Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jan 31.
Published in final edited form as: Environ Toxicol Pharmacol. 2018 Dec 23;66:24–35. doi: 10.1016/j.etap.2018.12.018

Organochlorine pesticides and polychlorinated biphenyls (PCBs) in early adulthood and blood lipids over a 23-year follow-up

Jose R Suarez-Lopez a,*, Chase G Clemesha a, Miquel Porta b, Myron D Gross c, Duk-Hee Lee d
PMCID: PMC12858084  NIHMSID: NIHMS2137346  PMID: 30594847

Abstract

Background:

Some evidence in humans suggests that persistent organic pollutants (POPs), including organochlorine pesticides (OCPs) and polychlorinated biphenyls (PCBs), may alter the blood lipid composition. This study analyzed associations between serum POPs concentrations in young adulthood with blood lipid levels up to 23 years later.

Methods:

Serum POPs were measured in year 2 of follow-up (n = 180 men and women, ages: 20-32y), and plasma lipids in follow-up years 2, 7, 10, 15, 20 and 25. 32 POPs were detectable in ≥75% of participants (23 PCBs, 8 OCPs and PBB-153). We created summary scores for PCBs and OCPs for both wet-weight, and lipid standardized (LP) concentrations. We used repeated measures regression adjusting for demographic factors, BMI, smoking, diabetes status, among others.

Results:

We observed positive associations of the 23 LP-PCB score with total cholesterol (βper SD increase [95%CI]: 5.0 mg/dL [0.7, 9.2]), triglycerides (7.8 mg/dL [−0.9, 16.5]), LDL (4.2 mg/dL [0.2, 8.2]), oxidized LDL 3.4 U/L (−0.05, 6.8), and cholesterol/HDL ratio (0.2 [0.02, 0.3]). The associations for triglycerides (14.7 mg/dL [0.4, 20.1]), cholesterol/HDL (0.33 [0.09, 0.56]) and, to some extent, LDL (4.7 md/dL [−1.6, 10.9]) were only observed among participants in the upper 50th percentile of BMI. Non-dioxin-like PCBs had stronger associations that dioxin-like PCBs. OCPs and PBB-s had positive associations with most outcomes.

Conclusions:

PCBs and PBB-153 measured in young adulthood were positively associated with prospective alterations in most blood lipid components, with evidence of effect modification by BMI. Further longitudinal studies with multiple measures of POPs overtime are needed.

Keywords: Organochlorine, PCB, Lipids, Longitudinal, Adults

1. Introduction

Persistent organic pollutants (POPs), including polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), polybrominated biphenyls (PBBs) and organochlorine pesticides (OCPs), are chemicals that have been used for pest control in agriculture, as lubricants, and in the manufacture of building materials, textiles among others (Contreras López, 2003). Although many POPs, including PCBs and OCPs have been banned or discontinued in the US and most countries worldwide between the 1970s and the 2000s, POPs are still detectable at significant levels in the food supply (Schecter et al., 2010) and in the general population of the US (Lee et al., 2007) and many other countries. This is due to the stable and highly-lipophilic chemical structures of POPs, with half-lives that can range from under a year, to over a decade in fat tissue (Milbrath et al., 2009).

Chronic exposure to background levels of POPs in humans has been linked to negative health outcomes, which may stem from endocrine disruption (Elobeid et al., 2010), inflammation (Ibrahim et al., 2011) and mitochondrial dysfunction (Ruzzin et al., 2010), among other processes. Substantial evidence links exposure to POPs with type II diabetes (Ibrahim et al., 2012; Lee et al., 2011a, 2006; Magliano et al., 2014; Remillard and Bunce, 2002; Tang et al., 2014) and cardiovascular disease (Goncharov et al., 2010; Lee et al., 2012; Lind and Lind, 2012), including risk factors such as alterations in blood lipids. Low levels of high-density lipoprotein cholesterol (HDL), and high total cholesterol, high low-density lipoprotein cholesterol (LDL), high oxidized LDLs, and high cholesterol/HDL ratio have been associated with increased atherogenesis and cardiovascular disease (Navab et al., 2011; Parthasarathy et al., 2010; Prospective Studies Collaboration et al., 2007; The Emerging Risk Factors Collaboration* et al., 2009; Trpkovic et al., 2015). Feeding mice and rats with fish or fish oil with background levels of POPs was found to induce greater increases in blood cholesterol and triglycerides than feedings with fish or fish oil without POPs (Ibrahim et al., 2011; Ruzzin et al., 2010). In humans, only a few studies have prospectively assessed associations between blood concentrations of POPs and lipids. A prospective study of older adults in Sweden found that serum PCB levels were positively associated with cholesterol and LDL 5-years later; these associations were weaker among participants who gained the greatest amount of weight in the 5-year period (Penell et al., 2014). In prospective analyses among healthy young adults of the Coronary Artery Risk Development in Young Adults (CARDIA) study in the US, non-monotonic associations were observed between various PCB congeners and HDL (U-shape) and triglycerides (inverse U-shape) measured in serum samples collected 18 years after the PCB measurement (Lee et al., 2011b). The remaining studies are cross-sectional and have described positive linear associations between plasma PCBs or OCPs and total serum lipids, cholesterol and/or triglycerides (Aminov et al., 2013; Baker et al., 1980; Goncharov et al., 2008; Moysich et al., 2002; Singh and Chan, 2018; Tokunaga and Kataoka, 2003). There is also some evidence of inverse U-shaped associations cross-sectionally (Arrebola et al., 2014). There is a need to expand longitudinal studies on this topic, particularly because cross-sectional analyses can be challenging to interpret considering the affinity that PCBs, OCPs, PBDEs and PBBs have with all blood lipids, including triglycerides and all cholesterol components.

The objective of this investigation was to characterize the associations of background exposures to POPs in young adulthood with blood lipid levels over a 23-year follow-up, among a subset of participants with and without diabetes from the CARDIA study. We hypothesized that greater concentrations of organochlorine pesticides and PCBs would be positively associated with all blood lipid levels, and inversely with HDL, considering the previously described positive associations between POPs and cardiovascular disease.

1.1. Participant selection

The present study included 180 participants from a case-control study nested within the CARDIA study (Lee et al., 2010). The CARDIA study examined 5115 Caucasian and Afro-descendant participants aged 18–30 years in 1985–1986. Men and women were recruited in equal numbers from each of four locations: Birmingham, AL; Chicago, IL; Minneapolis, MN; and Oakland, CA. Additional details of the study design and recruitment have been previously described in detail (Friedman et al., 1988). Participants were reexamined at follow-up years 2 (1987–1988), 5 (1990–1991), 7 (1992–1993), 10 (1995–1996), 15 (2000–2001), 20 (2005–2006) and 25 (2010–2011). With respect to the initial (1985–1986) population, the participant retention proportions were 90%, 86%, 81%, 79%, 74%, 72% and 72% for each follow-up examination, respectively. Informed consent was obtained from each study participant at every examination. The CARDIA study was approved by the institutional review boards of the University of Minnesota, University of Alabama at Birmingham, Northwestern University, and the Division of Research at Kaiser Permanente Health Care Plan.

Participants of the present nested case-control study included 90 diabetics (cases) who were randomly selected from CARDIA participants and who did not have diabetes during year 0 and 2 of follow-up. Diabetes status was defined as taking glucose-lowering medications or having fasting glucose levels ≥126 mg/dL at two or more follow-up examinations after the year 2 exam (CARDIA years 7, 10, 15 and 20). Ninety participants without diabetes (controls) were selected from those who had fasting glucose levels below 100 mg/dL at follow-up years 0, 7, 10, 15, and 20 and had not been diagnosed with diabetes before follow-up year 20 and were not receiving glucose-lowering medications. Controls were frequency matched to cases, selected at random within several year-0 body mass index (BMI) categories (< 20, 20–24.9, 25–29.9, 30–39.9, and 40+ kg/m2).

Since the selection of controls required attendance at study years 0, 7, 10, 15, and 20, they had a small amount of missing data in the current analyses: only 6% (n = 5) of controls did not attend the year 25 examination. Missing follow-up examinations occurred in 11% (n = 10) at year 7, 10% (n = 9) at year 10, 17% (n = 15) at year 15, 12% (n = 11) at year 20 and 20% (n = 18) at year 25.

1.2. Measures

Participants were asked to fast for 12 h prior to each examination. Plasma samples were processed by centrifugation within an hour of blood draw and stored at −70 °C. The samples were analyzed at the Northwest Lipid Research Center in Seattle, Washington within six weeks of sample collection. Total cholesterol, and triglyceride concentrations were measured using an enzymatic assay at the Abbot Laboratories in North Chicago, Illinois. HDL concentrations were measured via precipitation using dextran sulfate-magnesium chloride on the ABA 200 Biochromatic device. LDL concentrations were calculated using the Friedewald equation (Friedewald et al., 1972). Oxidized LDL concentrations were measured using a monoclonal antibody mAb-4E6–based competition ELISA (Mercodia, Uppsala, Sweden). The analytical performance of this assay has been described elsewhere (Holvoet et al., 2006). Levels of total cholesterol, LDLs, HDLs and triglycerides were measured in years 2, 7, 10, 15, 20 and 25; levels of oxidized LDL were measured in year 15 and 20 blood samples.

Concentrations of POPs in stored CARDIA year 2 serum samples (collected in 1987 and 1988) were measured in 2008 as part of the Young Adults Longitudinal Trends in Atherosclerosis (YALTA). The YALTA study is an ancillary study to the CARDIA study. Analyses of samples were performed using solid-phase extraction and final determination using gas chromatography isotope dilution high-resolution mass spectrometry (Barr et al., 2003; Sjödin et al., 2004) at the CDC Environmental Health Laboratory. A total of 55 POPs were measured: 9 OCPs, 35 PCB congeners, 10 PBDE congeners, and 1 PBB congener. Of these 55 POPs, 32 which had at least 75% of values above the detection limit were selected for analysis in the present study (8 OCPs, 23 PCBs, and PBB-153). Among these, non-detectable concentrations were replaced with half of the detection limit. The levels of detection were dependent on the blood sample amount available for each participant and specific for each POP measured.

1.3. Statistical methods

Because the general population is exposed to a mixture of many POPs (and other environmental compounds) (Porta et al., 2008), we assessed the associations of composites of various POPs with blood lipid levels over time. Analyses of POPs using composite variables substantially reduces the number of associations tested, thus reducing the potential for spurious findings. Since concentrations of individual POPs varied substantially (e.g., the sum of the concentrations of the 8 most prevalent POPs were 25.6 times greater than the 8 least prevalent), we created a summary variable for the 32 POPs detectable in ≥75% of participants (32 POP summary score) as the sum of the participants’ log-transformed concentrations of each of the 32 POPs divided by the groups’ standard deviation of the corresponding log-transformed POP (Σ[logPOPindividual / logPOP standard deviationgroup]). We created similar summary variables for 8 organochlorine pesticides (8 OCP summary score), 23 PCBs (23 PCB summary score), 7 dioxin-like PCBs (7 DL-PCB summary score) and 16 non-dioxin-like PCBs (16 NDL-PCB summary score). A list of POPs included in these analyses is presented elsewhere (Suarez-Lopez et al., 2019). Overall, the crude Pearson correlation coefficients of the 32 POP summary score with the 23 PCB and 8 OCP summary scores were 0.98 and 0.80, respectively. Because of the high correlation between the 32 POP and 23 PCB summary scores, we limited the use of the 32 POP summary score to the display of participant characteristics across quartiles of POPs concentrations in Table 1. The correlation between 23 PCB and 8 OCP summary scores was 0.70. We also log-transformed PBB-153, and analyzed associations with it separately because it was the only congener of its class.

Table 1.

Participant characteristics by quartiles of POPs wet weights measured in year-2 of follow-up.

Variable 32 POP Summary quartiles
N All 1
n = 45a 		2
n = 45a 		3
n = 45a 		4
n = 45a 		P-trend
Age at year 2 180 27.2 (3.5) 25.6 (3.4) 26.3 (3.5) 28.1 (3.2) 28.8 (2.9) < 0.001
Gender (% Female) 180 62 67 67 60 53 0.08
Race (% Caucasian) 180 41 47 42 33 42 0.54
Smoking Y25 (%) 156
Never 52 46 63 55 44 0.94
Former 29 30 24 33 31 0.93
Current 19 24 13 12 26 0.84
Case-control status (% diabetic) 180 50 42 62 44 51 0.19
BMI Y2 (Kg/m2) 180 30.1 (6.8) 31.8 (7.5) 31.4 (6.8) 27.9 (6.5) 29.1 (5.7) 0.03
BMI Y15 (Kg/m2) 165 33.6 (7.1) 35.5 (7.7) 35.8 (6.8) 31.1 (6.2) 32.0 (6.4) 0.02
BMI Y25 (Kg/m2) 157 34.1 (8.2) 35.2 (9.0) 36.7 (8.4) 32.8 (7.4) 32.0 (7.2) 0.05
Cholesterol Y2 (mg/dL) 176 189.67 (36.8) 169.8 (31.4) 182.9 (33.8) 190.7 (30.0) 216.5 (36.4) < 0.001
Cholesterol Y15 (mg/dL) 165 182.6 (34.3) 168.5 (32.7) 178.5 (35.7) 189.1 (30.8) 195.3 (32.4) < 0.001
Cholesterol Y25 (mg/dL) 156 180.9 (37.3) 169.6 (40.2) 170.4 (31.4) 193.6 (37.2) 188.8 (34.8) 0.004
HDL Y2 (mg/dL) 176 51.1 (13.9) 51.5 (13.3) 49.7 (11.4) 52.3 (10.2) 51.1 (19.4) 0.46
HDL Y15 (mg/dL) 165 46.6 (12.5) 48.5 (13.5) 42.1 (11.3) 48.2 (10.3) 47.3 (13.8) 0.36
HDL Y25 (mg/dL) 156 56.1 (16.2) 58.2 (15.5) 49.8 (12.6) 57.6 (15.3) 58.9 (19.3) 0.54
LDL Y2 (mg/dL) 176 119.9 (33.8) 103.2 (29.4) 115.2 (33.9) 123.5 (30.5) 138.6 (32.3) < 0.001
LDL Y15 (U/L) 164 114.4 (31.5) 99.8 (29.2) 113.4 (33.9) 121.9 (27.7) 123.6 (30.0) < 0.001
LDL Y25 (mg/dL) 152 100.6 (31.4) 92.3 (34.1) 92.3 (29.2) 111.7 (31.4) 105.5 (27.7) 0.007
Oxidized LDL Y15 (mg/dL) 135 82.9 (25.0) 77.5 (26.5) 83.6 (27.1) 81.7 (21.3) 89.5 (24.5) 0.005
Oxidized LDL Y20 (mg/dL) 153 83.6 (23.3) 71.4 (21.6) 84.9 (22.2) 87.0 (23.4) 90.2 (22.4) 0.001
Triglycerides Y2 (mg/dL) 176 94.3 (62.5) 76.6 (33.5) 91.2 (51.9) 75.4 (46.5) 134.8 (87.8) < 0.001
Triglycerides Y15 (mg/dL) 165 107.4 (67.4) 102.8 (74.0) 113.4 (66.3) 93.2 (49.6) 119.8 (74.3) 0.051
Triglycerides Y25 (mg/dL) 156 123.8 (108.1) 95.0 (51.7) 143.1 (108.9) 120.4 (107.0) 135.2 (141.0) 0.081
Cholesterol-HDL Ratio Y2 176 4.00 (1.43) 3.49 (1.05) 3.87 (1.09) 3.80 (1.01) 4.85 (2.00) < 0.001
Cholesterol-HDL Ratio Y15 165 4.21 (1.42) 3.73 (1.29) 4.47 (1.18) 4.09 (1.06) 4.57 (1.87) < 0.001
Cholesterol-HDL Ratio Y25 156 3.48 (1.32) 3.10 (1.08) 3.61 (1.02) 3.61 (1.32) 3.59 (1.70) 0.007
32 POP Summary-WW Y2 (logPOPi / logSDg) 180 217.9 (26.7) 183.4 (14.1) 211.0 (4.9) 226.2 (4.9) 251.1 (13.7) –
23 PCB Summary-WW Y2 (logPOPi / logSDg) 180 147.7 (21.4) 120.7 (12.0) 141.7 (5.3) 154.3 (5.2) 174.0 (11.5) –
23 PCB Summary-LP Y2 (logPOPi / logSDg) 180 107.4 (20.0) 83.5 (10.9) 101.8 (8.7) 114.7 (9.2) 129.6 (13.1) –
8 OCP Summary-WW Y2 (logPOPi / logSDg) 180 57.29 (5.5) 51.3 (4.6) 56.7 (3.4) 58.4 (3.2) 62.8 (3.4) –
8 OCP Summary-LP, Y2 (logPOPi / logSDg) 180 43.7 (4.58) 39.1 (4.1) 43.5 (3.2) 45.1 (3.1) 47.2 (3.5) –
Log(PBB-153) (pg/g), Y2 180 2.94 (0.76) 2.59 (0.65) 3.01 (0.88) 3.28 (0.63) 3.37 (0.76) –
Log(PBB-153) (pg/g), Lipid Standardized Y2 180 1.49 (0.63) 1.47 (0.62) 1.81 (0.88) 2.03 (0.70) 1.95 (0.75)
Open in a new tab

Presented values are percentage or mean (standard deviation).

POP summary scores were calculated as: Σ(log[POPi] / log[SDg]). POPi: Concentration of each POP of each individual.

SDg: Standard deviation of the corresponding POP of all participants.

WW: Wet weight.

LP: Lipid standardized.

a

n at baseline.

We present results of analyses of POPs summary scores derived from both wet-weight concentrations and lipid-standardized concentrations (POP concentration divided by the total blood lipid content [total cholesterol + triglycerides]), to account for the presence of POPs in blood lipids. We tested longitudinal associations between POPs concentration at year 2 and lipid concentrations, including LDLs, HDLs, triglycerides, and total cholesterol (years 2, 7, 10, 15, 20 and 25), and oxidized LDL (years 15 and 20) using repeated measures regression (generalized linear model, Toeplitz covariance matrix) to account for within individual correlations of outcome measures over time. Considering our recently reported effect modification by age of the associations between POPs and alterations in glucose homeostasis in this study population (Suarez-Lopez et al., 2015), we tested for interactions with age the associations of POPs with all outcomes. We also assessed effect modification by BMI, diabetes status, and exam period, and conducted analyses stratified by categories of the potential effect modifier. Statistical models adjusted for age, race, gender, BMI, smoking status (current, former, never), exam center, exam year, and diabetes diagnosis. Because this study was originally designed as a nested case-control study (participants with and without diabetes), we assessed whether diabetes status was an effect modifier on any of the main exposure-outcome associations; we observed no indication of effect modification; we thus pooled the data of both groups and adjusted for diabetes status. We excluded participant observations when there was concomitant use of lipid lowering medications. Analyses were conducted using SAS 9.4 (Cary, NC).

2. Results

2.1. Participant characteristics

At year 2, the mean age of participants was 27.2 years, 62% were female, 59% Afro-descendant, 41% Caucasian (Table 1). The mean BMI increased from 30.1 kg/m2 in year 2 to 34.1 kg/m2 in year 25. The means (range) of BMI across all time points in the upper and lower 50th percentile of BMI were 26.5 kg/m2 (17.9–31.99) and 38.6 kg/m2 (32.0, 58.8), respectively. Total cholesterol, LDL cholesterol and cholesterol-HDL ratio decreased between year 2 and 25, whereas triglyceride levels increased. Averages for most variables for years 7, 10 and 20 are not shown. The median concentrations of each POP across quartiles of the POPs summary scores and overall have been published (Suarez-Lopez et al., 2019).

The concentrations of organochlorine pesticides and PCBs in CARDIA participants in 1987–1988 were approximately 2–4 times and 1.5–7 times greater than participants of similar ages in 2003–2004 of the National Health and Nutrition Examination Survey (NHANES), respectively (Lee et al., 2010). The distributions of wet-weight and lipid-standardized POPs summary scores are described in Table 1. Unadjusted estimates in Table 1 show positive associations between POPs levels and age, total cholesterol, LDL, oxidized LDL and triglycerides, whereas negative associations were observed with BMI.

2.2. Year 2 PCBs and blood lipids over 23 years

Fig. 1 presents repeated-measure linear associations: positive associations were observed of the 23 PCB summary scores calculated using wet-weights (23 PCB summary-WW) and lipid-standardized concentrations (23 PCB summary-LP) with total cholesterol, triglycerides, LDLs, oxidized LDLs and cholesterol/HDL ratio. The associations with the 23 PCB summary-LP were 2–3 times weaker than the 23 PCB summary-WW and had borderline statistical (non)significance with oxidized LDL and triglycerides. Nonetheless, most associations remained statistically significant using 23 PCB summary-LP. Quadratic terms were non-significant, indicating that there was no evidence of curvature in the associations. We observed no interaction with age, or diabetes on any of the associations between PCBs and blood lipids. There was no exam period effect (interaction) with any of the outcomes, so we present pooled associations for all years in Figs. 1–3. Supplementary Fig. 1 presents associations between PCBs and blood lipids stratified by participant’s age at blood lipid assessment. The associations were consistent across all ages and support the non-significant interaction term of age. The associations for all lipid outcomes were stronger for non-dioxin-like PCBs than dioxin-like PCBs (Table 2). On average, the associations for non-dioxin-like PCB scores were stronger than dioxin-like PCB scores by 22% using the wet-weight summary scores and by 63% using the lipid-standardized score.

Fig. 1.

Fig. 1.

Fig. 1.

Open in a new tab

Longitudinal associations between quartiles of PCB summary scores (year 2) and blood lipids (years 2–25). Excludes observations of participants receiving lipid-lowering medication during blood lipid measurements.

Adjusted for age, race, gender, concurrent BMI, exam center, exam year, diabetes and smoking status. Lines between quartile points (diamonds) were added as a visual aid.

1β: Blood lipid difference (mg/dL) per SD increase of PCB Summary variable. No statistically significant quadratic associations were present.

Fig. 3.

Fig. 3.

Fig. 3.

Open in a new tab

Longitudinal associations between quartiles of PBB-153 (year 2) and blood lipids (year 2–25). Excludes observations of participants receiving lipid-lowering medication during blood lipid measurements.

Adjusted for age, race, gender, concurrent BMI, exam center, exam year, diabetes and smoking status.

Lines between quartile points (diamonds) were added as a visual aid.

aβ: Blood lipid difference (mg/dL) per SD increase of log PBB-153.

bPquadratic = 0.02, β = 53.1, βquadratic =−4.9.

cPquadratic = 0.03, β = 19.2, βquadratic =−2.6.

dPquadratic = 0.03, β = 66.9, βquadratic =−5.7.

ePquadratic = < 0.001, β = 46.7, βquadratic =−4.3.

fPquadratic = 0.002, β = 31.7, βquadratic =−2.9.

gPquadratic = 0.04, β = 14.5, βquadratic =−1.8.

hPquadratic = < 0.001, β = 1.72, βquadratic =−0.16.

iPquadratic = 0.05, β = 0.64, βquadratic = −0.08.

Table 2.

Longitudinal associations between dioxin-like (DL) and non-dioxin-like (NDL) PCB summary scores (year 2) and blood lipids (year 2–25). Excludes observations of participants receiving lipid-lowering medication during blood lipid measurements. N = 180.

Wet Weight Score βa
(95% CI)
 Lipid Standardized Score βa
(95% CI)
7 DL-PCBs 16 NDL-PCBs 7 DL-PCBs 16 NDL-PCBs
Cholesterol 12.4 (7.8, 17) 14.5 (9.8, 19.2) 3.9 (−1.1, 8.9) 6.5 (1.4, 11.7)
Triglycerides 20.3 (10.7, 29.9) 22.9 (13, 32.9) 7.0 (−3.1, 17.1)b 9.9 (−0.6, 20.4)e
HDL −1.1 (−2.8, 0.6) −1.3 (−3.0, 0.5) −0.3 (−2.1, 1.4) −0.5 (−2.3, 1.3)
LDL 10.0 (5.5, 14.5) 12.0 (7.4, 16.5) 3.1 (−1.6, 7.9)c 5.6 (0.7, 10.4)f
Oxidized LDL 5.9 (2.0, 9.8) 8.1 (4.1, 12.1) 2.4 (−1.6, 6.4) 4.5 (0.4, 8.7)
Cholesterol/HDL 0.42 (0.25, 0.60) 0.47 (0.30, 0.65) 0.16 (−0.02, 0.41)d 0.22 (0.03, 0.41)g
Open in a new tab

Adjusted for age, race, gender, concurrent BMI, exam center, exam year, diabetes and smoking status.

a

β Blood lipid difference (mg/dL) per SD increase of PCB summary score.

b

β (95% CI) for: a) BMI > 50th percentile: −1.2 (−11.8, 9.4), b) BMI ≤50th percentile: 12.2 (−2.1, 26.5).

c

β (95% CI) for: a) BMI > 50th percentile: 0.6 (−5.0, 6.1), b) BMI ≤50th percentile: 4.0 (−2.2, 10.1).

d

β (95% CI) for: a) BMI > 50th percentile: −0.04 (−0.26, 0.17), b) BMI ≤50th percentile: 0.30 (0.01, 0.07).

e

β (95% CI) for: a) BMI > 50th percentile: −1.6 (−12.9, 9.8), b) BMI ≤50th percentile: 15.3 (0.8, 29.8).

f

β (95% CI) for: a) BMI > 50th percentile: 2.2 (−3.8, 8.1), b) BMI ≤50th percentile: 4.8 (−1.5, 11.1).

g

β (95% CI) for: a) BMI > 50th percentile: −0.07 (−0.30, 0.16), b) BMI ≤50th percentile: 0.32 (0.08, 0.56).

We observed significant effect modification by BMI in the PCB-LP associations with some of the lipid outcomes (see “Interactions with BMI”, below).

2.3. Year 2 organochlorine pesticides and blood lipids over 23 years

The associations between the 8 OCP summary score based on wetweights (8 OCP summary-WW) and blood lipids were similar in magnitude to those of the PCB summary-WW, with positive and statistically significant associations with total cholesterol, triglycerides, LDLs, oxidized LDLs and cholesterol/HDL ratio (Fig. 2). However, all associations with lipids were non-significant when using the lipid adjusted summary score (8 OCP summary-LP). There was no evidence of curvilinear associations with any of the outcomes. We also observed no interaction with age, diabetes, BMI or examination period on any of the associations.

Fig. 2.

Fig. 2.

Fig. 2.

Open in a new tab

Longitudinal associations between quartiles of OCP summary scores (year 2) and blood lipids (years 2–25). Excludes observations of participants receiving lipid-lowering medication during blood lipid measurements.

Adjusted for age, race, gender, concurrent BMI, exam center, exam year, diabetes and smoking status. Lines between quartile points (diamonds) were added as a visual aid.

1β: Blood lipid difference (mg/dL) per SD increase of OCP Summary variable. No statistically significant quadratic associations were present.

2.4. Year 2 PBB-153 and blood lipids over 23 years

The associations of log PBB-153 wet-weight (log PBB-153-WW) with blood lipids followed the same general pattern as for the 23 PCB summary variables (Fig. 3). There were significant positive associations between log PBB-153-WW with total cholesterol, triglycerides, LDL, oxidized LDL and cholesterol/HDL ratio; however, all statistically significant associations were curvilinear. There was no association with HDL cholesterol. As expected, associations using the lipid standardized log PBB-153 (log PBB-153-LP) were weaker (and non-significant) than those of log PBB-153-WW. Significant positive quadratic associations were observed with cholesterol and oxidized LDLs with the lipid standardized variable. No interaction was observed with age, diabetes, or examination period on any of the associations. We observed significant effect modification by BMI in the PCB-LP associations with some of the lipid outcomes (see “Interactions with BMI”, below).

2.5. Interactions with BMI

We observed significant effect modification by BMI (averaged across all examination periods) on the association between 23-PCB-LP and LDL (interaction term p-value = 0.04), and between log-PBB-153-LP with oxidized LDL (p < 0.01) and cholesterol/HDL ratio (p = 0.01). As a result, we stratified the associations by a median split of BMI for both the 23-PCB-LP and log-PBB-153-LP. The associations of 23-PCB-LP among participants in the upper 50th percentile of BMI had stronger positive associations with LDL, although non-statistically significant, than participants in the lower 50th percentile. (Fig. 1). We also observed significant positive associations of PCBs with triglycerides and cholesterol/HDL ratio among participants in the upper 50th percentile of BMI, but null associations among participants in the lower 50th percentile. Similarly, effect modification by BMI was observed on the associations of triglycerides, LDL and Cholesterol/HDL ratio with the lipid-standardized 16 NDL-PCB and 7 DL-PCB scores. The BMI-stratified associations were very similar to those of the 23-PCB-LP score (Table 2).

For log PBB-153-LP, participants in the upper vs the lower 50th percentile of BMI (averaged across all time periods) had a stronger, although non-significant, linear association with oxidized LDL. The associations with cholesterol/HDL ratio, were equivalent but in opposite directions (both non-statistically significant) across the categories of BMI.

3. Discussion

We observed significant prospective associations between concentrations of PCBs measured in young adulthood and alterations in total cholesterol, triglycerides, LDLs, oxidized LDLs and cholesterol/HDL ratio measured at up to 7 points in time until the participants’ 5th or 6th decade of life. These associations, which were also partly observed for PBB 153, did not change as participants aged, unlike our previous analyses in this cohort in which we observed strong positive associations between year 2 POPs levels and alterations in glucose homeostasis as participants entered the 5th decade of life but not earlier (Suarez-Lopez et al., 2015). The present finding suggests that the disturbance of lipid metabolism due to exposure to POPs starts earlier than the disturbance of glucose metabolism. If replicated, the finding would be very relevant to understand POP-related pathogenesis.

Furthermore, the associations between most lipid measures and PCBs, were stronger among non-dioxin-like PCBs than among dioxin-like PCBs, a finding that was also partly observed in a large cross-sectional study in Canada (Singh and Chan, 2018).

Although the 8 OCP summary-WW had significant associations of a magnitude similar to that of the 23 PCB-WW, we observed no associations with any of the lipid outcomes when we tested associations with the 8 OCP-LP. This lack of association of OCPs when using the lipid standardized summary score may indicate that the significant associations observed with the wet weight OCP summary score may be due to the relationship between blood POPs levels with concurrent blood lipid levels, given the lipophilic properties of these POPs, rather than POP-induced disturbances of lipid metabolism.

We also observed notable effect modification by BMI on the associations of triglycerides, LDLs, and cholesterol/HDL ratio with PCBs, and to a lesser extent with PBB-153. Participants in the upper 50th percentile of BMI had stronger positive associations between: a) PCBs (lipid standardized) with triglycerides, LDL, and cholesterol/HDL ratio, and b) PBB-153 (lipid standardized) with oxidized LDL and cholesterol/HDL ratio, compared to those in the lower 50th percentile. This effect modification is not surprising considering that both POPs and obesity are each associated with systemic inflammation, and with alterations in hormone levels and in the metabolism of glucose and blood lipids (Greenberg and Obin, 2006; Horwich and Fonarow, 2010; Parekh and Anania, 2007; Penell et al., 2014; Vandenberg et al., 2012). In previous analyses of participants in our cohort, BMI was found to be an effect modifier for the associations between POPs and diabetes, in which associations (inverted U-shape) were only observed among obese participants (Lee et al., 2010). In the present study, we did not explore associations within multiple categories of BMI given the limited sample size of our study. Nonetheless, we were able to observe substantial differences in the associations between the upper and lower categories of BMI.

The present findings are the first to describe positive associations of POPs with oxidized LDL and the cholesterol/HDL ratio. The cholesterol/HDL ratio is a measure that has been found to predict best the risk of cardiovascular disease compared to other blood lipid measures (Prospective Studies Collaboration et al., 2007; The Emerging Risk Factors Collaboration* et al., 2009). Oxidized LDL plays an important role in atherogenesis by promoting an inflammatory environment and lipid deposition in the arterial wall, and elevated levels have been found in patients with coronary, carotid and femoral artery disease, as well as patients with metabolic syndrome and diabetes (Trpkovic et al., 2015).

Our study findings are concordant with other studies worldwide, although few prospective studies have assessed this to our knowledge. The 5-year prospective analyses from the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) cohort of elderly adults in Sweden (Penell et al., 2014) also showed positive relationships of POPs with cholesterol and LDL consistently among various PCB congeners; however, PCBs or OCPs were not associated with triglycerides. Like in our study, these associations were observed using both wet-weights and lipid-standardized values; however, the associations observed in PIVUS were slightly stronger using the lipid-standardized values, which is contrary to what we observed. Similar to our findings, OCPs also did not have consistent associations with blood lipids. Finally, effect modification by BMI was also observed in the PIVUS study; however, increased BMI between baseline and 5 years was related to a smaller increase in total cholesterol and LDL.

Previous analyses among non-diabetics in our study population showed U-shaped associations between various plasma PCB congeners measured on year 2 and HDLs at year 20, inverted U-shaped associations with triglycerides at year 20 and no associations with LDLs at year 20 (Lee et al., 2011b) in models adjusting for year 2 cholesterol and triglycerides. In the present expanded analyses, which incorporated participants with and without diabetes who were not receiving cholesterol lowering medications, we observed no curvature on the associations with PCBs and OCPs, but did observe some evidence of curvilinear associations with PBB-153.

Other studies, albeit cross-sectional, have reported positive associations between POPs and triglycerides among participants who did not use lipid-lowering medication, as in our study. A study of patients examined in Japan observed positive associations of PCBs with total cholesterol and triglycerides, and no associations with HDLs (as we observed) (Tokunaga and Kataoka, 2003). A cross-sectional study of residents of Anniston, USA, reported positive associations with total cholesterol and triglycerides with summary scores of PCBs and organochlorine pesticides. However, no associations were observed with LDLs. A cross-sectional study of 335 Native Americans also found positive associations of PCB wet-weights with total cholesterol and triglycerides (Goncharov et al., 2008). Finally, a large cross-sectional study in Canada also observed positive associations between PCBs and cholesterol, triglycerides, LDLs but not HDLs (Singh and Chan, 2018). As in our study, they observed slightly stronger associations for high cholesterol (OR for quartile 4 vs quartile 1) with non-dioxin-like PCBs (4.61) compared to dioxin-like PCBs (4.05); however, the linear associations did not seem to differ between non-dioxin-like PCBs and dioxin-like PCBs.

Studies have observed significant positive associations of POPs with cardiovascular diseases (Goncharov et al., 2010; Lee et al., 2012; Lind and Lind, 2012). Considering that high total cholesterol and LDLs and low HDLs are established risk factors for cardiovascular disease (Prospective Studies Collaboration et al., 2007; The Emerging Risk Factors Collaboration* et al., 2009), it is plausible that blood lipid alterations may mediate the association between serum POPs levels and cardiovascular disease as observed by Goncharov et al. (2008).

In-vitro and human studies have proposed that POPs and other endocrine disruptors may alter the regulation of blood lipids at various levels including heightened stimulation of liver synthesis of cholesterol and triglycerides (Goncharov et al., 2008), down-regulation of genes that regulate lipid levels (Ruzzin et al., 2010), and altering nuclear receptor signaling leading to heightened fat deposition and synthesis, and obesity (Grün and Blumberg, 2009, 2007).

Analyses of POPs have generally used POPs values standardized for lipids (POPs concentration divided by total lipid content) or statistical models that adjust for concurrent blood lipid levels, given that POPs are predominantly carried in the lipid component of blood (Porta et al., 2009). Unfortunately, the high correlation over time of all blood lipid components precludes us from adjusting for year-2 blood lipids, considering that blood lipids over time are the outcomes of this study. The correlations between year-2 and years 7, 10, 15 and 20 ranged from 0.71 to 0.51 for total cholesterol and from 0.67 to 0.58 for triglycerides. To circumvent this issue, we present analyses using POP summary scores based on wet-weights and lipid standardized values (wet-weights / total lipids). The wet-weight POPs variables do not account for the levels of a baseline outcome value (e.g. cholesterol and triglycerides), thus avoiding over adjustment. However, this analysis fails to account for the correlation between blood POPs and blood lipid levels attributable to the lipophilicity of POPs. The lipid-standardized POPs variable analyses are a good middle ground between analyses involving wet-weights and analyses adjusting for blood lipids, as the lipid-standardized variables account for varying levels of circulating POPs across blood lipid levels (although less efficiently than adjusting for total lipids), and the statistical model does not adjust for a lipid outcome variable. The true exposure-outcome associations are probably closest to the results using the lipid-standardized variables in this context. The associations were strongest among the wet-weight POP summary scores, but the significance and direction of the associations were generally concordant with the results using the lipid standardized POP summary scores.

We preferred to present associations for POP summary scores rather than for specific POPs because it more realistically estimates exposure given that background environmental concentrations are composed of a mix of POPs; hence, summary variables also account for potential interactions between POPs to some extent. However, composite variables do not allow us to assess interactions between specific POPs. The present summary variables also tend to be less variable than individual compounds and allowed us to dramatically reduce the number of regressions analyzed (compared to individual POPs) which reduced the likelihood of type-1 errors: finding significant associations when in fact there are none. A drawback of composite variables is that they include information of POPs that could be unrelated or inversely related to the outcomes of interest, which therefore, could weaken the strength of the associations of POPs that are related.

A limitation of this study is that its sample population originates from a nested case-control study which included participants who became diabetic during follow-up (cases) and controls who were selected for having normal fasting glucose concentrations through year 20 of follow-up. Although people with and without diabetes have different metabolisms, we did not observe effect modification by case-control status, which allowed us to pool the data of both groups.

An additional limitation of our study is the single measurement of POPs which pre-dated outcome measurements by up to 23 years. It is estimated that the concentrations of POPs in adults in the U.S. have progressively decreased by 80% since the 1980’s (Aylward and Hays, 2002; Centers for Disease Control and Prevention, 2009; Lorber, 2002). We did not observe effect modification as participants became older, which may be an indication that the group POPs levels may not have had drastic changes over time. Subsequent longitudinal studies should include quantifications of POPs at various ages. Although speculative, it is plausible that other unmeasured POPs that may be correlated with PCB levels may be affecting lipid metabolism.

The relatively small sample size is a limitation, although statistical power seemed sufficient for most outcomes. An important strength of this study is its prospective design with data points for most participants across 7 examinations over a 23-year follow-up. This strengthened our findings and power (a larger number of observations) compared to other study designs with similar sample sizes.

4. Conclusions

Lipid standardized PCBs and PBB-153 measured in young adulthood were positively associated with alterations in total cholesterol, triglycerides, LDLs, oxidized LDLs and the cholesterol-HDL ratio measured in serum samples concurrently, and prospectively up to 23 years later. These associations did not vary as participants aged but they were stronger among participants with higher BMI. Further longitudinal studies with multiple measures of POPs overtime are needed.

Supplementary Material

supplementary material

Acknowledgments

We would like to thank Dr. David R Jacobs Jr. for his prior work on persistent organic pollutants in the CARDIA study which allowed us to conduct the present analyses. The Coronary Artery Risk Development in Young Adults Study (CARDIA) is supported by contracts HHSN268201300025C, HHSN268201300026C, HHSN268201300027C, HHSN268201300028C, HHSN268201300029C, and HHSN268200900041C from the National Heart, Lung, and Blood Institute (NHLBI), the Intramural Research Program of the National Institute on Aging (NIA), and an intra-agency agreement between NIA and NHLBI (AG0005). The Young Adult Longitudinal Trends in Atherosclerosis (YALTA) study is supported by R01HL53560. The development of this manuscript was supported by a JPB Environmental Health Fellowship award granted by the JPB Foundation and managed by the Harvard T. H. Chan School of Public Health.

Abbreviations:

POPs

persistent organic pollutants

PCB

polychlorinated biphenyls

OCP

organochlorine pesticides

HDL

High Density Lipoprotein

LDL

Low Density Lipoprotein

BMI

Body mass index

YALTA

Young Adults Longitudinal Trends in Atherosclerosis

CARDIA

Coronary Artery Risk Development in Young Adults

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.etap.2018.12.018.

Footnotes

Competing financial interests

None.

Transparency document

The Transparency document associated with this article can be found in the online version.

References

  1. Aminov Z, Haase RF, Pavuk M, Carpenter DO, Anniston Environmental Health Research Consortium, 2013. Analysis of the effects of exposure to polychlorinated biphenyls and chlorinated pesticides on serum lipid levels in residents of Anniston, Alabama. Environ. Health 12, 108. 10.1186/1476-069X-12-108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arrebola JP, Ocaña-Riola R, Arrebola-Moreno AL, Fernández-Rodríguez M, Martin-Olmedo P, Fernández MF, Olea N, 2014. Associations of accumulated exposure to persistent organic pollutants with serum lipids and obesity in an adult cohort from Southern Spain. Environ. Pollut 195, 9–15. 10.1016/j.envpol.2014.08.003. [DOI] [PubMed] [Google Scholar]
  3. Aylward L, Hays S, 2002. Temporal trends in human TCDD body burden: decreases over three decades and implications for exposure levels. J. Expo. Anal. Environ. Epidemiol 12, 319–328. [DOI] [PubMed] [Google Scholar]
  4. Baker EL, Landrigan PJ, Glueck CJ, Zack MM, Liddle JA, Burse VW, Housworth WJ, Needham LL, 1980. Metabolic consequences of exposure to polychlorinated biphenyls (PCB) in sewage sludge. Am. J. Epidemiol 112, 553–563. [DOI] [PubMed] [Google Scholar]
  5. Barr JR, Maggio VL, Barr DB, Turner WE, Sjödin A, Sandau CD, Pirkle JL, Needham LL, Patterson DG, 2003. New high-resolution mass spectrometric approach for the measurement of polychlorinated biphenyls and organochlorine pesticides in human serum. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci 794, 137–148. [DOI] [PubMed] [Google Scholar]
  6. Centers for Disease Control and Prevention, 2009. Fourth National Report on Human Exposure to Environmental Chemicals, Environmental Health. Centers for Disease Control and Prevention, Atlanta, GA. [Google Scholar]
  7. Contreras López MC, 2003. Determination of potentially bioaccumulating complex mixtures of organochlorine compounds in wastewater: a review. Env. Int 28, 751–759. 10.1016/s0160-4120(02)00120-4. [DOI] [PubMed] [Google Scholar]
  8. Elobeid MA, Padilla MA, Brock DW, Ruden DM, Allison DB, 2010. Endocrine disruptors and obesity: an examination of selected persistent organic pollutants in the NHANES 1999-2002 data. Int. J. Environ. Res. Public Health 7, 2988–3005. 10.3390/ijerph7072988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Friedewald WT, Levy RI, Fredrickson DS, 1972. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem 18, 499–502. [PubMed] [Google Scholar]
  10. Friedman GD, Cutter GR, Donahue RP, Hughes GH, Hulley SB, Jacobs DR Jr., Liu K, Savage PJ, 1988. CARDIA: study design, recruitment, and some characteristics of the examined subjects. J. Clin. Epidemiol 41, 1105–1116. [DOI] [PubMed] [Google Scholar]
  11. Goncharov A, Haase RF, Santiago-Rivera A, Morse G, McCaffrey RJ, Rej R, Carpenter DO, 2008. High serum PCBs are associated with elevation of serum lipids and cardiovascular disease in a Native American population. Environ. Res 106, 226–239. 10.1016/j.envres.2007.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Goncharov A, Bloom M, Pavuk M, Birman I, Carpenter DO, 2010. Blood pressure and hypertension in relation to levels of serum polychlorinated biphenyls in residents of Anniston, Alabama. J. Hypertens 28, 2053–2060. 10.1097/HJH.0b013e32833c5f3e.. [DOI] [PubMed] [Google Scholar]
  13. Greenberg AS, Obin MS, 2006. Obesity and the role of adipose tissue in inflammation and metabolism. Am. J. Clin. Nutr 83, 461S–465S. [DOI] [PubMed] [Google Scholar]
  14. Grün F, Blumberg B, 2007. Perturbed nuclear receptor signaling by environmental obesogens as emerging factors in the obesity crisis. Rev. Endocr. Metab. Disord 8, 161–171. 10.1007/s11154-007-9049-x. [DOI] [PubMed] [Google Scholar]
  15. Grün F, Blumberg B, 2009. Endocrine disrupters as obesogens. Mol. Cell. Endocrinol 304, 19–29. 10.1016/j.mce.2009.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Holvoet P, Macy E, Landeloos M, Jones D, Jenny NS, Nancy JS, Van de Werf F, Tracy RP, 2006. Analytical performance and diagnostic accuracy of immunometric assays for the measurement of circulating oxidized LDL. Clin. Chem 52, 760–764. 10.1373/clinchem.2005.064337. [DOI] [PubMed] [Google Scholar]
  17. Horwich TB, Fonarow GC, 2010. Glucose, obesity, metabolic syndrome, and diabetes. J. Am. Coll. Cardiol 55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ibrahim MM, Fjære E, Lock E-JJ, Naville D, Amlund H, Meugnier E, Le Magueresse Battistoni B, Frøyland L, Madsen L, Jessen N, Lund S, Vidal H, Ruzzin J, Fjaere E, Lock E-JJ, Naville D, Amlund H, Meugnier E, Le Magueresse Battistoni B, Froyland L, Madsen L, Jessen N, Lund S, Vidal H, Ruzzin J, 2011. Chronic consumption of farmed salmon containing persistent organic pollutants causes insulin resistance and obesity in mice. PLoS One 6, e25170. 10.1371/journal.pone.0025170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ibrahim MM, Fjære E, Lock E-JJ, Frøyland L, Jessen N, Lund S, Vidal H, Ruzzin J, Fjaere E, Lock E-JJ, Froyland L, Jessen N, Lund S, Vidal H, Ruzzin J, 2012. Metabolic impacts of high dietary exposure to persistent organic pollutants in mice. Toxicol. Lett 215, 8–15. 10.1016/j.toxlet.2012.09.022. [DOI] [PubMed] [Google Scholar]
  20. Lee D-H, Lee I-K, Song K, Steffes M, Toscano W, Baker Ba, Jacobs DR, 2006. A strong dose-response relation between serum concentrations of persistent organic pollutants and diabetes: results from the National Health and Examination Survey 1999-2002. Diabetes Care 29, 1638–1644. 10.2337/dc06-0543. [DOI] [PubMed] [Google Scholar]
  21. Lee DH, Lee IK, Jin SH, Steffes M, Jacobs DR Jr., 2007. Association between serum concentrations of persistent organic pollutants and insulin resistance among nondiabetic adults: results from the National Health and Nutrition Examination Survey 1999-2002. Diabetes Care 30, 622–628. 10.2337/dc06-2190. [DOI] [PubMed] [Google Scholar]
  22. Lee D-H, Steffes MW, Sjödin A, Jones RS, Needham LL, Jacobs DR, 2010. Low dose of some persistent organic pollutants predicts type 2 diabetes: a nested case-control study. Environ. Health Perspect 118, 1235–1242. 10.1289/ehp.0901480. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lee D-H, Lind PM, Jacobs DR, Salihovic S, van Bavel B, Lind L, 2011a. Polychlorinated biphenyls and organochlorine pesticides in plasma predict development of type 2 diabetes in the elderly: the prospective investigation of the vasculature in Uppsala Seniors (PIVUS) study. Diabetes Care 34, 1778–1784. 10.2337/dc10-2116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lee D-HH, Steffes MW, Sjodin A, Jones RS, Needham LL, Jacobs DR Jr., Sjödin A, Jones RS, Needham LL, Jacobs DR, 2011b. Low dose organochlorine pesticides and polychlorinated biphenyls predict obesity, dyslipidemia, and insulin resistance among people free of diabetes. PLoS One 6, e15977. 10.1371/journal.pone.0015977. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lee D-H, Lind PM, Jacobs DR, Salihovic S, van Bavel B, Lind L, 2012. Background exposure to persistent organic pollutants predicts stroke in the elderly. Environ. Int 47, 115–120. 10.1016/j.envint.2012.06.009. [DOI] [PubMed] [Google Scholar]
  26. Lind L, Lind PM, 2012. Can persistent organic pollutants and plastic-associated chemicals cause cardiovascular disease? J. Intern. Med 271, 537–553. 10.1111/j.1365-2796.2012.02536.x. [DOI] [PubMed] [Google Scholar]
  27. Lorber M, 2002. A pharmacokinetic model for estimating exposure of Americans to dioxin-like compounds in the past, present, and future. Sci. Total Environ 288, 81–95. [DOI] [PubMed] [Google Scholar]
  28. Magliano DJ, Loh VH, Harding JL, Botton J, Shaw JE, 2014. Persistent organic pollutants and diabetes: a review of the epidemiological evidence. Diabetes Metab. 40, 1–14. 10.1016/j.diabet.2013.09.006. [DOI] [PubMed] [Google Scholar]
  29. Milbrath MO, Wenger Y, Chang CW, Emond C, Garabrant D, Gillespie BW, Jolliet O, 2009. Apparent half-lives of dioxins, furans, and polychlorinated biphenyls as a function of age, body fat, smoking status, and breast-feeding. Environ. Health Perspect 117, 417–425. 10.1289/ehp.11781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Moysich KB, Ambrosone CB, Mendola P, Kostyniak PJ, Greizerstein HB, Vena JE, Menezes RJ, Swede H, Shields PG, Freudenheim JL, 2002. Exposures associated with serum organochlorine levels among postmenopausal women from western New York State. Am. J. Ind. Med 41, 102–110. [DOI] [PubMed] [Google Scholar]
  31. Navab M, Reddy ST, Van Lenten BJ, Fogelman AM, 2011. HDL and cardiovascular disease: atherogenic and atheroprotective mechanisms. Nat. Rev. Cardiol 8, 222–232. 10.1038/nrcardio.2010.222. [DOI] [PubMed] [Google Scholar]
  32. Parekh S, Anania FA, 2007. Abnormal lipid and glucose metabolism in obesity: implications for nonalcoholic fatty liver disease. Gastroenterology 132, 2191–2207. 10.1053/j.gastro.2007.03.055. [DOI] [PubMed] [Google Scholar]
  33. Parthasarathy S, Raghavamenon A, Garelnabi MO, Santanam N, 2010. Oxidized low-density lipoprotein. Methods Mol. Biol 610, 403–417. 10.1007/978-1-60327-029-8_24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Penell J, Lind L, Salihovic S, van Bavel B, Lind PM, 2014. Persistent organic pollutants are related to the change in circulating lipid levels during a 5 year follow-up. Environ. Res 134, 190–197. 10.1016/j.envres.2014.08.005. [DOI] [PubMed] [Google Scholar]
  35. Porta M, Puigdomènech E, Ballester F, Selva J, Ribas-Fitó N, Llop S, López T, 2008. Monitoring concentrations of persistent organic pollutants in the general population: the international experience. Environ. Int 34, 546–561. 10.1016/j.envint.2007.10.004. [DOI] [PubMed] [Google Scholar]
  36. Porta M, Jariod M, López T, Pumarega J, Puigdomènech E, Marco E, Malats N, Grimalt JO, Real FX, 2009. Correcting serum concentrations of organochlorine compounds by lipids: alternatives to the organochlorine/total lipids ratio. Environ. Int 35, 1080–1085. 10.1016/j.envint.2009.06.004. [DOI] [PubMed] [Google Scholar]
  37. Prospective Studies Collaboration, Lewington S, Whitlock G, Clarke R, Sherliker P, Emberson J, Halsey J, Qizilbash N, Peto R, Collins R, 2007. Blood cholesterol and vascular mortality by age, sex, and blood pressure: a meta-analysis of individual data from 61 prospective studies with 55,000 vascular deaths. Lancet (London, England) 370, 1829–1839. 10.1016/S0140-6736(07)61778-4. [DOI] [PubMed] [Google Scholar]
  38. Remillard RBJ, Bunce NJ, 2002. Linking dioxins to diabetes: epidemiology and biologic plausibility. Environ. Health Perspect 110, 853–858. 10.1289/ehp.02110853. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ruzzin J, Petersen R, Meugnier E, Madsen L, Lock E-J, Lillefosse H, Ma T, Pesenti S, Sonne SB, Marstrand TT, Malde MK, Du Z-Y, Chavey C, Fajas L, Lundebye A-K, Brand CL, Vidal H, Kristiansen K, Froyland L, Frøyland L, 2010. Persistent organic pollutant exposure leads to insulin resistance syndrome. Environ. Health Perspect 118, 465–471. 10.1289/ehp.0901321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Schecter A, Colacino J, Haffner D, Patel K, Opel M, Papke O, Birnbaum L, Päpke O, Birnbaum L, 2010. Perfluorinated compounds, polychlorinated biphenyls, and organochlorine pesticide contamination in composite food samples from Dallas, Texas, USA. Environ. Health Perspect 118, 796–802. 10.1289/ehp.0901347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Singh K, Chan HM, 2018. Association of blood polychlorinated biphenyls and cholesterol levels among Canadian Inuit. Environ. Res 160, 298–305. 10.1016/J.ENVRES.2017.10.010. [DOI] [PubMed] [Google Scholar]
  42. Sjödin A, Jones RS, Lapeza CR, Focant J-F, McGahee EE, Patterson DG, 2004. Semiautomated high-throughput extraction and cleanup method for the measurement of polybrominated diphenyl ethers, polybrominated biphenyls, and polychlorinated biphenyls in human serum. Anal. Chem 76, 1921–1927. 10.1021/ac030381+. [DOI] [PubMed] [Google Scholar]
  43. Suarez-Lopez JR, Lee D-H, Porta M, Steffes MW, Jacobs DR Jr., 2015. Persistent organic pollutants in young adults and changes in glucose related metabolism over a 23-year follow-up. Environ. Res 137, 485–494. 10.1016/j.envres.2014.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Suarez-Lopez JR, Gross MD, Lee D-H, 2019. Summary data of serum concentrations of 32 persistent organic pollutants in young adults in relation to summary scores of persistent organic pollutants. Data Br. 103720. 10.1016/J.DIB.2019.103720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tang M, Chen K, Yang F, Liu W, 2014. Exposure to organochlorine pollutants and type 2 diabetes: a systematic review and meta-analysis. PLoS One 9, e85556. 10.1371/journal.pone.0085556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. The Emerging Risk Factors Collaboration*, Di Angelantonio E, Sarwar N, Perry P, Kaptoge S, Ray KK, Thompson A, Wood AM, Lewington S, Sattar N, Packard CJ, Collins R, Thompson SG, Danesh J, 2009. Major lipids, apolipoproteins, and risk of vascular disease. JAMA 302, 1993. 10.1001/jama.2009.1619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Tokunaga S, Kataoka K, 2003. A longitudinal analysis on the association of serum lipids and lipoproteins concentrations with blood polychlorinated biphenyls level in chronic “Yusho” patients. Fukuoka Igaku Zasshi 94, 110–117. [PubMed] [Google Scholar]
  48. Trpkovic A, Resanovic I, Stanimirovic J, Radak D, Mousa SA, Cenic-Milosevic D, Jevremovic D, Isenovic ER, 2015. Oxidized low-density lipoprotein as a biomarker of cardiovascular diseases. Crit. Rev. Clin. Lab. Sci 52, 70–85. 10.3109/10408363.2014.992063. [DOI] [PubMed] [Google Scholar]
  49. Vandenberg LN, Colborn T, Hayes TB, Heindel JJ, Jacobs DR, Lee D-H, Shioda T, Soto AM, vom Saal FS, Welshons WV, Zoeller RT, Myers JP, 2012. Hormones and endocrine-disrupting chemicals: low-dose effects and nonmonotonic dose responses. Endocr. Rev 33, 378–455. 10.1210/er.2011-1050. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementary material
NIHMS2137346-supplement-supplementary_material.docx (36.6KB, docx)

Data Availability Statement

The Transparency document associated with this article can be found in the online version.

RESOURCES