
Keywords: organochlorine; obesity; polychlorinated biphenyl; weight loss
INTRODUCTION
Organochlorine pollutants, including polychlorinated biphenyls
(PCBs) and organochlorine pesticides, are lipid-soluble compounds that persist in the environment and bioaccumulate in
the food chain. In humans, factors affecting the homeostasis
between organochlorines stored in adipose tissue and circulating
blood concentrations are not clear. Higher blood concentrations
have been measured in obese vs lean individuals, even after the
typical ‘‘correction’’ of the organochlorine concentration for blood
lipids;1--4 however, inverse associations between organochlorines
and body mass index (BMI) have also been observed.4,5 It is likely
that the association between body composition and organochlorine concentration differs by the compound, due to differing
chemical properties and excretion rates as well as patterns of
historical use and current exposure,6 but there are few such
published data on specific organochlorines. Circulating organochlorine concentrations are also affected by weight change in that
weight loss results in reduction of the adipose tissue compartment
and thus leads to increased organochlorine concentrations in both
adipose tissue and blood --- at least in the short term7--9--- however,
it is not well described how weight change throughout the adult
life or historical weight loss episodes affect in vivo organochlorine
concentrations at older ages.
We investigated aspects of body composition and weight change
in relation to organochlorines in postmenopausal, overweight, and
obese women living in the Seattle, Washington metropolitan area.
We measured a broad panel of organochlorine analytes in plasma
samples to describe cross-sectional associations of body composition characteristics and weight change history, including weight
loss episodes, with concentrations of specific individual PCB congeners and organochlorine pesticides/pesticide metabolites. We
hypothesized that indicators of adiposity, including higher weight,
BMI, and fat mass would be associated with lower organochlorine
concentrations, as would weight gain since younger ages.
METHODS
Study Population
The study population included participants in a previously conducted
exercise intervention trial and ancillary study of immune function. The
Physical Activity for Total Health study,10 conducted at the Fred Hutchinson
Cancer Research Center and the University of Washington (UW), was a
randomized controlled trial comparing the effects of a yearlong moderateintensity aerobic exercise intervention vs stretching control among 173
sedentary and overweight/obese postmenopausal women in the greater
Seattle, WA area. Women were ages 50 to 75 years, non-smokers, with alcohol
consumption of fewer than two drinks per day, sedentary (o60 min per
week of moderate-to-vigorous intensity exercise or maximal oxygen
consumption [VO2 max] o25.0 ml/kg per min), overweight or obese (BMI
Z25.0 or between 24.0 and 24.9 with percentage body fat 433%), weight
1
Epidemiology Program, Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA; 2Department of Epidemiology, School of Public
Health, University of Washington, Seattle, Washington, USA; 3Cancer Prevention Program, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA; 4National Center
for Tumor Diseases and German Cancer Research Center, Heidelberg, Germany; 5Centers for Disease Control and Prevention (CDC), National Center for Environmental Health
(NCEH), Division of Laboratory Sciences (DLS), Atlanta, Georgia, USA; 6School of Medicine, Department of Medicine, University of Washington, Seattle, Washington, USA.
Correspondence to: Dr. Anneclaire J. De Roos, Epidemiology Program, Public Health Sciences Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue N,
M4-B874, Seattle, WA 98109-1024, USA.
Tel.: þ 1 206 667 7315. Fax: þ 1 206 667 4787.
E-mail: [email protected]
Received 24 March 2011; accepted 28 November 2011; published online 16 May 2012
Body composition, weight change, and plasma organochlorines
De Roos et al
618
stable for the past 3 months, postmenopausal and not taking hormone
replacement therapy in the past 6 months, and with no clinical diagnosis
of diabetes and fasting blood glucose levels o140 mg/dl. Women were
ineligible if they were volunteering for the study to lose weight, had a
history of surgery for weight loss, or were currently attempting or planning
to attempt weight loss by taking diet pills or entering a structured weightloss program. The ancillary study of Immune Function and Exercise (IMEX)
was conducted among 115 of these participants.11 Additional eligibility
criteria for IMEX were: no history of invasive cancer, cardiovascular disease,
or asthma; no current serious allergies; no regular (Z2 times/week) use of
aspirin or other non-steroidal anti-inflammatory medications; and not
using corticosteroids or other medications known to affect immune
function. Women were enrolled in 1998 through 2000.
Questionnaire data at the study baseline included demographics, body
weight history, medical history, smoking, medication use, and dietary
intake over the past 3 months using a food frequency questionnaire.12
During a clinic visit, study staff measured body weight, height, and waist
and hip circumferences. Total body fat and body fat percentage were
assessed using a DXA whole-body scanner. Intra-abdominal and subcutaneous fat areas were measured using computer tomography scan.
Fasting blood samples took place at the UW Department of Laboratory
Medicine between 0730 and 0830 hours, and were processed within 1 h.
183, 187, 189, 194, 195, 199, 206, 209, 138/158, 196/203)13 and 9
organochlorine pesticides/metabolites (hexachlorobenzene (HCB), betahexachlorocyclohexane (BHCCH), g-hexachlorocyclohexane (lindane), oxychlordane, trans-nonachlor, p,p0 -DDE, o,p0 -DDT, p,p0 -DDT, and mirex) by
high-resolution gas chromatography/isotope-dilution high-resolution mass
spectrometry.14 The plasma sample amount available for the assay ranged
from 0.225 to 0.843 g with a median of 0.382 g. The analytic results were
reported on both a wet-weight basis and lipid-standardized basis (using
previously measured triglyceride and total cholesterol values from the
same blood draw).15
Measurements were successfully conducted on 109 of 111 samples from
the study baseline that were shipped to the CDC laboratory (those
available from the 115 IMEX participants). Detection rates for most of the
PCBs and several of the pesticides/metabolites were high among this
group of postmenopausal women; we list in Table 1 those analytes
detected in 480% of samples. For these organochlorines, we imputed
values below the limit of detection (LOD) with the LOD divided by 2. We
created a variable for the PCB sum (nmol/g-lipid), by adding the molar
concentrations of PCBs with detection frequencies over 80% (Table 1).
Repeatability of the measurements was excellent, with coefficients of
variation between blinded QC replicate samples ranging from 2.4 to 11.2
for the organochlorine analytes in Table 1.
Organochlorine Pollutant Measurement
Statistical Analysis
PCB and organochlorine pesticides/metabolites were measured at the
Combustion Products and Persistent Pollutants Biomonitoring Laboratory
of the Centers for Disease Control and Prevention (CDC) in Atlanta, GA. We
measured 36 PCB congeners (International Union of Pure and Applied
Chemistry (IUPAC) scheme numbers 18, 28, 44, 49, 52, 66, 74, 87, 99, 101,
105, 110, 118, 128, 146, 149, 151, 153, 156, 157, 167, 170, 172, 177, 178, 180,
Statistical analyses were conducted using SAS version 9.3 (Cary, NC).
Associations of body composition and weight-related characteristics
with plasma organochlorine concentrations at the study baseline were
evaluated using linear regression, with the natural log of the measured
organochlorine concentration as the dependent variable. We modeled
each specific organochlorine analyte and the PCB sum separately. We
Table 1.
Organochlorine concentrations in plasma at the study reference datea.
Organochlorine
analyteb
PCB 74
PCB 99
PCB 105
PCB 118
PCB 138/158
PCB 146
PCB 153
PCB 156
PCB 170
PCB 180
PCB 183
PCB 187
PCB 194
PCB 196/203
PCB 199
PCB 206
PCB 209
p,p0 -DDE
HCB
BHCCH
Oxychlordane
t0 -Nonachlor
c
102
108
93
109
109
102
109
105
109
109
90
107
100
109
108
107
93
109
109
95
90
94
Wet-weight concentration
(pg/g)
Lipid-standardized concentration [ng/g-lipid]
N (of 109 total)
Mediand
Minimum detected e
Maximum detected
Average detection limit f
Mediand
13.7
9.6
3.8
19.4
34.5
5.2
45.1
6.1
12.6
33.7
3.6
10.7
8.7
8.5
8.7
5.6
4.0
488
57.4
13.1
18.6
20.3
4.9
2.2
1.5
4.1
12.4
2.2
15.1
2.6
5.1
13.6
2.0
4.3
2.9
3.5
2.9
2.0
1.5
68.5
9.4
6.2
9.4
10.2
68.6
105
63.1
185
235
41.8
258
41.0
37.0
111
12.3
33.6
35.0
31.6
27.3
16.1
26.4
6540
202
650
58.3
78.7
5.0
1.9
1.9
2.7
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
5.1
6.7
8.7
8.7
8.7
100
64.5
26.1
136
239
36.2
322
44.5
88.5
235
24.5
74.5
59.3
60.7
58.5
39.0
28.0
3227
417
93.4
134
140
a
Samples with successful measurements for both organochlorines and lipids.
Analytes shown if 480% of samples had measured values above the limit of detection.
Number of samples with measured values above the limit of detection.
d
Median of all samples (N ¼ 109 including those with non-detected values).
e
Minimum of detected (non-imputed) values.
f
Average detection limit of all samples analyzed; note that minimum detected can be lower than average detection limit if the sample’s detection limit was
lower than average.
b
c
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
& 2012 Nature America, Inc.
Body composition, weight change, and plasma organochlorines
De Roos et al
619
present results as the average percentage change in the organochlorine
concentration per unit change in the body composition or weight history
characteristic.
Characteristics evaluated included measured current (reference date)
weight (kg), height (cm), BMI (kg/m2), fat mass as percentage of total body
mass (%), subcutaneous abdominal fat area (cm3), intra-abdominal fat area
(cm3), waist circumference (cm), hip circumference (cm), and waist-to-hip
ratio, as well as historical weight (kg) at ages 18, 35, and 50 years, and
maximum adult weight (kg). We also calculated net weight changes from
ages 18 and 35 years to the reference date. Where appropriate, we
categorized continuous variables using percentiles or predefined cutpoints. Participants were asked about weight loss episodes with the
question, ‘‘Within the last 20 years, when you were not pregnant or sick,
did you ever lose 10 pounds or more on purpose?’’, and were then asked
to specify the number of episodes in amounts of 10--19 pounds (4.5 to
o9.1 kg), 20--49 pounds (9.1 to o22.7 kg), and 50 or more pounds (22.7 kg
or more). We combined the eight women who reported a weight loss
episode of Z50 pounds with the next-lowest category to examine weight
loss episodes of 20 or more pounds (Z9.1 kg). The timing of these weight
loss episodes within the last 20 years was not specified. However, an
additional question asked, ‘‘How long have you been within 10 pounds of
your current weight (do not count times when you were pregnant or
sick)?’’, and we conducted a subanalysis of our main results stratified by
the number years a woman reported she was within 10 pounds (4.5 kg) of
her current weight, in categories of r5 years and 45 years.
Covariates selected a priori as potential confounders of the crosssectional associations between body composition/weight-related characteristics and plasma organochlorine concentrations were age (years),
education (rhigh school diploma; vocational school, some college, or
associate’s degree; bachelor’s degree, master’s degree or doctorate), race/
ethnicity (white, non-Hispanic vs other), annual household income
(o$20,000, $20,000 to o$35,000, $35,000 to o$50,000K, $50,000 to
o$75,000, $75,000 to o$100,000, Z$100,000), marital status (presently
married vs not), smoking status (ever vs never), number of live births (0, 1,
2, 3, 4, Z5), and lifetime total months of breastfeeding (0, 1--12, 412). In
secondary models, we estimated the associations of interest with adjustment for additional potential confounders expected to be more proximal
in time to the observed associations of interest, and therefore potentially
intermediate in the causal pathway and questionable in terms of true
confounding: typical physical activity (MET minutes per week), alcohol
consumption (drinks per week), and dietary intakes in the past year of fish
(o1.5, 1.5 to o2.5, and Z2.5 servings per week), and total fat, animal
protein, and vegetables (categorized by quartiles).
RESULTS
Our study population organochlorine concentrations (Table 1)
were similar to the National Health and Nutrition Examination
Survey (NHANES) medians of participants age 60 years and older
for several analytes (in ng/g-lipid) including PCB 74 (13.7 (our
study) vs 13.0 (NHANES)) and PCB 118 (19.4 vs 14.7).16 However,
median levels were somewhat lower in our study for other
analytes including PCB 180 (33.7 vs 49.6). Per the inclusion criteria
of the parent study, the 109 women in our study (Table 2) were
postmenopausal (mean age 60.6 years) and overweight or obese
(mean BMI 30.3). Participants were also frequently highly educated
(41% with a bachelor’s degree or higher), had relatively high
incomes (29% with annual incomes Z$75,000), and were of white
race (89%). Age was the covariate most consistently associated
with plasma organochlorine concentrations; the levels of PCB 118,
PCB 180, and p,p0 -DDE were increased by 3.9% (95% CI: 1.2, 6.7),
3.1% (95% CI: 1.8, 4.5), and 2.3% (95% CI: 0.41, 5.0), respectively,
per 1-year increase of age, when modeled with the full set of a
priori covariates. Married status, higher education, and higher
income were associated with higher concentrations of several
organochlorine analytes. Breastfeeding longer than a year (vs no
breastfeeding) was associated higher concentrations of PCBs 74,
99, 105, 118, and p,p0 -DDE, but breastfeeding was not associated
& 2012 Nature America, Inc.
Table 2.
Characteristics of participants at the study reference date
(N ¼ 109)a.
Mean (SD) or N (%)
Age (years)
60.6 (6.9), mean (SD)
Education
High school diploma or less
Vocational school, some college, or
associate’s degree
Bachelor’s degree
Master’s degree or doctorate
25 (22.9%)
20 (18.4%)
White, non-Hispanic race/ethnicityb
96 (88.9%)
16 (14.7%)
48 (44.0%)
b
Annual household income
o$20,000
$20,000 to o$35,000
$35,000 to o$50,000
$50,000 to o$75,000
$75,000 to o$100,000
Z$100,000
12
23
18
21
16
15
Presently married
60 (55.1%)
Ever smoked
(11.4%)
(21.9%)
(17.2%)
(20.0%)
(15.2%)
(14.3%)
55 (50.5%)
Cigarettes/day among ever-smokersb
14.8 (13.6), mean (SD)
Body mass index (kg/m2)
o25
25 to o30
30 to o35
Z35
30.3 (3.9), mean (SD)
7 (6.4%)
51 (46.8%)
36 (33.0%)
15 (13.8%)
Ever pregnant
96 (88.1%)
Number of live birthsb
0
1
2
3
4
Z5
15
13
27
28
16
10
Ever breastfedb
66 (61.1%)
Months breastfed
0
1--12
412
42 (39.3%)
37 (34.5%)
28 (26.2%)
(13.7%)
(11.9%)
(24.8%)
(25.7%)
(14.7%)
(9.2%)
a
Study population includes participants with successful measurements for
both organochlorines and lipids.
b
Variable contains missing values for our study population.
with higher-chlorinated PCBs. Women with two or three live births
(vs none) had significantly lower concentrations of several organochlorines, but there were no associations with greater number of
births. Neither race/ethnicity nor smoking was associated with
organochlorines with any consistency; these variables were
dropped from further consideration as potential confounders.
Associations of body composition and weight change characteristics with plasma organochlorine concentrations differed by
specific organochlorine analyte. Results for all the analytes in
relation to fat mass percent, weight gain since age 35, and weight
loss episodes Z20 pounds (Z9.1 kg) are shown in Figures 1--3,
and detailed results for PCB 180, PCB 118, and p,p0 -DDE are shown
in Table 3. The patterns of association were very similar when
modeling the wet-weight organochlorine concentrations (pg/gplasma) with serum lipids as covariates (results not shown).
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
Body composition, weight change, and plasma organochlorines
De Roos et al
620
10%
5%
HCB
nach
lo
t'-No
CH
Oxyc
hlord
BHC
r
ane
DDE
p,p'-
194
209
199
206
196/2
03
187
183
156
146
170
153
58
138/1
-10%
180
105
118
74
PCB
-5%
sum
99
0%
-15%
Organochlorine analyte
Figure 1. Association of fat mass (% of total body mass) with plasma organochlorines at the study reference date (% difference in mean
organochlorine concentration [ng/g-lipid] per unit increase in fat mass, adjusted for age, education, marital status, income, number of live
births, and total months breastfeeding).
1.5%
1.0%
0.5%
0.0%
-0.5%
ne
t'-No
nach
BHC
lor
lorda
CH
Oxyc
h
DDE
209
194
p,p'-
199
206
196/2
03
1 87
183
180
170
156
146
153
58
138/1
74
-2.5%
118
105
PCB
99
sum
-1.5%
-2.0%
HCB
-1.0%
-3.0%
Organochlorine analyte
Figure 2. Association of net weight change since age 35 (kg) with plasma organochlorines at the study reference date (% difference in mean
organochlorine concentration [ng/g-lipid] per unit increase in weight, adjusted for age, education, marital status, income, number of live
births, and total months breastfeeding).
120%
100%
80%
60%
40%
lor
nach
ane
CH
Oxyc
hlord
HCB
BHC
t'-No
DDE
p,p'-
209
206
199
196/2
194
187
183
03
180
156
170
153
146
58
118
138/1
74
105
-40%
99
-20%
PCB
0%
sum
20%
-60%
Organochlorine analyte
Figure 3. Association of weight loss episode of Z20 pounds within the last 20 years with plasma organochlorines at the study reference date
(% difference in mean organochlorine concentration [ng/g-lipid] in women reporting at least one episode vs none, adjusted for age,
education, marital status, income, number of live births, and total months breastfeeding).
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
& 2012 Nature America, Inc.
Body composition, weight change, and plasma organochlorines
De Roos et al
621
Table 3.
Associations of body composition and weight history characteristics with plasma organochlorine concentrations at the study reference date
(percent difference in plasma organochlorine concentration [ng/g-lipid] per unit increase in characteristic)a.
Characteristic
N
PCB 180
Association
Current body composition
Height (cm)
Weight (kg)
P-value
103
103
0.29 (1.2, 0.67)
0.82 (1.3, 0.35)
103
(25.2%)
(23.3%)
(24.3%)
(27.2%)
3.3
0
18.1
22.0
31.5
(4.8, 1.7)
(Referent)
(30.9, 2.9)
(34.3, 7.5)
(42.6, 18.1)
o0.0001
Fat mass (% of total body mass)
Subcutaneous fat (cm3), per 100 unitsb
Intra-abdominal fat (cm3), per 100 unitsb
Waist circumference (cm)
Hip circumference (cm)
Waist-to-hip ratio, per 0.1 unit
103
101
102
103
103
103
5.0
6.3
16.2
0.99
1.3
11.0
(7.0, 3.0)
(11.7, 0.60)
(26.6, 4.2)
(1.6, 0.43)
(2.0, 0.54)
(19.5, 1.5)
Weight history
Weight (kg), age 18 years
Weight (kg), age 35 years
Weight (kg), age 50 years
Weight (kg), maximum
103
103
103
103
0.27
0.14
0.24
0.26
(0.63,
(0.40,
(0.50,
(0.46,
103
(16.5%)
(45.6%)
(24.3%)
(13.6%)
0.93
0
13.7
18.9
31.6
(1.5, 0.34)
(Referent)
(28.4, 4.1)
(34.3, 0.01)
(45.8, 13.8)
Body mass index (kg/m2)
r27.5
427.5--29.6
429.6--32.8
432.8
26
24
25
28
p,p0 -DDE
PCB 118
0.55
0.0007
Association
P-value
Association
P value
0.40 (2.4, 1.6)
0.36 (0.70, 1.4)
0.70
0.51
0.62 (1.4, 2.7)
0.19 (0.87, 1.3)
0.55
0.73
(3.4, 3.9)
(Referent)
(38.3, 34.6)
(31.1, 50.9)
(35.0, 46.8)
0.91
(1.6, 5.8)
(Referent)
(35.9, 39.6)
(21.9, 70.5)
(24.1, 71.2)
0.28
0.02
0.004
o0.0001
2.0
0
5.4
15.4
14.0
0.78
0.47
0.53
0.20
0
8.8
2.0
2.3
o0.0001
0.03
0.01
0.0006
0.0008
0.02
2.7
19.4
37.4
1.2
0.54
28.5
(2.3, 7.8)
(5.6, 35.0)
(3.5, 82.5)
(0.04, 2.5)
(1.1, 2.3)
(3.8, 59.2)
0.29
0.005
0.03
0.06
0.53
0.02
0.43
1.4
36.5
0.54
0.14
18.9
(4.4, 5.5)
(10.7, 15.1)
(2.8, 81.2)
(0.72, 1.8)
(1.8, 1.6)
(4.2, 47.6)
0.86
0.83
0.03
0.40
0.87
0.12
0.64
0.92
0.91
0.13
0.32
0.06
0.01
0.03
0.65
0.68
0.42
(0.78, 0.73)
(0.09, 1.2)
(0.15, 1.2)
(0.01, 0.86)
0.94
0.02
0.01
0.06
0.45
0.35
0.58
0.13
(0.30, 1.2)
(0.21, 0.92)
(0.05, 1.12)
(0.32, 0.57)
0.24
0.22
0.03
0.58
0.002
(2.3, 0.28)
(Referent)
(55.4, 1.1)
(48.9, 27.7)
(48.5, 42.0)
0.13
0.06
0.36
0.55
0.56
0
11.0
9.4
27.5
(1.9, 0.74)
(Referent)
(41.1, 34.5)
(43.0, 43.8)
(56.5, 21.0)
0.40
0.12
0.05
0.001
1.0
0
32.9
19.2
14.5
0.58
0.67
0.22
Weight loss episodes in past 20 years (vs no loss Z10 pounds (Z4.5 kg))b,c
Ever lost 10--19 lbs (4.5 to o9.1 kg)
69 (67.7%)
27.7 (39.5, 13.7)
Ever lost Z20 lbs (Z9.1 kg)
43 (42.2%)
27.9 (9.3, 49.6)
0.0003
0.002
14.7 (42.8, 27.2)
25.7 (11.7, 79.0)
0.43
0.20
6.7 (37.3, 38.9)
31.4 (7.5, 86.7)
0.73
0.13
Number of weight loss episodes Z20 pounds (Z9.1 kg) in past 20 yearsb
0
59 (57.8%)
0 (Referent)
1--2
25 (24.5%)
25.0 (4.7, 49.3)
Z3
18 (17.7%)
31.7 (8.8, 59.5)
0.01
0.005
0 (Referent)
23.9 (16.9, 84.9)
28.1 (16.7, 97.0)
0.29
0.26
0 (Referent)
22.8 (17.4, 82.6)
43.6 (6.3, 120)
0.31
0.10
Net weight change since age 35 years (kg)
Net change o10 kg (loss or gain)
Net gain 10--20 kg
Net gain 20--30 kg
Net gain Z30 kg
17
47
25
14
0.08)
0.13)
0.01)
0.06)
a
Adjusted for age, education, marital status, income, number of live births, and total months breastfeeding.
Variable contains missing values for our study population.
Categories are not mutually exclusive and were therefore modeled together.
b
c
Higher body weight and other body composition indicators of
increased adiposity were generally associated with lower baseline
plasma concentrations of PCBs with six or more chlorine atoms
(IUPAC numbers 146 and higher). This pattern was observed for
current weight, BMI, fat mass percent (Figure 1), subcutaneous
abdominal fat, intra-abdominal fat, waist circumference, hip
circumference, waist-to-hip ratio, and maximum adult weight.
For example, each unit increase in BMI (kg/m2) was associated
with 3.3% lower average PCB 180 concentration, and obese
women with BMI 432.8 kg/m2 had, on average, 31.5% lower PCB
180 than women who were only modestly overweight (BMI
r27.5 kg/m2; Table 3). Furthermore, each 1% increase in total
body mass as fat was associated with 5% lower average PCB 180
(Table 3). Of the pesticides/metabolites, BHCCH and transnonachlor showed similar directions of association as PCB 180
with BMI and fat mass percent (Figure 1), but were not associated
with intra-abdominal fat or waist-to-hip ratio.
Penta-PCBs (five chlorine atoms) and PCBs with lower chlorination generally showed either no association with the body composition characteristics, or associations in the opposite direction
than PCBs with higher chlorination. For example, greater subcutaneous abdominal fat, intra-abdominal fat, waist circumference, waist-to-hip ratio, weight at ages 35 and 50 years, and
maximum adult weight were associated with higher PCB 118
& 2012 Nature America, Inc.
concentration. Each 100-cm3 increase in intra-abdominal fat was
associated with 37% higher PCB 118 concentration, and each
0.1-unit increase in waist-to-hip ratio was associated with 29%
higher PCB 118 concentration (Table 3). We observed similar
patterns of association for PCB 105 and p,p0 -DDE. Weight at age
50 years was significantly associated with both PCB 118 and p,p0 DDE (Table 3), whereas these organochlorines were not associated with current weight. Given the differing results among the
different PCB congeners, the PCB sum was not surprisingly
unassociated with many of the body composition characteristics
we studied.
Weight gain since age 35 in this group of postmenopausal
women was associated (significantly or non-significantly) with
lower plasma concentrations of almost every organochlorine we
studied (except HCB, Figure 2). Each kilogram of weight gained
since age 35 in this group of postmenopausal women was
associated with B0.9% lower average PCB 180 concentration
(Table 3), and net gain of 30 or more kilograms was associated
with 32% lower PCB 180 compared with little weight change
(o10 kg change). Consistent decreases in organochlorine concentrations across increasing categories of weight gain since age
35 were observed for PCBs 170, 180, 187, 196, 199, and 206.
Similar associations were observed for net weight change since
age 18 (not shown), except the magnitudes of effect were less
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
Body composition, weight change, and plasma organochlorines
De Roos et al
622
often statistically significant, and PCBs 99, 105, and 118 showed
non-significant, positive associations with weight gain.
Weight loss episodes of at least 20 pounds (Z9.1 kg) in the last
20 years were generally associated with higher organochlorine
concentrations at the study reference date (except for HCB and
BHCCH, Figure 3), and concentrations increased with the
frequency of such episodes. For example, women who had lost
20 or more pounds (Z9.1 kg) at least once in the past 20 years
had 28% higher average PCB 180 concentration than those
reporting no weight loss (o10 pounds only (o4.54 kg)), and this
varied by whether the woman reported one or two such weight
loss episodes (25% higher PCB 180) or three or more episodes
(32% higher PCB 180). Statistically significant organochlorine
increases with at least one Z20-pound (Z9.1 kg) weight loss
episode ranged from 25% (PCB 199) to 46% (PCB 146), and with
three or more episodes ranged from 28% (PCB 170) to 71%
(PCB 146). These associations remained after adjustment for
current BMI.
Adjustment beyond the a priori covariates for physical activity,
alcohol consumption, and dietary intakes of fish, total fat, animal
protein, and vegetables did not notably change our results (not
shown). For example, we observed similar associations for fat mass
percent with PCB 180 (5.2% with full adjustment vs 5.0% with a
priori adjustment, per unit increase in fat mass), for waist-to-hip
ratio with PCB 118 (29.6% vs 28.5%), for net weight change since
age 35 with trans-nonachlor (1.7% vs 1.2%), and for ever lost
Z20 pounds (Z9.1 kg) with the PCB sum (30.3% vs 32.9%). Of the
secondary covariates, fish consumption was most consistently
associated with plasma organochlorine concentrations. Consumption of Z2.5 servings of fish per week (vs o1.5 servings per week)
was significantly associated with higher concentrations of several
organochlorines, such as a 35% increase in the PCB sum and a
55% increase in trans-nonachlor, when modeled with the other
covariates. Despite this, it was not an important confounder of the
associations we studied.
In analyses stratified by the number years a woman reported
she was at her current weight (within 10 pounds, or 4.5 kg),
associations between current weight-related characteristics and
organochlorines were generally stronger among women who had
been at their current weight for 5 years or less (n ¼ 59 women)
compared with those within their current weight for 45 years
(n ¼ 42 women). For example, each increase in fat mass percent
was associated with a 6.6% decrease (95% CI: 9.6%, 3.4%) in
PCB 180 concentration among women at their current weight for
r5 years, and a 3.2% decrease (95% CI: 7.4%, 0.26%) among
women at their current weight for 45 years (Supplemental
Figures 1a and 1b). Net weight gain since age 35 was associated
with significantly lower plasma organochlorine concentrations
among women at their current weight for r5 years, for the PCB
sum and PCBs 138/158, 153, 156, 170, 180, 196/203, 199, and 206.
In contrast, only PCB 206 was significantly decreased in association with weight gain since age 35 among women who were at
their current weight for longer than 5 years, and there were nonsignificant increases for many of the organochlorines in this
analysis (Supplemental Figures 2a and 2b). Results for weight loss
episodes of Z20 pounds (9.1 kg) in the past 20 years did not
depend strongly on the number of years within current weight.
One or more such episodes was associated with significantly
higher concentrations of PCBs 170, 187, 206, and 209 among
women at their current weight for r5 years, and with increases in
PCBs 146, 153, 170, 180, 196/203 among women who reported
being at their current weight for 45 years (Supplemental Figures
3a and 3b).
DISCUSSION
Our study is the first to present detailed data on in vivo plasma
concentrations of multiple, specific organochlorine pollutants in
relation to body composition characteristics and weight change
history. Our findings indicating lower concentrations of circulating
organochlorines with higher current adiposity (as reflected by
higher weight, BMI, fat mass percent, etc.) are in our hypothesized
direction, based on a scenario of a greater adipose tissue compartment with higher adiposity allowing in vivo dilution of these
lipid-soluble compounds. We found the most consistent support
of this hypothesis for PCBs with six or more chlorine atoms. In
addition, weight gain since age 35 was associated (although not
always statistically significant) with lower concentrations of nearly
every organochlorine analyte examined (except HCB and oxychlordane), indicating that the principle of in vivo dilution may
apply fairly universally across the different compounds. However,
several of our observations of higher organochlorine concentrations in relation to higher adiposity (such as for PCB 118 and
p,p0 -DDE) are counter to our hypothesis.
Our analyte-specific findings are in agreement with the existing
literature --- of inverse associations of BMI with PCB 18017--19 and
summed PCBs,5,18 and a positive association between BMI and
PCB 118.5,19 Obesity has also been associated in previous studies
with higher serum concentrations PCB 138/158,5 BHCCH,18 and
oxychlordane,19 in contrast to our null findings for these
compounds. Our results for weight gain since early adulthood
agree with a study of women in Québec, Canada, in which weight
gain since age 18 was inversely associated with plasma PCB 153
concentration measured postmenopause,4 and with two longitudinal studies in which BMI that increased over time was
associated with lowering of PCB 153, summed PCBs, and p,p0 -DDE,
but not HCB.20--23 We found that the associations of net weight
gain in adulthood with lower organochlorine concentrations were
mostly limited to women who reported that they had been at
their current weight for 5 years or less, suggesting that an increase
in the amount of adipose tissue with weight gain may initially
provide a dilution effect for these pollutants, but that other
competing factors affect the organochlorine concentrations as
time passes since the weight gain occurred.
Based on our hypothesis, we had no a priori reason to expect
that measures of adiposity would be inversely associated with
certain organochlorine analytes and not associated, or positively
associated, with others. Differences observed between the
organochlorines are at least partially due to differing chemical
properties affecting their propensity for storage in adipose tissue,
as reflected by the octanol-water partition coefficients (Kow values)
for these chemicals, which for PCBs tend to increase with the
number of chlorine atoms.24 A lower propensity for PCB 118 to
dissolve in lipids (reflected by logKow of 6.7970) than PCB 180
(logKow of 7.2070) would contribute to less dependence of its in
vivo concentration on the amount of body fat, as reflected in the
results shown in Figure 1. Another reason for different results
between organochlorine analytes could be the timing of maximum exposure. Although DDT and PCBs were banned in the
United States during a similar time period (1970s), we can assume
that the use of DDT as a pesticide would have ended almost
immediately, whereas PCBs continued to be present as a
component of dielectric fluids in older industrial equipment such
as transformers, and therefore even today present a source of
exposure with leakage from such equipment. Therefore, DDT
exposure declined more rapidly than PCB exposure,25,26 and may
be less likely to show associations with current body composition
and weight characteristics.
Increased organochlorine concentrations with greater adiposity
could occur if adiposity affects the rate of metabolism and
excretion of organochlorines, as hypothesized by Wolff et al.,6
particularly if this effect differs by analyte. Thus, the positive
associations we observed for PCB 118 and p,p0 -DDE measured at
the study reference date may reflect slowed excretion of these
compounds in obese women following maximum exposure. This
idea is supported by our data in that positive associations were
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
& 2012 Nature America, Inc.
Body composition, weight change, and plasma organochlorines
De Roos et al
limited to weight at younger ages (age 50 for both analytes and
age 35 for PCB 118) and did not appear with current (reference
date) weight. Alternatively, although the interpretation relating to
our main hypothesis is that body composition characteristics
affected the organochlorine concentration, the opposite direction
of effect --- that organochlorines affected body composition --- is
also possible and may underlie some of the positive crosssectional associations we observed. For example, PCBs 118 and
105 and p,p0 -DDE were positively associated with several factors
indicating current central adiposity. Few previous studies have
examined the influence of organochlorines on development of
obesity or changes in body composition, although these
chemicals have been suspected as obesogens27 due to their
endocrine-disrupting properties in addition to an observed
negative association between PCB 153 and the metabolismmodulating hormone, adiponectin.28 A recent study reported that
maternal serum DDE levels were associated with rapid infant
weight gain and elevated 14-month BMI in their offspring.29
Certainly, more research in adults is needed from longitudinal
studies of the effects of these chemicals on weight gain and on
patterns of weight gain.
A striking finding for most of the organochlorine analytes was
that women who reported at least one weight loss episode of 20
pounds or more (Z9.1 kg) during the past 20 years had higher
average organochlorine concentrations at the study reference
date. These results did not differ substantially according to the
number of years a woman was at her current weight, suggesting
that weight loss episodes that occurred more than 5 years in the
past (and perhaps up to 20 years in the past) affected current
organochlorine concentrations. Furthermore, organochlorine increases were greatest when such weight loss episodes were
frequent (three times or more). These findings may illustrate the
converse of the dilution scenario with a reduced adipose tissue
compartment from weight loss resulting in increased organochlorine concentrations; however, the associations remained with
additional adjustment for current (reference date) BMI, suggesting
that any organochlorine increase from past weight loss episodes
did not decline to previous levels with weight regain. This could
occur if weight-loss-associated organochlorine increases may in
fact promote weight regain, as proposed by Tremblay and
Chaput;27 however, this scenario was not investigated in our
study because we did not have data on the timing of weight loss
episodes or subsequent weight regain. An alternative explanation
for observed associations is that frequent weight loss (‘‘weight
cycling’’) does not affect in vivo organochlorine levels, and that
observed associations may be due confounding by a related factor
such as diet. Nevertheless, these associations were robust to
adjustment for recent dietary factors. Although there are welldemonstrated benefits of weight reduction such as improvement
of cardiovascular risk factors (e.g., blood pressure, lipids) and
decreased diabetes incidence among high-risk individuals,30--32
increases in organochlorine concentrations are an apparent
understudied consequence of weight loss. Increased pollutant
exposure with weight loss is a potentially important issue because
of the high frequency of weight loss episodes (usually with regain)
among adults,33 however, the public health significance of the
small increases we observed is unknown.
Although the current body composition variables we studied
were measured in the clinic, the weight history variables were selfreported, and thus subject to misreporting. A comparison in
NHANES of self-reported historical weight to measured values
from an earlier survey found that elapsed time contributed to
greater misreporting; however, the degree of bias due to recall
was modest, estimated as 1.8 kg (4 pounds) after 20 years.34 There
was a tendency of women to underreport weight, and this bias
increased with BMI. It is unclear how such reporting bias would
have affected our results; however, we assume that the results for
historical weight are less accurate than those for current weight,
& 2012 Nature America, Inc.
and results may be particularly biased for women of higher BMI
(either historically or currently). In addition, the questions we used
on weight loss episodes and years at current weight have not
been validated for accuracy of recall. These are issues for more
detailed examination in longitudinal studies with careful tracking
of weight over time.
Our study findings have implications for interpretation of
epidemiologic research on persistent pollutants. Our study
population consisted of postmenopausal women --- a group in
which body burden of organochlorine pollutants may have health
consequences including increased risk of certain cancers such as
non-Hodgkin lymphoma35--38 and non-cancer outcomes such as
diabetes.39--41 Given that diabetes etiology is clearly associated
with obesity, and NHL may be also be influenced by obesity
(particularly for certain NHL histologic types42), understanding
how organochlorine pollutant levels vary with body composition
and historical weight change is critical in studies of pollutant
health effects --- particularly when exposure is characterized by a
single measurement, often at the time of diagnosis. For example,
our data suggest that epidemiologic studies showing positive
associations between p,p0 -DDE and health outcomes such as
diabetes may be spurious findings due to confounding by a
possible positive correlation between p,p0 -DDE and indicators of
adiposity. Conversely, positive associations between PCB 180 and
health outcomes may be subject to negative confounding from
obesity-related characteristics due to PCB 180s generally inverse
association with obesity. It is important to note that our exclusive
focus on postmenopausal women limits the generalizability of our
findings, since associations may differ in men or younger women.
Nevertheless, our data illustrate the importance of further
characterization of how past and current body composition affect
(and are potentially affected by) in vivo pollutants, and the need
for analyte-specific consideration of such relationships when
interpreting epidemiologic data.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
ACKNOWLEDGEMENTS
This research was funded by a grant from the National Institute of Environmental
Health Sciences of the National Institutes of Health (R03ES015787).
REFERENCES
1 Falk C., Hanrahan L., Anderson H.A., Kanarek M.S., Draheim L., and Needham L.,
et al. Body burden levels of dioxin, furans, and PCBs among frequent consumers of Great Lakes sport fish. The Great Lakes Consortium. Environ Res 1999:
80(2 Pt 2): S19--S25.
2 Kang D., Tepper A., and Patterson D.G. Jr. Coplanar PCBs and the relative
contribution of coplanar PCBs, PCDDs, and PCDFs to the total 2,3,7,8-TCDD
toxicity equivalents in human serum. Chemosphere 1997: 35: 503--511.
3 Pelletier C., Despres J.P., and Tremblay A. Plasma organochlorine concentrations
in endurance athletes and obese individuals. Med Sci Sports Exerc 2002: 34:
1971--1975.
4 Sandanger T.M., Sinotte M., Dumas P., Marchand M., Sandau C.D., and Pereg D.,
et al. Plasma concentrations of selected organobromine compounds and
polychlorinated biphenyls in postmenopausal women of Quebec, Canada.
Environ Health Perspect 2007: 115: 1429--1434.
5 Agudo A., Goni F., Etxeandia A., Vives A., Millan E., and Lopez R., et al.
Polychlorinated biphenyls in Spanish adults: determinants of serum concentrations. Environ Res 2009: 109: 620--628.
6 Wolff M.S., Anderson H.A., Britton J.A., and Rothman N. Pharmacokinetic variability
and modern epidemiology---the example of dichlorodiphenyltrichloroethane,
body mass index, and birth cohort. Cancer Epidemiol Biomarkers Prev 2007: 16:
1925--1930.
7 Chevrier J., Dewailly E., Ayotte P., Mauriege P., Despres J.P., and Tremblay A. Body
weight loss increases plasma and adipose tissue concentrations of potentially
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
623
Body composition, weight change, and plasma organochlorines
De Roos et al
624
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
toxic pollutants in obese individuals. Int J Obes Relat Metab Disord 2000: 24:
1272--1278.
Pelletier C., Doucet E., Imbeault P., and Tremblay A. Associations between weight
loss-induced changes in plasma organochlorine concentrations, serum T(3)
concentration, and resting metabolic rate. Toxicol Sci 2002b: 67: 46--51.
Walford R.L., Mock D., MacCallum T., and Laseter J.L. Physiologic changes in
humans subjected to severe, selective calorie restriction for two years in
biosphere 2: health, aging, and toxicological perspectives. Toxicol Sci 1999: 52(2
Suppl): 61--65.
McTiernan A., Ulrich C.M., Yancey D., Slate S., Nakamura H., and Oestreicher N.,
et al. The Physical Activity for Total Health (PATH) Study: rationale and design.
Med Sci Sports Exerc 1999: 31: 1307--1312.
Shade E.D., Ulrich C.M., Wener M.H., Wood B., Yasui Y., and LaCroix K., et al.
Frequent intentional weight loss is associated with lower natural killer cell
cytotoxicity in postmenopausal women: possible long-term immune effects. J Am
Diet Assoc 2004: 104: 903--912.
Patterson R.E., Kristal A.R., Tinker L.F., Carter R.A., Bolton M.P., and Agurs-Collins T.
Measurement characteristics of the Women’s Health Initiative food frequency
questionnaire. Ann Epidemiol 1999: 9: 178--187.
Ballschmiter K., and Zell M. Baseline studies of the global pollution. I. Occurrence
of organohalogens in pristine European and antarctic aquatic environments. Int J
Environ Anal Chem 1980: 8: 15--35.
Turner W., DiPietro E., Cash T.P., McClure P.C., Patterson D.G Jr., and Shirkhan H.
An improved SPE extraction and automated sample cleanup method for
serum PCDDs, PCDFs, and coplanar PCBs. Organohalogen Compounds 1994: 19:
31--35.
Mohanka M., Irwin M., Heckbert S.R., Yasui Y., Sorensen B., and Chubak J., et al.
Serum lipoproteins in overweight/obese postmenopausal women: a one-year
exercise trial. Med Sci Sports Exerc 2006: 38: 231--239.
Patterson D.G. Jr., Wong L.Y., Turner W.E., Caudill S.P., Dipietro E.S., and McClure
P.C., et al. Levels in the US population of those persistent organic pollutants
(2003--2004) included in the Stockholm Convention or in other long range
transboundary air pollution agreements. Environ Sci Technol 2009: 43:
1211--1218.
Cerna M., Maly M., Grabic R., Batariova A., Smid J., and Benes B. Serum
concentrations of indicator PCB congeners in the Czech adult population.
Chemosphere 2008: 72: 1124--1131.
Dirinck E., Jorens P.G., Covaci A., Geens T., Roosens L., Neels H., Mertens I., and Van
Gaal L. Obesity and persistent organic pollutants: possible obesogenic effect of
organochlorine pesticides and polychlorinated biphenyls. Obesity (Silver Spring)
2011: 19: 709--714.
Hue O., Marcotte J., Berrigan F., Simoneau M., Dore J., and Marceau P., et al.
Plasma concentration of organochlorine compounds is associated with age and
not obesity. Chemosphere 2007: 67: 1463--1467.
Hagmar L. Polychlorinated biphenyls and thyroid status in humans: a review.
Thyroid 2003: 13: 1021--1028.
Hagmar L., Wallin E., Vessby B., Jonsson B.A., Bergman A., and Rylander L. Intraindividual variations and time trends 1991--2001 in human serum levels of PCB,
DDE and hexachlorobenzene. Chemosphere 2006: 64: 1507--1513.
Vo T.T., Gladen B.C., Cooper G.S., Baird D.D., Daniels J.L., and Gammon M.D., et al.
Dichlorodiphenyldichloroethane and polychlorinated biphenyls: intraindividual
changes, correlations, and predictors in healthy women from the southeastern
United States. Cancer Epidemiol Biomarkers Prev 2008: 17: 2729--2736.
Wolff M.S., Britton J.A., Teitelbaum S.L., Eng S., Deych E., and Ireland K., et al.
Improving organochlorine biomarker models for cancer research. Cancer
Epidemiol Biomarkers Prev 2005: 14: 2224--2236.
24 Jantschi L., and Bolboaca S. Molecular descriptors family on structure activity
relationships 6. Octanol-water partition coefficient of polychlorinated biphenyls.
Leonardo Electronic J Pr Technol 2006: 8: 71--86.
25 Bignert A., Olsson M., Persson W., Jensen S., Zakrisson S., and Litzen K., et al.
Temporal trends of organochlorines in Northern Europe, 1967--1995. Relation to
global fractionation, leakage from sediments and international measures. Environ
Pollut 1998: 99: 177--198.
26 Loganathan B.G., Tanabe S., Hidaka Y., Kawano M., Hidaka H., and Tatsukawa R.
Temporal trends of persistent organochlorine residues in human adipose tissue
from Japan, 1928--1985. Environ Pollut 1993: 81: 31--39.
27 Tremblay A., and Chaput J.P. About unsuspected potential determinants of
obesity. Appl Physiol Nutr Metab 2008: 33: 791--796.
28 Mullerova D., Kopecky J., Matejkova D., Muller L., Rosmus J., and Racek J., et al.
Negative association between plasma levels of adiponectin and polychlorinated
biphenyl 153 in obese women under non-energy-restrictive regime. Int J Obes
(Lond) 2008: 32: 1875--1878.
29 Mendez M.A., Garcia-Esteban R., Guxens M., Vrijheid M., Kogevinas M., and Goni F.,
et al. Prenatal organochlorine compound exposure, rapid weight gain and
overweight in infancy. Environ Health Perspect 2011: 119: 272--278.
30 Anderson J.W., and Konz E.C. Obesity and disease management: effects of weight
loss on comorbid conditions. Obes Res 2001: 9(Suppl 4): 326S--334S.
31 Fujimoto W.Y., Jablonski K.A., Bray G.A., Kriska A., Barrett-Connor E., and Haffner S.,
et al. Body size and shape changes and the risk of diabetes in the diabetes
prevention program. Diabetes 2007: 56: 1680--1685.
32 Torgerson J.S., and Sjostrom L. The Swedish Obese Subjects (SOS) study--rationale and results. Int J Obes Relat Metab Disord 2001: 25(Suppl 1): S2--S4.
33 Serdula M.K., Mokdad A.H., Williamson D.F., Galuska D.A., Mendlein J.M., and
Heath G.W. Prevalence of attempting weight loss and strategies for controlling
weight. JAMA 1999: 282: 1353--1358.
34 Kovalchik S. Validity of adult lifetime self-reported body weight. Public Health Nutr
2009: 12: 1072--1077.
35 De Roos A.J., Hartge P., Lubin J.H., Colt J.S., Davis S., and Cerhan J.R., et al.
Persistent organochlorine chemicals in plasma and risk of non-Hodgkin’s
lymphoma. Cancer Res 2005: 65: 11214--11226.
36 Engel L.S., Laden F., Andersen A., Strickland P.T., Blair A., and Needham L.L., et al.
Polychlorinated biphenyl levels in peripheral blood and non-Hodgkin’s lymphoma: a report from three cohorts. Cancer Res 2007: 67: 5545--5552.
37 Rothman N., Cantor K.P., Blair A., Bush D., Brock J.W., and Helzlsouer K., et al. A
nested case-control study of non-Hodgkin lymphoma and serum organochlorine
residues. Lancet 1997: 350: 240--244.
38 Spinelli J.J., Ng C.H., Weber J.P., Connors J.M., Gascoyne R.D., and Lai A.S., et al.
Organochlorines and risk of non-Hodgkin lymphoma. Int J Cancer 2007: 121:
2767--2775.
39 Everett C.J., Frithsen I.L., Diaz V.A., Koopman R.J., Simpson W.M. Jr., and Mainous
A.G. III. Association of a polychlorinated dibenzo-p-dioxin, a polychlorinated
biphenyl, and DDT with diabetes in the 1999--2002 National Health and Nutrition
Examination Survey. Environ Res 2007: 103: 413--418.
40 Fierens S., Mairesse H., Heilier J.F., De Burbure C., Focant J.F., and Eppe G., et al.
Dioxin/polychlorinated biphenyl body burden, diabetes and endometriosis:
findings in a population-based study in Belgium. Biomarkers 2003: 8: 529--534.
41 Vasiliu O., Cameron L., Gardiner J., Deguire P., and Karmaus W. Polybrominated
biphenyls, polychlorinated biphenyls, body weight, and incidence of adult-onset
diabetes mellitus. Epidemiology 2006: 17: 352--359.
42 De Roos A.J., Ulrich C.M., Ray R.M., Mossavar-Rahmani Y., Rosenberg C.A., and
Caan B.J., et al. Intentional weight loss and risk of lymphohematopoietic cancers.
Cancer Causes Control 2010: 21: 223--236.
Supplementary Information accompanies the paper on the Journal of Exposure Science and Environmental Epidemiology website (http://
www.nature.com/jes)
Journal of Exposure Science and Environmental Epidemiology (2012), 617 -- 624
& 2012 Nature America, Inc.