Disparities in EnvironmentalExposures to Endocrine-DisruptingChemicals andDiabetesRisk in Vulnerable Populationshttps://doi.org/10.2337/dc16-2765

Burgeoning epidemiological, animal, and cellular data link environmental endocrine-disrupting chemicals (EDCs) to metabolic dysfunction. Disproportionate exposure todiabetes-associated EDCs may be an underappreciated contributor to disparities inmetabolic disease risk. The burden of diabetes is not uniformly borne by Americansociety; rather, this disease disproportionately affects certain populations, includingAfrican Americans, Latinos, and low-income individuals. The purpose of this studywas to review the evidence linking unequal exposures to EDCswith racial, ethnic, andsocioeconomic diabetes disparities in the U.S.; discuss social forces promoting thesedisparities; andexplore potential interventions. Articles examining the links betweenchemical exposures and metabolic disease were extracted from the U.S. NationalLibrary ofMedicine for the period of 1966 to 3 December 2016. EDCs associatedwithdiabetes in the literature were then searched for evidence of racial, ethnic, andsocioeconomic exposure disparities. Among Latinos, African Americans, and low-income individuals, numerous studies have reported significantly higher exposuresto diabetogenic EDCs, including polychlorinated biphenyls, organochlorine pesti-cides, multiple chemical constituents of air pollution, bisphenol A, and phthalates.This review reveals that unequal exposure to EDCs may be a novel contributor todiabetes disparities. Efforts to reduce the individual and societal burden of diabetesshould include educating clinicians on environmental exposures that may increasedisease risk, strategies to reduce those exposures, and social policies to addressenvironmental inequality as a novel source of diabetes disparities.

Diabetes is a complex and devastating metabolic disease that arises from impairmentsin insulin production and/or action with consequential derangements in global energymetabolism. In the U.S., diabetes is the leading cause of adult blindness, kidney failure,and nontraumatic amputations; moreover, it is a central driver of cardiovascular dis-ease, the leading cause of death among people with diabetes. Diabetes disproportion-ately afflicts African Americans, Latinos, and low-income individuals. Compared withnon-Hispanic whites, the risk of developing diabetes is estimated to be 66% higher forHispanics and77%higher for AfricanAmericans (1). Indeed, 17.9%ofAfricanAmericansand 20.5% of Mexican Americans have diabetes compared with only 9.1% of non-Hispanic whites, and these disparities in diabetes prevalence have been amplifiedover the past decade (2). Furthermore, age-adjusted diabetes mortality rates aresignificantly higher amongHispanics and non-Hispanic blacks than non-Hispanic whites(3). Understanding the complete array of factors that contribute to racial and ethnicdifferences in the pathogenesis of metabolic disease is critical for addressing thedisproportionate burden of diabetes in communities of color.

1Committee on Molecular Metabolism and Nu-trition, University of Chicago, Chicago, IL2College of Food, Agriculture and EnvironmentalSciences, School of Environment and Natural Re-sources, Ohio State University, Columbus, OH3Environmental and Occupational Health Sci-ences Division, School of Public Health, Univer-sity of Illinois at Chicago, Chicago, IL4Division of Endocrinology, Diabetes, and Me-tabolism, Department of Medicine, Universityof Illinois at Chicago, Chicago, IL

Corresponding author: Robert M. Sargis, rsargis@uic.edu.

Received 27 December 2016 and accepted 23September 2017.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-2765/-/DC1.

© 2017 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notfor profit, and the work is not altered. More infor-mation is available at http://www.diabetesjournals.org/content/license.

Daniel Ruiz,1 Marisol Becerra,2

Jyotsna S. Jagai,3 Kerry Ard,2 and

Robert M. Sargis4

Diabetes Care 1

REV

IEW

Diabetes Care Publish Ahead of Print, published online November 15, 2017

Although physical inactivity and caloricexcess undoubtedly are risk factors,emerging evidence implicates environ-mental endocrine-disrupting chemicals(EDCs) as contributors to the diabetes ep-idemic. The Endocrine Society defines anEDC as an exogenous chemical ormixtureof chemicals that interferes with any as-pect of hormone action (4). Of note, thedramatic rise in U.S. diabetes rates corre-lates closely with synthetic chemical pro-duction (5), and these associations arenow supported by epidemiological, ani-mal, and cellular data that demonstratethat EDCs can interferewith insulin secre-tion and action aswell as with other path-ways that regulate glucose homeostasis.Despite a history of environmental pollu-tion disproportionately affecting commu-nities of color in the U.S. (6), the potentialcontribution of environmental toxicantsto racial and ethnic differences in diabe-tes risk is underappreciated.The issue of environmental injustice

first entered widespread consciousness in1982 when residents of the predomi-nantly African American community ofWarren County, North Carolina, madenational news by laying themselves acrossa rural road to prevent encroaching trucksfrom dumping dirt laden with polychlori-nated biphenyls (PCBs) in their community(6). This media attention prompted em-pirical examination of the community’sclaim that toxicwaste facilitieswere beingdisproportionately sited in low-incomecommunities and communities of color.This research has grown substantiallysince the 1980s, with the majority of evi-dence showing racial and socioeconomicdisparities in exposures tomyriad environ-mental hazards (6). In addition to higherexposure to air pollution nationwide (7),unequal exposures amongpeople of colorare also rooted in patterns of occupation,housing conditions, and neighborhood in-frastructure (8,9). This article reviews thestate of the evidence linking ethnic, racial,and socioeconomic disparities in pollut-ant exposure in the U.S. to EDCs linkedto diabetes.

UNEQUAL ENVIRONMENTALEXPOSURES AND DIABETES RISK

Scientific evidence linking EDCs with thedevelopment of diabetes and other meta-bolic disorders continues to grow. Of note,exposures to several toxicants have beenprospectively linked to diabetes risk, includ-ing PCBs, organochlorine (OC) pesticides,

various chemical constituents of air pol-lution, bisphenol A (BPA), and phtha-lates (Table 1); moreover, exposure tothese EDCs is higher amongAfrican Amer-icans, Latinos, and low-income individuals(Supplementary Table 1). These unequalexposures raise the possibility that EDCsare underappreciated contributors to di-abetes disparities.

PCBsIntroduced in the U.S. in the 1930s for avariety of industrial purposes, PCBs area class of synthetic compounds wherevarious combinations of hydrogen atomson the biphenyl (C12H10) structure aresubstituted with chlorine, resulting in209 congeners designated by a uniquenumber reflecting the extent and positionof their chlorination (e.g., PCB 153). Al-though bannedby theU.S. EnvironmentalProtection Agency in 1977, PCBs remaindetectable in human tissues as a resultof their environmental and biologi-cal persistence (37). Higher PCB expo-sures among African Americans havebeen documented since the 1960s (38)(Supplementary Table 1). Ongoing humanexposure to PCBs is due to the legacy ofcontamination in food, including certainfish (39); however, additional exposuresources include leaching from contami-nated industrial sites and indoor con-struction materials (40,41). PCB waste isfound in Superfund and toxic waste sitesthat are concentrated in neighborhoodsof color (6). Although catfish consump-tion has been suggested as the main con-tributor to increased PCB levels in AfricanAmericans (42), the historical siting ofPCB production and disposal sites in pre-dominantly black communities is likely asignificant additional contributor to in-creased contamination of locally sourcedfoods. Oneexample of this phenomenon isAnniston, Alabama, a PCB manufacturingcity from 1929 to 1971. African Americansnot only lived closer to a formerMonsantoPCB manufacturing plant but also hadPCB levels three times higher thanwhites living in Anniston (43). Consump-tion of local fish and livestock were thestrongest predictors of higher serum PCBlevels among African Americans (44),whereas consumption of local dairy prod-ucts and dredging near another PCB-contaminatedSuperfund site alsopredictedhigher cord blood PCB levels in infants (45).

A large body of evidence, including pro-spective epidemiological studies, supports

the hypothesis that PCBs are metabolicdisease risk factors. For example, residen-tial proximity to PCB-contaminatedwastesites is associated with higher diabeteshospitalization rates (46). Among femaleresidents of Anniston, serum PCB levelswere significantly associated with diabe-tes (47), whereas in a separate studywith 25 years of follow-up, women withhigher PCB levels exhibited increased di-abetes incidence (incidence density ratio2.33 [95% CI 1.25–4.34]) (10). Similarly,women exposed to PCB-laced rice branoil during the Yucheng poisoning eventin Taiwan also had an increased riskof developing diabetes (odds ratio [OR]2.1 [95% CI 1.1–4.5]), with markedlyhigher risk among those who developedchloracne, a cutaneous manifestation ofdioxin-like PCB exposure (OR 5.5 [95% CI2.1–13.4]) (11). A meta-analysis thatpooled data from the Nurses’ HealthStudy (NHS) with six prospective studiesshowed that total PCBs were associatedwith incident diabetes (OR 1.70 [95% CI1.28–2.27]) (13). Further supportingthese prospective links between PCB ex-posure and diabetes are data from cohortstudies, including the Prospective Investi-gation of the Vasculature in Uppsala Se-niors (PIVUS) (16) and a group followedfor nearly 20 years (15). Finally, althoughnot reaching statistical significance, astudy of Swedish women suggested thathigher levels of PCB 153 were similarlyassociated with increased rates of type 2diabetes diagnosed after .6 years offollow-up (OR 1.6 [95% CI 0.61–4.0]) (14).Collectively, these data suggest an associ-ation between PCBs and diabetes risk, es-pecially among women; however, somediscrepant findings exist in the literature.In a study of Great Lakes sport fish con-sumers, PCB 118 and total PCBs were notassociated with diabetes (12), and in aFlemish study that adjusted for correlatedexposures, PCBs showed a negative asso-ciation with self-reported diabetes (18).Despite these discrepancies, a meta-analysis of both cross-sectional and pro-spective studies published before March2014 showed that in aggregate, totalPCBs are associated with increased dia-betes risk (relative risk [RR] 2.39 [95% CI1.86–3.08]) (17). Takenwithin the contextof animal and cellular data demonstrat-ing that PCBs alter metabolic function(Supplementary Table 2), this evidencecollectively suggests that PCBs contributeto diabetes risk and disparities.

2 Environmental Inequality and Diabetes Risk Diabetes Care

Tab

le1—

Prosp

ective

studiesdocu

men

tingasso

ciationsbetwee

nEDC

exposu

rean

ddiabetes

risk

Reference

Population

Outcomeandcomparison

Effect

estimate(95%

CI)

PCBs

Vasiliuet

al.,2006

(10)

1,38

4subjectswithoutdiabetes

intheMichigan

polybrominated

biphen

ylscohortfollowed

for25

years

Inciden

tdiabetes

inwomen

withthehighestvs.lowest

serum

PCBlevels

IDR:2.33(1.25–4.34)*

Wanget

al.,2008

(11)

378subjectsand37

0matched

referentsfrom

theYuchen

gpoisoningin

Taiwan

inthe19

70s

Inciden

tdiabetes

inwomen

whoconsumed

rice

branoil

lacedwithPC

Bsas

wellasasubgroupwhodeveloped

chloracne,aman

ifestationofdioxin-likePC

Bexposure

OR:2.1(1.1–4.5)*;C

hloracne

OR:5.5(2.1–13.4)*

Turyket

al.,2009

(12)

471Great

Lakessportfish

consumerswithoutdiabetes

followed

from

1994

/199

5to

2005

Inciden

tdiabetes

amongthehighestvs.lowesttertile

ofPC

Blevels

TotalP

CBsIRR:1.8(0.6–5.0);PC

B118IRR:1.3

(0.5–3.0)

Wuet

al.,2013

(13)

Twocase-controlstudiesofwomen

withoutdiabetes

from

theNHSandameta-an

alysisofpooleddata

withsixadditionalprospective

studies

Inciden

tdiabetes

afterpoolingofdataandcomparing

highestPC

Bexposure

groupwiththereferent

PooledOR:1.70(1.28–2.27)*

Rignell-Hydbo

met

al.,2009

(14)

Case-controlstudyofwomen

age50

–59yearsin

southern

Sweden

Inciden

tdiabetes

in39

patientsandmatched

control

subjectsafter$6yearsof

follow-upcomparing

the

highestqu

artileof

PCBlevelswiththereferent

OR:1.6(0.61–4.0)

Leeet

al.,2010

(15)

90patientsandcontrolsubjectsinanestedcase-control

studyfollowed

for;18

years

Inciden

tdiabetes

comparingsecondsextile

orquartile

withthereferentforasummarymeasure

of16

POPs,

including12

PCBsas

wellasindividualPC

Bs

PCBsum

OR:5.3*;PC

B187OR:2.8(1.1–7.4)*

Leeet

al.,2011

(16)

725participan

tsfrom

thePIVUSstudy

Inciden

tdiabetes

comparingasummarymeasure

of14

PCBsacross

quintileswiththereferent

Quintile2OR:4.5(0.9–23.5);Q

uintile

3OR:

5.1(1.0–26.0);Q

uintile

4OR:8.8(1.8–42.7)*;

Quintile

5OR:7.5(1.4–38.8)*;P trend,

0.01

Song

etal.,2016

(17)

Meta-an

alysisof13

cross-sectionaland8prospective

studiespublished

before

8March

2014exam

ining

links

betweenPC

Bsanddiabetes

risk

Pooleddiabetes

risk

inthehighestvs.lowestexposure

groupsforPC

Bs

RR:2.39(1.86–3.08)*

OCpesticides

Wuet

al.,2013

(13)

Twocase-controlstudiesofwomen

withoutdiabetes

from

theNHSandameta-an

alysisofpooleddata

withsixadditionalprospective

studies

Inciden

tdiabetes

comparinghighesttertile

ofplasm

aHCBlevelsin

NHSandhighestto

lowestexposure

groupin

pooledprospective

studies

NHSOR:3.14(1.28–7.67)*;P

ooledOR:

2.00

(1.13–3.53)*

Turyket

al.,2009

(12)

471Great

Lakessportfish

consumerswithoutdiabetes

followed

from

1994

/199

5to

2005

Inciden

tdiabetes

comparingtertilesofserumDDElevels

withthereferent

Tertile

2IRR:5.5(1.2–25.1)*;Tertile

3IRR:

7.1(1.6–31.9)*

Rignell-Hydbo

met

al.,2009

(14)

Case-controlstudyofw

omen

age50

–59yearsinsouthern

Sweden

Inciden

tdiabetes

in39

patientsandmatched

control

subjectsafter$6yearsof

follow-upcomparing

thehighestqu

artileof

DDElevelswiththereferent

OR:5.5(1.2–25)*

Leeet

al.,2010

(15)

90patientsandcontrolsubjectsinanestedcase-control

studyfollowed

for;18

years

Inciden

tdiabetes

comparingsecondsextile

withthe

referentforsummarymeasure

of16

POPs

(including

3OCpesticides)orsecondquartilewiththereferent

fortrans-no

nachlor

Sum

OR:5.4(1.6–18.4)*;Trans-no

nachlorOR:

4.3(1.5–12.6)*

Leeet

al.,2011

(16)

725participan

tsfrom

thePIVUSstudy

Inciden

tdiabetes

comparingquintilesofO

Cpesticides

or

summarymeasure

ofthreeOCpesticides

withthe

referent

Quintile5sum

OCpesticides

OR:3.4(1.0–11.7);

P trend5

0.03;Q

uintile

4trans-no

nachlorOR:

4.2(1.3–13.3)*;P trend5

0.03

Van

Larebeke

etal.,2015

(18)

973participan

tsoftheFlem

ishEnvironmen

tand

HealthSurvey

Riskofinciden

tdiabetes

calculatedforadoubling

ofserum

orcomparing90

thpercentilewith10

thpercentileoflevelsforHCB(m

enandwomen

)orDDE

(men

only)

DoubledHCBOR:1.61(1.07–2.42)*;90thvs.10th

percentileHCBOR:6.27*;D

oubled

DDEOR:1.66

(1.09–2.53)*;90thvs.10thpercentileDDEOR:

5.39*

Continuedon

p.4

care.diabetesjournals.org Ruiz and Associates 3

Tab

le1—

Continued

Reference

Population

Outcomeandcomparison

Effect

estimate(95%

CI)

Starlinget

al.,2014

(19)

13,637

women

from

theAgriculturalH

ealthStudy

Inciden

tdiabetes

foreveruse

oftheOCpesticidedieldrin

HR:1.99(1.12–3.54)*

Song

etal.,2016

(17)

Meta-an

alysisof11

cross-sectionaland6prospective

studiespublished

before

8March

2014exam

ining

links

amongvariouspesticides

anddiabetes

risk

Pooleddiabetes

risk

inthehighestvs.lowestexposure

groupsforpesticides

RR:2.30(1.81–2.93)*

Chem

icalconstituentsof

airpo

llution

Brook

etal.,2016

(20)

65adultsw

ithmetabolic

syndromeandinsulin

resistan

cefrom

theAirPo

llutionandCardiometabolic

Diseases

ChinaStudy

Chan

gein

HOMA-IRper

SDincrease

inpersonal-level

black

carbonorPM

2.5expo

sure

during

thefourth

and

fifthdays

ofassessment

Day

4black

carbon:0.18(0.01–0.36)*;D

ay5black

carbon

:0.22(0.04–0.39)*;D

ay4PM

2.5:

0.18

(0.02–0.34)*;D

ay5

PM2.5:0.22(0.08–0.36)*

Weinm

ayret

al.,2015

(21)

3,60

7individualsfrom

theHeinzNixdorfRecallStudy

inGermanyfollowed

foran

averageof5.1years

Inciden

tdiabetes

relative

IQRincrease

inPM

10,PM

2.5,

traffic-specificPM

10,and

traffic-specificPM

2.5as

wellas

comparisonof

living,100m

vs.200

mfrom

abu

syroad

PM10RR:1.20(1.01–1.42)*;P

M2.5RR:1.08(0.89–

1.29);TrafficPM

10RR:1.11(0.99–1.23);Traffic

PM2.5RR:1.10(0.99–1.23);,100m

RR:1.37

(1.04–1.81)*

Toet

al.,2015

(22)

29,549

women

from

theCan

adianNationalBreast

Screen

ingStudy

Chan

geindiabetes

prevalence

per

10mg/m

3increase

inPM

2.5expo

sure

PRR:1.28

(1.16–1.41)*

Park

etal.,2015

(23)

5,83

9subjectsin

theMulti-EthnicStudyof

AtherosclerosisCohort

Prevalen

tandinciden

tdiabetes

risk

per

IQRincrease

inresiden

tialconcentrationsofPM

2.5or

NOx

Prevalen

tDM

PM2.5OR:1.09(1.00–1.17)*;Prevalent

DM

NOxOR:1.18(1.01–1.38)*;IncidentDM

PM2.5

HR:1.02(0.95–1.10);Incident

DM

NOxHR:1.00

(0.86–1.16)

Pope

etal.,2015

(24)

669,04

6participan

tsfrom

theAmerican

Can

cerSociety

Can

cerPreven

tionStudyII

Riskofdiabetes-associated

death

ondeath

certificates

per

10mg/m

3increase

inPM

2.5

HR:1.13(1.02–1.26)*

Brook

etal.,2013

(25)

2.1millionadultsfrom

the19

91Can

adianCen

sus

MortalityFollow-upStudy

Diabetes-associated

mortalityper

10m/m

3increase

inPM

2.5

HR:1.49(1.37–1.62)*

Thiering

etal.,2013

(26)

Fastingbloodfrom

39710

-year-old

childrenin

two

prospective

German

birth

cohortstudies

Chan

gein

HOMA-IRper

2-SD

increase

inam

bientNO2

andPM

10andforevery500m

tonearestmajor

road

NO2:17.0%

(5.0–30.3%

)*;PM

10:18.7%

(2.9–36.9%

)*;

Roadproximity:7.2%

(0.8–14.0%

)*Coo

ganet

al.,2012

(27)

3,99

2black

women

livingin

LosAngelesfollowed

for

10years

Inciden

tdiabetes

per

IQRincrease

inNOxor

10mg/m

3

increase

inPM

2.5

NOxIRR:1.25(1.07–1.46)*;P

M2.5IRR:1.63

(0.78–3.44)

Coo

ganet

al.,2016

(28)

43,003

participan

tsin

theBlack

Women

’sHealthStud

yfollowed

from

1995

to2011

Inciden

tdiabetes

per

IQRincrease

inNO2by

usingbo

thland

userepression

anddispersion

mod

els

Landuse

HR:0.96(0.88–1.06);DispersionHR:0.94

(0.80–1.10)

Kram

eret

al.,2010

(29)

1,77

5women

withoutdiabetes

age54

–55followed

for16

yearsinWestGermany

Inciden

tdiabetes

per

IQRincrease

inexposure

onthe

basisofdatafrom

monitoringstations,em

ission

inventories,orlanduse

regressionmodelsas

wellas

distance

from

busy

road

(,100m)relative

toeducationstatus

MonitoredPM

10HR:1.16(0.81–1.65);Mon

itored

NO2

HR:1.34(1.02–1.76)*;InventorytrafficPM

HR:1.15

(1.04–1.27)*;Inventory

trafficNO2HR:1.15(1.04–

1.27)*;Landusesoot

HR:1.27(1.09–1.48)*;Land

useNO2HR:1.42(1.16–1.73)*;,

100m/low

educationHR:2.54(1.31–4.91)*;,

100m/high

educationHR:0.92(0.58–1.47)

Schn

eideret

al.,2008

(30)

22peo

plewithtype2diabetes

livingin

NorthCarolina

Chan

gesinvascularparam

etersper

10m/m

3increase

inPM

2.5accoun

ting

forlagperiod

(indays)

Lag0FM

D:2

17.3(2

34.6to

0.0)*;Lag1SA

EI:2

17.0

(227.5to

26.4)*;Lag3SA

EI:2

15.1

(229.3to

20.9)*

O’Don

nellet

al.,2011

(31)

9,20

2patientshospitalized

withischem

icstroke

Riskofischem

icstroke

amongpatientswithdiabetes

per

10mg/m

3increase

inPM

2.5

11%(1–22%

)*

Brook

etal.,2013

(32)

25healthyadultsfrom

ruralM

ichigan

brought

toan

urban

locationfor5consecutive

days

Chan

gein

HOMA-IRper

10mg/m

3increase

inPM

2.5

0.7(0.1–1.3)*

Continuedon

p.5

4 Environmental Inequality and Diabetes Risk Diabetes Care

OC PesticidesOC pesticides were extensively used inthe U.S. until the 1970s when mostwere banned because of their environ-mental persistence and toxicity to hu-mans and wildlife; however, several OCpesticides and their metabolites are stillmeasurable in the U.S. population. Levelsof these compounds are greater in Mex-ican Americans and African Americanscompared with whites (SupplementaryTable 1). The prolonged use of OC pesti-cides outside of the U.S. for agriculturalpurposes or vector control is believed tocontribute to higher levels in Latino pop-ulations (37,48,49). Indeed, on the basis ofNational Health and Nutrition Examina-tion Survey data, people born outsidethe U.S. are more likely to be exposedto OC pesticides (50). However, the over-representation of Mexican Americans inU.S. agriculture may also play a role inexposure disparities (51). In addition, di-rect exposures to OC pesticides beforetheir phaseout may have been passeddown to offspring through breast milkand cord blood (52), likely resulting inhigher body burdens at the start of lifethat persist into adulthood.

In concordance with animal and cellu-lar data demonstrating the capacity ofOC pesticides to disrupt multiple aspectsof cellular and systemic glucose regula-tion (Supplementary Table 2), epidemio-logical studies from various regions of theworld have associated OC pesticide expo-sure with diabetes and the metabolicsyndrome (Table 1). For example, plasmahexachlorobenzene (HCB) was positivelyassociated with incident type 2 diabetes,an effect confirmed in an accompanyingmeta-analysis (OR 2.00 [95% CI 1.13–3.53]) (13). Among Great Lakes sportfish consumers, levels of dichlorodiphe-nyldichloroethylene (DDE), a metaboliteof dichlordiphenyltrichloroethane (DDT),were associated with incident diabetes(12), whereas a study in Swedish womenshowed that being in the highest quartileof DDE levels relative to the lowest quar-tile was associated with incident diabetes(OR 5.5 [95% CI 1.2–25]) (14). In a nestedcase-control cohort of individuals fol-lowed for nearly 20 years, the OC pesti-cides trans-nonachlor, oxychlordane, andmirex were nonlinearly associated withnew-onset diabetes (15). In the PIVUSstudy, trans-nonachlor and a summary in-dex of three OC pesticides also were pos-itively associated with diabetes at age

Tab

le1—

Continued

Reference

Population

Outcomeandcomparison

Effect

estimate(95%

CI)

BPA Su

net

al.,2014

(33)

971inciden

ttype2diabetes

case-controlp

airsfrom

theNHSandNHSII

Inciden

tdiabetes

afteradjustingforBMIcomparing

highestw

iththereferentq

uartileofurinaryBPAlevels

NHSOR:0.81(0.48–1.38);NHSIIOR:2.08(1.17–3.69)

*Bietal.,2016

(34)

2,20

9middle-agedandelderlysubjectswithoutdiabetes

followed

for4years

Inciden

tdiabetes

risk

inhighestquartile

vs.lowest

quartileofurinaryBPA

levelforeach10

-pointincrease

inadiabetes

geneticrisk

score

OR:1.89(1.31–2.72)*

Huet

al.,2015

(35)

121patients

withtype2diabetes

followed

for6years

Inciden

tchronickidney

disease

inpatientswithdiabetes

comparinghighestwithreferenttertile

ofurinary

BPA

level

OR:6.65(1.47–30.04)*

Song

etal.,2016

(17)

Meta-an

alysisoffive

cross-sectionalandprospective

studiespublished

before

8March

2014exam

ining

links

betweenBPA

anddiabetes

risk

Pooleddiabetes

risk

inthehighestvs.lowestexposure

groupsforBPA

RR:1.45(1.13–1.87)*

Phthalates

Sunet

al.,2014

(33)

971inciden

ttype2diabetes

case-controlp

airsfrom

theNHSandNHSII

Inciden

tdiabetes

afteradjustingforBMIcomparing

highestwiththereferentquartileofurinary

phthalatelevels

NHSDEH

POR:1.34(0.77–2.30);NHSbu

tylphthalates

OR:0.91(0.50–1.68);NHStotalphthalatesOR:0.87

(0.49–1.53);NHSIIDEH

POR:1.91(1.04–3.49)*;

NHSIIbu

tylphthalatesOR:3.16(1.68–5.95)*;N

HSII

totalphthalatesOR:2.14(1.19–3.85)*

Watkins

etal.,2016

(36)

250childrenofw

omen

enrolledintheEarlyLifeExposure

inMexicoto

Environmen

talToxicantscohort

Chan

gein

insulin

secretionas

assessed

byaC-pep

tide

index

per

IQRincrease

ineither

inutero

MEP

levelsfor

pubertalb

oys

orperipubertalD

EHPforprepubertal

girls

Pubertalb

oys:217%(2

29to

23.3%

)*;P

repu

bertal

girls:20%(2.5–41%

)*

Song

etal.,2016

(17)

Meta-an

alysisoffourcross-sectionalandprospective

studiespublished

before

8March

2014exam

ining

links

betweenphthalates

anddiabetes

risk

Pooleddiabetes

risk

inthehighestvs.lowestexposure

groupsforBPA

RR:1.48(0.98–2.25)

Dataarefrom

stud

iesfrom

arou

ndtheworld(Sup

plem

entary

Fig.1).D

M,diabetes;FM

D,flow

-mediateddilatation

;IDR,incidence

densityratio;

PRR,prevalencerate

ratio;

SAEI,small-arteryelasticity

index.

*P,

0.05.

care.diabetesjournals.org Ruiz and Associates 5

75 years (16). In a Flemish biomonitoringprogram, OC pesticide levels measured in2004–2005 were associated with self-reported diabetes in 2011; this includedHCB as well as DDE inmen (18). Finally, inthe Agricultural Health Study, a large pro-spective cohort of pesticide applicatorsand their spouses, the OC pesticide diel-drin was associated with incident diabetes(hazard ratio [HR] 1.99 [95%CI 1.12–3.54])(19). By compiling data across studies, ameta-analysis comparing the highest tolowest exposure groups demonstrated astrong positive correlation between OCpesticide exposure and diabetes rates(RR 2.30 [95% CI 1.81–2.93]) (17).

Traffic-Related Air Pollution andParticulate MatterTraffic-related air pollution comprisesvarious chemical components, includingnitric oxides (NOx), ozone, and particulatematter (PM), which is a mixture of parti-cles and liquids classified by their diam-eter (e.g., ,10 mm [PM10] or ,2.5 mm[PM2.5]). Nationwide studies have shownthat African Americans and Latinos areexposed to significantly more PM2.5

(7,53), and ethnic and racial disparitiesin exposure to traffic-related air pollutionexceed those between income groups(54) (Supplementary Table 1). NO2 levelscorrelate closelywith PM2.5, ultrafinepar-ticles, and black carbon and thus serveas a proxy for traffic-related air pollution(55). Exposure to NO2 is 38% higher forpeople of color than for non-Hispanicwhites and 10% higher for people belowthe poverty line (54). Among nonwhiteindividuals living in poverty, childrenage ,5 years are exposed to 23% higherNO2 concentrations than the rest of thepopulation (54). Of note, racial differ-ences inNO2 exposure are greater in largemetropolitan centers compared withsmall-to-medium urban areas, likely re-flecting racial and ethnic segregationaround traffic corridors inmajor U.S. cities.Increasing evidence implicates air pol-

lution in glucose dysregulation, includinginsulin resistance (56) (Table 1). In asmall, but elegant study of residents ofrural Michigan, exposure to urban air foronly 4–5 h daily for 5 consecutive days in-creasedHOMAinsulin resistance (HOMA-IR)for each 10 mg/m3 increase in PM2.5 (32).Similarly, in adults with the metabolicsyndrome living in the Beijing metropol-itan area, variations in black carbonand PM2.5 have been associated with

worsening insulin resistance (20). In Ger-many, long-term exposure to PM10 andNO2 was associated with greater insulinresistance in 10-year-old children (26). Inaddition, several studies have linked poorair quality with progression to diabetes.In one study of individuals without dia-betes followed for 5.1 years, eachinterquartile range (IQR) increase in totalPM10 was associated with a 20% in-creased risk of developing type 2 diabetes(RR 1.20 [95% CI 1.01–1.42]) (21).Living ,100 m (relative to .200 m)from a busy road was associated with a37% increased risk of developing diabetes(RR 1.37 [95% CI 1.04–1.81]). Further-more, higher levels of PM2.5, traffic-specific PM10, and traffic-specific PM2.5

were associated with increased diabetesrisk; however, these measures failed toreach statistical significance. In a studyof black women living in Los Angeles, Cal-ifornia, followed for 10 years, incident di-abetes rates were increased for each IQRincrease in NOx (incidence rate ratio [IRR]1.25 [95% CI 1.07–1.46]), whereas PM2.5

was associated with a nonsignificant in-crease in incident diabetes (IRR 1.63[95% CI 0.78–3.44]) (27). Among womanwithout diabetes from the Study of theInfluence of Air Pollution on Lung, Inflam-mation, and Aging cohort followed for16 years, incident diabetes increased by15–42% per IQR of PM10 or traffic-relatedair pollution (29). The data from prospec-tive studies are not, however, uniform. IntheMulti-Ethnic Study of Atherosclerosis,NOx was associated with prevalent diabe-tes, and PM2.5 trended toward an associ-ation (OR 1.09 [95% CI 1.00–1.17]), butno air pollution measure was associatedwith incident diabetes over 9 years offollow-up (23). In a long-term analysis ofthe BlackWomen’s Health Study with ad-justment formultiplemetabolic stressors,NO2 was not associated with diabetes in-cidence (28). Despite this heterogeneity,epidemiological studies linking variouschemical constituents of air pollution todiabetes risk coupled with animal studiesdemonstrating that exposures to air pol-lutants such as PM2.5 and polyaromatichydrocarbons disrupt metabolism andpromote inflammation (SupplementaryTable 2) suggest that differential exposureto air pollution may augment diabetes riskin low-income communities of color.

In addition to effects on diabetes de-velopment per se, air pollutants may alsopromote adverse outcomes in those with

the disease. For example, PM2.5 levelsmodeled for home addresses were linkedto diabetes on death certificates (24),whereas a prospective analysis of .2million adults revealed that a 10 mg/m3 in-crease in PM2.5 was associated with in-creased diabetes-related mortality (25).These findings may be related to adversevascular effects in individuals with diabe-tes. In 22 patients with type 2 diabetesliving in North Carolina, daily measuresof flow-mediated vasodilatation were de-creased in association with PM2.5 levels(30). The clinical significance of this find-ing may be reflected in data showingthat each 10 mg/m3 increase in PM2.5 wasassociated with an 11% increased risk ofischemic stroke in individuals with diabe-tes (31).

BPABPA is a ubiquitous synthetic chemicalused in themanufacturing of polycarbon-ate and other plastics commonly used inconsumer products; moreover, BPA is acomponent of sales receipts and epoxyresins lining food and beverage cans aswell as water pipes. BPA exposure in theU.S. population is nearly universal (57).Although BPA is rapidly cleared from thebody and single measurements may notreflect cumulative exposure (58), AfricanAmericans and people with lower in-comes have higher BPA levels than thepopulation at large (SupplementaryTable 1). The reasons for these disparitiesare not clear, but reduced access to freshfood and consequential consumption ofprocessed foods may partly explain theseassociations (59) because consumingfoods packaged in plastic or cans in-creases BPA exposure (60). Moreover,among individuals with low food security,BPA levels are higher if they receive emer-gency food assistance, which includescanned foods (61). For example, 6–11-year-old children receiving emergencyfood assistance had BPA levels thatwere 54% higher than age-matched child-ren from more affluent families (61).

Disparities in BPA exposure may con-tribute to metabolic disease burden be-cause increasing evidence associates BPAwith diabetes. Analyses that explore theassociation between urinary BPA levelsand metabolic disease are complicatedby BPA’s rapid excretion (62); moreover,although no definitive evidence exists thaturinary excretion of BPA is influenced byrace/ethnicity, lack of adjustment for

6 Environmental Inequality and Diabetes Risk Diabetes Care

renal function can complicate urinaryassessments in population studies (63).Despite these caveats, the NationalToxicology Program concluded that BPAcould exert effects on glucose homeosta-sis and insulin release on the basis of an-imal and in vitro studies (64). Althoughsome heterogeneity exists across studies,the literature that supports this conclu-sion demonstrates myriad BPA-inducedmetabolic disruptions across multiple an-imal and cellular model systems, includ-ing alterations in body weight regulation,insulin action, and insulin secretion aswell as specific disruptions in b-cell,a-cell, hepatocyte, and adipocyte func-tion and development (SupplementaryTable 2). This conclusion is further sup-ported by limited prospective humanstudies (Table 1). In data from the NHS,extremes of BPA quartiles were associ-ated with incident diabetes after adjust-ing for BMI (OR 2.08 [95% CI 1.17–3.69])in NHS II but not NHS (33), suggesting thatagemodifies BPA-associated diabetes riskbecause the mean age in NHS II was 45.6years versus 65.6 years in NHS. Alterna-tively, these differences may have arisenfrom period-cohort effects in which theextent, diversity, or timing of exposuresmay have been greater or more deleteri-ous in NHS II. Furthermore, evidence thatthe BPA-diabetes association is modifiedby a diabetes genetic risk score (34) sug-gests that some populations are moresensitive to the diabetogenic effects ofBPA. Of note, BPA may exacerbate diabe-tes complications because high levels ofBPA have been associated with a mark-edly increased rate of developing chronickidney disease (OR 6.65 [95% CI 1.47–30.04]) (35). In onemeta-analysis that ag-gregated cross-sectional and prospectivestudies, a comparison of the highest tolowest exposure groups demonstrated apositive association between BPA and di-abetes (RR 1.45 [95% CI 1.13–1.87])(17), a finding similar to a second, morerecent meta-analysis of prevalent diabe-tes in three cross-sectional studies (OR1.47 [95% CI 1.21–1.80]) (65). Thus, onthe basis of reasonable evidence, differ-ential BPA exposure may promote diabe-tes disparities.

PhthalatesPhthalates are a diverse class of widelyused synthetic compounds. High-molecular-weight (HMW) phthalates are mainlyused as plasticizers in food packaging,

toys, and building materials, such as poly-vinyl chloride (PVC); low-molecular-weightphthalates are used in pharmaceuticals,personal care products, and solvents.Phthalates arenot covalentlyboundwithinproducts and, thus, can volatilize or leachout, thereby facilitating absorptionthrough dermal contact, ingestion, andinhalation. Higher phthalate exposureamong people of color and people withlow income have been documented invarious studies (Supplementary Table 1),although the sources of these exposuredifferences are difficult to discern giventhewidespread commercial use of phtha-lates. Reduced access to fresh fruits andvegetables and increased consumption offat-rich foods in low-income populationsmay augment exposure differences be-cause certain high-fat foods are a majorsource of HMW phthalates (66). Weather-ing of older constructionmaterials in low-income households may increase inhala-tional phthalate exposure (67). Further-more, purchasing inexpensive productslikely contributes to disproportionatephthalate exposures according to an eval-uation of products at dollar stores thatrevealed that 32% of PVC-containingproducts exceed phthalate limits estab-lished for children’s products by the Con-sumer Product Safety Commission (68).Of note, personal care products and cos-metics also contribute to phthalate expo-sure (69), especially in women, whotypically have the highest concentrationsof phthalates (70). Indeed, certain femi-nine hygiene products were found to beat least partially responsible for higherlevels of monoethyl phthalate (MEP) inAfrican American women (71). Thesedata provide provocative evidence of ra-cial, ethnic, and socioeconomic disparitiesin phthalate exposure; however, addi-tional studies are needed to further illu-minate the sources of these differences.

Several epidemiologic studies havelinked higher phthalate exposure with di-abetes (Table 1). In data fromNHS II, totalurinary phthalate metabolites were asso-ciated with diabetes (33). In this analysis,metabolites of butyl phthalates anddiethylhexyl phthalate (DEHP) were asso-ciated with diabetes (OR 3.16 [95% CI1.68–5.95] and 1.91 [95% CI 1.04–3.49],respectively). Similar to BPA, these asso-ciations may be age-related or a conse-quence of period-cohort effects becausesimilar findings were not observed withthe older, original NHS. In the Early Life

Exposure in Mexico to EnvironmentalToxicants cohort, in utero levels of MEPwere associated with reduced insulinsecretion in pubertal boys (36). In themeta-analysis of Song et al. (17), urinaryconcentrations of phthalates were nearlysignificantly associated with diabetes (RR1.48 [95% CI 0.98–2.25]).With supportivecellular and animal data demonstratingthat various phthalates have the capacityto promote dysfunction in multiple met-abolic tissues (Supplementary Table 2),further prospective studies are justifiedtodefine the relationship between phtha-late exposures and diabetes risk, particu-larly among vulnerable populations.

LINKING ENVIRONMENTALEXPOSURES TO DIABETES RISKIN VULNERABLE POPULATIONS

Most studies examining links betweenEDCs and diabetes have done so withoutconsideration of race, ethnicity, or socio-economic status; however, recent reportshave begun to interrogate these impor-tant interactions. In a cross-sectionalstudy investigating the associations be-tween phthalates and insulin resistance,an interactionwith racedemonstrated thatMexican American (P = 0.001) and non-Hispanic black adolescents (P = 0.002)had significant increments in HOMA-IRwith higher levels of HMW phtha-lates or DEHP that were not observedin non-Hispanic whites (P $ 0.74) (72).Similarly, in stratified models, HMWphthalates and DEHP were more stronglyassociated with HOMA-IR in adolescentsfrom households with lower income. Inanother cross-sectional study, phthalatelevels were positively associated withfasting blood glucose, fasting insulin, orHOMA-IR; however, the dose-responserelationship was stronger among AfricanAmericans and Mexican Americans thanamong whites (73). In the meta-analysisof Song et al. (17), the impact of PCBson diabetes risk was higher in nonwhitepopulations (RR 2.91 [95% CI 1.60–5.30])compared with their white counterparts(RR 1.94 [95% CI 1.42–2.62]); similarly,associations between OC pesticides anddiabetes were stronger in nonwhites (RR2.64 [95% CI 1.56–4.49]) than in whites(RR 1.95 [95% CI 1.40–2.71]). Althoughthese associations are likely partially at-tributable to higher EDC exposures, thesefindings also suggest that African Ameri-cans and Latinos have heightened sensi-tivity to the diabetogenic effects of

care.diabetesjournals.org Ruiz and Associates 7

Tab

le2—

Interven

tional

studiesthat

lowered

leve

lsofnonpersisten

tan

dpersisten

tdiabetogen

icEDCsin

human

s

Reference

Population

Interven

tionandassessmen

tResults

Non

persistent

pollutants

Harleyet

al.,2016

(84)

100Latinaadolescen

tsfrom

theHealthand

Environmen

talR

esearchonMakeu

pofSalinas

Adolescen

tsstudy

Meanpercentchan

ge(95%

CI)in

urinary

concentrationsafter3-day

interven

tionwith

personalcare

productsdevoidofchem

icalsunder

study

MEP:227.4%(2

39.3to

213.2)*;M

ethylparaben:2

43.9%

(261.3to

218.8)*;Propylparaben:2

45.4%(2

63.7

to217.9)*;Triclosan:2

35.7%(2

53.3to

211.6)*;

Benzoph

enon

e-3:236.0%(2

51.0to

216.4)*

Rud

elet

al.,2011

(85)

10childrenand10

adultsfrom

theSanFran

cisco

Bay

Area,California

MeanurinaryconcentrationsofBPA

andphthalates

before

andduring3-day

dietary

interven

tionwith

freshandorganicfoodsthat

werenotcanned

or

packagedin

plastic

BPA

:3.7vs.1.2ng/mL;266%*;MEH

P:7.1vs.3.4ng/m

L;253%*;MEO

HP:27

vs.12ng/m

L;255%*;MEH

HP:57

vs.

25ng/m

L;256%*

Chenet

al.,2015

(86)

30Taiwan

esegirlswithpreviouslyrecorded

high

urinaryphthalatemetaboliteconcentrations

Meanurinaryconcentrations(mg/g)

ofcreatinine

(95%

CI)of

eightph

thalatemetabolites

before

and

after1weekof

sevendifferentintervention

s:hand

washing,not

usingplasticcontainers,not

eating

food

wrapp

edinplastic,no

tmicrowavingfood

,not

taking

nutritionalsup

plem

ents,reducingtheuseof

cosm

etics,andredu

cing

theuseof

person

alcare

prod

ucts(resultsareforthosewho

werecompliant

withtheintervention

)

MMP:10

.4(3.49–29.7)vs.4.54(2.97–17.3)*;M

EP:58.6

(9.08–650)

vs.16.4(4.66–200)*;MBP:123(57.9–482)

vs.

84.7(36.3–236)*;MBzP:8.52(1.87–58.2)vs.4.67

(1.25–17.8)*;M

EHP:14.4(4.34–38.3)vs.6.95

(3.42–24.7)*;

MEO

HP:

55.2(21.2–207)

vs.26.9(15.5–85.6)*;M

EHHP:

115(40.3–398)

vs.61.2(29.1–202)*;MECPP:124

(34.7–320)

vs.52.9(30.1–161)*;SDEH

P:0.98

(0.36–3.21)

vs.0.49(0.27–1.57)*

Sathyanarayana

etal.,2013

(87)

21individualsfrom

Seattle,

Washington,w

ithhigh

potentialforBPA

andphthalateexposures

Geo

metricmeanurinaryDEH

Pconcentrations

(nmol/gcreatinine)

(95%

CI)before

andat

completionof5-day

interven

tionwithcomplete

dietary

replacemen

twithfreshandorgan

icfoods

prepared

withoutplastics

DEH

P:28

3.7

(154

.6–520.8)vs.7,027.5(4,428.1–11,152.68)*

POPs Geusauet

al.,1999

(88)

2femalepatients

withchloracne

Fecalexcretionof2,3,7,8-tetrachlorodiben

zo-

p-dioxinbefore

andaftera38-day

intervention

ofdietarysupp

lementation

witholestrachipsby

using

five

differentdo

sing

regimens(15–66

golestra

daily)

Patien

t1:13

4vs.1,350ng/day;Patient2:29

vs.240

ng/day

Jand

acek

etal.,2014

(89)

23participantsfrom

Anniston,A

labam

a,withPC

Blevelsabove

thenational50

thpercentile

Elim

inationrate

(ng/glipid/year;mean6

SEM)of

37serum

PCBsbefore

andaftera1-year

doub

le-

blindplacebo-controlledtrialof15

g/daydietary

olestravs.p

lacebo

(vegetableoil)

Olestra:20.008646

0.0116

vs.2

0.0829

60.0357/year*;

Placebo:

20.0283

60.0096

vs.2

0.0413

60.0408/year

Redgraveet

al.,2005

(90)

1obesemalepatientwithdiabetes

Adipose

levelsofthePC

Bmixturearoclor125

4before

andafter2yearsofd

ietary

supplemen

tationwith

olestra

(16g/day)

Aroclor1

254:3,20

0vs.56mg/kg;Bodyweight:10

1vs.83kg;

BMI:33

.0vs.27

.1kg/m

2;C

holesterol:8.6vs.3.7mmol/L;

Triglycerides:11.8vs.1.4mmol/L;B

lood

glucose:17

vs.

5.3mmol/L

Arguinet

al.,2010

(91)

37obesemen

undergo

ingweigh

tloss

trial

Plasmaconcentrations(mg/L)of

theOCpesticide

b-HCH(m

ean6

SD)before

andaftera3-mon

thweightloss

intervention

;sub

jectsrand

omized

tostandard

treatm

ent(n

=13),fat-redu

ceddiet

(n=

14),andolestra-substituteddiet(33%

ofdietaryfat)

(n=10)

Stan

dardtreatm

ent:0.00

96

0.019;Fat-redu

cedgrou

p:0.0156

0.035;Olestra

grou

p:20.0096

0.034*

Continuedon

p.9

8 Environmental Inequality and Diabetes Risk Diabetes Care

environmental contaminants because ofpotential synergy with other diabetes riskfactors.

The current literature provides evi-dence that five classes of environmentaltoxicants are linked to diabetes risk in hu-mans, and for each, vulnerable popula-tions are disproportionately exposed.The strength of epidemiological evidencefor these five classes varies, with themostconsistent findings observed for the per-sistent pollutants (PCBs and OCs), likelyreflecting their long biological half-lives,the stability of their quantitation, andthe longer time they have been studied.Among all five classes, however, is pro-vocative evidence of diabetes-promotingeffects as well as disparities in exposure.Thus, although further study is required,the unequal burden of environmental riskfactors in ethnically, racially, andeconom-ically segregated neighborhoods of colormay contribute to interethnic differencesin metabolic health.

ORIGINS OF DIFFERENTIALENVIRONMENTAL EXPOSURES

Addressing disparities in environmentalhealth necessitates understanding thesociological forces that shape society.Segregation profoundly influences indi-vidual socioeconomic status, reinforcesunhealthy neighborhood environments,and modifies individual behaviors (74),all of which influence metabolic diseasesusceptibility. Reduced access to afford-able healthy foods, as seen in many Afri-can American and Latino neighborhoods,promotes unhealthy eating habits (59),whereas lack of safety and reduced accessto green space can limit physical activity(75). Thus, the built environment in manycommunities of color potentiates two keydrivers of diabetes risk, namely diet andexercise.

In addition, historical economic andpolitical racialization of residential areasand the labor force has promoted today’sracial segregation and the codecline ofenvironmental health in these neighbor-hoods (76). Indeed, living in highly segre-gated metropolitan areas is associatedwith a greater health risk from industrialair pollution, with African Americans atenhanced risk relative to non-Hispanicwhites (77). Despite improvements in airquality over time, African Americans re-main exposed to significantly more airpollution than non-Hispanic whites (78).Accounts of the industrial divisionof labor

by race inmajor U.S. cities document howpeople of color were restricted to low-wage, hazardous occupations while simul-taneously being confined to low-incomehousing near these industries (76). Similarlabor divisions also occurred in agricul-ture (51). Grandfathering clauses allowolder industrial facilities, often locatedin America’s metropolitan centers, toopt out of the stricter environmental reg-ulations required of newer facilities,thereby clustering industrial toxins withinthese urban cores (77). Suburbanizationwas accompanied by expansion and clus-tering of highways near and throughneighborhoods of color (79), leading tohigher traffic-related air pollution expo-sure among African Americans and Lati-nos (54). A shifted focus to suburbaneconomic development with consequen-tial disinvestment in inner city neigh-borhoods has perpetuated a legacy ofenvironmental inequality (76). The cumu-lative effects of these cultural forcesenhance exposure to environmental tox-icants among African American, Latino,and low-income communities; addressingthis history is essential to eliminating dis-parities in metabolic health.

ENVIRONMENTAL HEALTH IN THEDIABETES CLINIC

With increasing evidence that pollutantspromotemetabolic dysfunction and likelycontribute to diabetes disparities, environ-mental health will become an importantcomponent of clinical practice. As such,physicians need to be acquainted withthesedata tomeaningfully address the con-cerns of their patients who are increasinglytroubled about these links. In addition, asthese data mature, policies to improveenvironmental quality should becomecom-ponents of comprehensive diabetes pre-vention and management strategies. Suchefforts may have significant benefits. Onthe basis of recent intriguing analyses ofthe PIVUS study, 25% reductions in repre-sentative compounds from several chemi-cal classes discussed herein (PCBs, OCpesticides, and phthalates) as well as per-fluoroalkyl substances are predicted to re-duce diabetes prevalence in Europe by 13%(95% CI 2–22%), with a projected costsavings of V4.51 billion/year (80). Thus,the identification of patient-specific expo-sures and implementation of exposurereduction strategies may reduce the bur-den of diabetes on both the individualand society at large.

Tab

le2—

Continued

Reference

Population

Interven

tionandassessmen

tResults

Guo

etal.,2016

(92)

15healthywomen

from

theSanFran

ciscoBay

Area,

California

Bloodlevelsoffi

vePC

Bsandtw

oOCpesticides

(ng/g

lipid)(m

ean6

SEM)before

andafter2mon

thsof

supp

lementation

with1,000mg/dayascorbicacid

(vitam

inC)

PCB74

:4.046

0.57

vs.4.006

0.62*;PC

B118:6.87

60.97

vs.6.796

1.01*;PC

B138:10.856

1.66

vs.10.52

61.65*;

PCB153:21.166

4.16

vs.20.68

64.01*;PC

B180:20.476

4.99

vs.20.07

64.85*;4,49-DDT:8.31

60.96

vs.8.186

1.11*;4,49-DDE:344.06

658.40vs.338.776

57.65*

b-HCH,b

-hexachlorocyclohexane;M

BP,mon

obutylph

thalate;MBzP,m

onob

enzylphthalate;M

ECPP,m

ono-2-ethyl-5-carbo

xypentylph

thalate;MEH

HP,mon

o-2-ethyl-5-hydoxyhexylph

thalate;MEH

P,mon

o-2-

ethylhexylph

thalate;MEO

HP,mon

o-2-ethyl-5-oxohexylphthalate;M

MP,mon

omethylphthalate.*P,

0.05.

care.diabetesjournals.org Ruiz and Associates 9

Identifying Patients With UniqueDiabetes Phenotypes

Astute clinicians revolutionized diabetescare by recognizing unique disease pheno-types that were subsequently linked tospecific genetic variants and targetedtherapeutics (i.e.,maturity onset diabetesof the young). Similarly, comprehensive

occupational and environmental historiesin patientswithout the classical clinical fea-tures of type 2 diabetes and without a ge-netic explanation, may identify uniquechemical exposures that promote diseasedevelopment. Similarly, patients whosemedication needs are greater than antici-pated may have background exposures

that exacerbate metabolic dysfunction.Aided by the development and implemen-tation of validated clinical questionnairesto estimate contact with diabetogenicchemicals, informed clinicians may beable to identify glucose-disrupting expo-sures and offer patients targeted interven-tions to improve diabetes outcomes.

Table 3—Sources of diabetes-promoting EDCs and exposure reduction strategies

Chemical Source Exposure Reduction Strategy

PCBs Contaminated fish, meat, and dairy products, including bottom-feeding freshwater fish that consume PCB-laden sediment

Consult local guidelines regarding which sport fish are safeto consume; Trim fat from meat and skin from fish andcook on a rack that allows fat to drain away

Dusts contaminatedwith low levels of PCBs can coat the surfacesof fruits and vegetables

Wash fruits and vegetables before consumption

Contaminated drinking water arising from PCB leaching fromtoxic waste sites or old submersible pumps containing PCBs(development of an oily film or fuel odor in water wells)

Check submersible pumps for failure and, if so, replacepumps and clean the well

Older fluorescent lights with transformers or ballasts containingPCBs

Replace old PCB-containing fluorescent bulbs

Deterioration of old building materials, including some paintsand caulking

Remove deteriorating building materials; Repair damagedareas with new, safer alternatives

OC pesticides Somehigh-fatmeats anddairy products aswell as some fattyfish Trim fat frommeat and skin from fish and cook on a rack thatallows fat to drain away

Dust and soil contaminated from historical use Regularly clean floors and remove dust with a damp cloth;Wash hands often, especially before eating or preparingfood; Wash fruits and vegetables before consumption

Air pollutant Burning of fossil fuels, including power plants, motorizedvehicles, lawn care equipment, chemical plants, factories,refineries, and gas stations

Check local airpollution forecastsandavoidoutdoorexercisewhen pollution levels are high; Avoid exercise near high-traffic areas; Use hand-powered or electric lawn careequipment; Encourage local schools andmunicipalities toreduce bus emissions by eliminating idling; Plant trees

Gas appliances, paints, solvents, tobacco smoke, and householdchemicals, including cleaning supplies

Choose electrical appliances and paints low in volatileorganic compounds; Limit use of household chemicals;Avoid places that permit smoking

Combustion of organic materials, including fireplaces, woodstoves, charcoal grills, and leaf burning

Do not burn wood, leaves, or trash

BPA Polycarbonate plastics, including some water and baby bottles,compact discs, impact-resistant safety equipment, andmedical devices

Avoidplastic containersdesignated#7onthebottom;Donotmicrowave polycarbonate plastic food containers; Opt forinfant formula bottles and toys that are labeled BPA-free; Opt for glass, porcelain, or stainless steel containerswhen possible, especially for hot foods and drinks

Epoxy resins coating metal products, such as food cans, bottletops, and water supply pipes

Eat fresh and frozen foods while reducing use of cannedfoods; Prepare more meals at home and emphasize freshingredients

Thermal paper, including sales receipts Minimize handling of receipts and thermal paper

Some dental sealants and composites Consult dentist about alternative options

Phthalates Plastic food and beverage containers Opt for glass, porcelain, or stainless steel containers whenpossible, especially for hot food and drinks

Personal care products, such as perfumes, hair sprays,deodorants, nail polishes, insect repellants, and mostconsumer products containing fragrances, includingshampoos, air fresheners, and laundry detergents

Read labels and avoid products containing phthalates;Choose products labeled phthalate-free; Avoid fragrancesand opt for cosmetics labeled no synthetic fragrance,scented only with essential oils, or phthalate-free

Contaminated food and water Purchase, if possible, organic produce, meat, and dairyproducts; Avoid food known to be especially high incontaminants; Consider using a water filter

Plastic toys; plastic coatings on wires, cables, and otherequipment; plastic showercurtains;PVC-containingproducts;carpeting and vinyl flooring; and medical devices, includingintravenous bags, tubing, and some extended-releasemedications

Choose nonplastic alternatives whenever possible,especially avoid plastics labeled #3 and #7; Avoid hand-me-down plastic toys

10 Environmental Inequality and Diabetes Risk Diabetes Care

Reducing Exposures to NonpersistentPollutantsFor patients exposed to nonpersistent dia-betogenic pollutants, practices that in-crease exposure to diabetogenic toxicantsoffer insights into potential interventions.For example, fast food intake increasesphthalate exposures (81), whereas con-sumingwater frompolycarbonate bottles(82) or soup from cans (83) increases uri-nary BPA levels. Built upon this knowl-edge, clinical trials have attempted tolower BPA and phthalate levels (Table 2).In an intervention focused on personalcare products, attention to product con-tents reduced levels of various chemicals,includingMEP (84). A trial focused on eat-ing food with limited packaging reducedlevels of DEHP and BPA (85), whereashand washing and reduction of the useof plastic cups lowered phthalate levelsin children (86). Although these studiesare encouraging, challenges in advisingpatients remain. For example, avoidingpackaged and fast foods may be imprac-tical in individuals with low food security.Moreover, even careful efforts can beconfounded by unexpected exposures asillustrated by a failed intervention duringwhich DEHP levels rose because of unex-pectedly high phthalate levels in milk andground coriander (87). Collectively, thesedata suggest that limiting contact withplastics and packaging, encouraginghand washing, and increasing awarenessof diabetogenic toxicants can reduce ex-posures; however, these efforts must besupported by regulatory action to ensureadequate labeling of consumer products.Finally, evidence that insulin sensitivityrapidly shifts with changes in air quality(20,32) suggests that advising patients toavoid exercise near busy streets or duringpeak traffic hours to limit contact with airpollutants may afford metabolic benefits,whereas community interventions to im-prove air quality (e.g., access to publictransit, reduced wood and leaf burning,expanded use of clean energy sources,tree planting) may reduce diabetes risk.

Clinical Strategies To Reduce the BodyBurden of Persistent PollutantsFor people exposed to persistent organicpollutants (POPs), evidence suggests thatinterventions can reduce diabetogenicEDC levels (Table 2). In several small stud-ies, thenonabsorbable fatolestra facilitatedelimination of lipophilic toxicants, includingthe dioxin 2,3,7,8-tetrachlorodibenzo-

p-dioxin in two patients with chloracne(88). Moreover, olestra was shown to ac-celerate the elimination of 37 noncopla-nar PCBs in 11 individuals from Anniston,Alabama (89). With regard to the meta-bolic impact of these changes, one casestudy of OC toxicity showed that 2 yearsof olestra resulted in weight loss and im-provements in glycemic control (90).Whether these metabolic improvementsresulted from the elimination of POPs orwere simply a consequence ofweight lossrequires further study. Olestra may notuniversally lower POP levels, however. Insubjects who underwent a weight lossintervention, olestra decreased levels ofb-hexachlorocyclohexane but did not at-tenuate the expected weight loss–induced rise in other OCs (91). Collec-tively, these findings raise the possibilitythat other agents that interrupt theenterohepatic circulationof lipophilic tox-icants similarly lower the body burden ofPOPs and mitigate their diabetogenic ef-fects. The glycemic benefits of the bile acidsequestrant colesevelamcouldpartially re-flect clearance of metabolism-disruptingchemicals, but this hypothesis requires for-mal testing. Other approaches to reducethe body burden of diabetogenic EDCsalso have been tried. Supported by cross-sectional data suggesting that fruit andvegetable consumption attenuates thePCB-diabetes association (93), a study of15 healthy women showed that 1,000mg/day of ascorbic acid for 2 monthsreduced levels of six PCBs and two OCpesticides (92).

Although furtherwork is needed, thesesmall intervention trials provide cliniciansand patientswith intriguing evidence thattherapeutic approaches may be devisedto mitigate exposures to diabetogenictoxicants and potentially reverse their ad-verse effects. On the basis of these dataand knowledgeof commonexposure sour-ces, physicians can aid patients wishing totake a precautionary approach by provid-ing guidance on exposure-reduction strat-egies (Table 3 and Healthcare ProviderGuide [Supplementary Fig. 2]).

CONCLUSIONS

African Americans, Latinos, and the socio-economically disadvantaged have longbeen recognized to bear a higher burdenof diabetes, but the reasons for these dis-parities are not completely understood.We provide evidence that higher expo-sure to diabetogenic pollutants is an

important contributor. Although furtherwork is required to validate the EDC-diabetes link and better quantify exposuredisparities, current evidence suggests thatimprovements in environmental healthcould reduce diabetes risk and disparities.As additional data accumulate and thefield matures, the practicing diabetologistand endocrinologist will be uniquely posi-tioned to address exposure to diabeto-genic environmental toxicants as part ofindividualized diabetes care plans to re-duce disease risk and to improve diabetesoutcomes across the population.

Acknowledgments. The authors gratefully ac-knowledge the constructive feedback of Dr. LouisH. Philipson, Director, Kovler Diabetes Center,University of Chicago. The authors also recognizethe insights and suggestions of Dr. Victoria Persky,University of Illinois at Chicago. Because of refer-ence limits, the authors regret the omission of anyother relevant studies.Funding. This work was supported by the Na-tional Institute of Child Health and Human Devel-opment (T32-HD-007009 to D.R.) and theNational Institute of Diabetes and Digestive andKidneyDiseases (P30-DK-092949 to J.S.J.) throughthe Chicago Center for Diabetes Translational Re-search as well as by the American Diabetes Asso-ciation (1-17-JDF-033 to R.M.S.).Duality of Interest. No potential conflicts of in-terest relevant to this article were reported.Author Contributions. D.R. researched andwrotethemanuscript.M.B. researchedandeditedthemanuscript. J.S.J. contributed tothediscussionandedited themanuscript. K.A. contributed to thediscussion and edited the manuscript. R.M.S.researched and wrote the manuscript. R.M.S. isthe guarantor of this work and, as such, had fullaccess to all the data in the study and takesresponsibility for the integrity of the data andthe accuracy of the data analysis.

References1. Centers for Disease Control and Prevention.National Diabetes Fact Sheet, 2011 Atlanta, Cen-ters for Disease Control and Prevention, 20112. Franco SJ. Age-adjusted percentage of adultsaged $ 20 years with diabetes, by race and His-panic ethnicitydNational Health and NutritionExamination Survey, United States, 1999-2002and 2009-2012. MMWR 2015;64(21);5993. Hunt BR, Whitman S, Henry CA. Age-adjusteddiabetes mortality rates vary in local communitiesin a metropolitan area: racial and spatial disparitiesand correlates. Diabetes Care 2014;37:1279–12864. Zoeller RT, BrownTR, Doan LL, et al. Endocrine-disrupting chemicals andpublic health protection:a statement of principles from the Endocrine So-ciety. Endocrinology 2012;153:4097–41105. Neel BA, Sargis RM. The paradox of progress:environmental disruption of metabolism and thediabetes epidemic. Diabetes 2011;60:1838–18486. Taylor D. Toxic Communities: EnvironmentalRacism, Industrial Pollution, and Residential Mo-bility. New York, New York University Press, 2014

care.diabetesjournals.org Ruiz and Associates 11

7. Bell ML, Ebisu K. Environmental inequalityin exposures to airborne particulate mattercomponents in the United States. Environ HealthPerspect 2012;120:1699–17048. Rauh VA, Landrigan PJ, Claudio L. Housing andhealth: intersectionof poverty and environmentalexposures. Ann N Y Acad Sci 2008;1136:276–2889. Murray LR. Sick and tired of being sick andtired: scientific evidence, methods, and researchimplications for racial and ethnic disparities in oc-cupational health. Am J Public Health 2003;93:221–22610. Vasiliu O, Cameron L, Gardiner J, Deguire P,Karmaus W. Polybrominated biphenyls, poly-chlorinatedbiphenyls, bodyweight, and incidenceof adult-onset diabetes mellitus. Epidemiology2006;17:352–35911. Wang SL, Tsai PC, Yang CY, Guo YL. Increasedrisk of diabetes andpolychlorinatedbiphenyls anddioxins: a 24-year follow-up study of the Yuchengcohort. Diabetes Care 2008;31:1574–157912. Turyk M, Anderson H, Knobeloch L, Imm P,Persky V. Organochlorine exposure and incidenceof diabetes in a cohort of Great Lakes sport fishconsumers. Environ Health Perspect 2009;117:1076–108213. Wu H, Bertrand KA, Choi AL, et al. Persistentorganicpollutantsandtype2diabetes:aprospectiveanal-ysis in the nurses’ health study and meta-analysis.Environ Health Perspect 2013;121:153–16114. Rignell-HydbomA, Lidfeldt J, KivirantaH, et al.Exposure to p,p9-DDE: a risk factor for type 2 di-abetes. PLoS One 2009;4:e750315. Lee DH, Steffes MW, Sjodin A, Jones RS,Needham LL, Jacobs DR Jr. Low dose of somepersistent organic pollutants predicts type 2 di-abetes: a nested case-control study. EnvironHealth Perspect 2010;118:1235–124216. LeeDH, Lind PM, JacobsDR Jr, Salihovic S, vanBavel B, Lind L. Polychlorinated biphenyls andorganochlorine pesticides in plasma predictdevelopment of type 2 diabetes in the elderly: theProspective Investigation of the Vasculaturein Uppsala Seniors (PIVUS) study. DiabetesCare 2011;34:1778–178417. Song Y, Chou EL, Baecker A, et al. Endocrine-disrupting chemicals, risk of type 2 diabetes, anddiabetes-related metabolic traits: a systematic re-viewandmeta-analysis. J Diabetes2016;8:516–53218. Van Larebeke N, Sioen I, Hond ED, et al. In-ternal exposure to organochlorine pollutants andcadmium and self-reported health status: a pro-spective study. Int J Hyg Environ Health 2015;218:232–24519. Starling AP, Umbach DM, Kamel F, Long S,Sandler DP, Hoppin JA. Pesticide use and incidentdiabetes among wives of farmers in the Agricul-tural Health Study. Occup Environ Med 2014;71:629–63520. Brook RD, Sun Z, Brook JR, et al. Extreme airpollution conditions adversely affect blood pres-sure and insulin resistance: the Air Pollution andCardiometabolic Disease Study. Hypertension2016;67:77–8521. Weinmayr G, Hennig F, Fuks K, et al.; HeinzNixdorf Recall Investigator Group. Long-term ex-posure to fine particulatematter and incidence oftype 2 diabetes mellitus in a cohort study: effectsof total and traffic-specific air pollution. EnvironHealth 2015;14:5322. To T, Zhu J, Villeneuve PJ, et al. Chronic dis-ease prevalence in women and air pollutionda

30-year longitudinal cohort study. Environ Int2015;80:26–3223. Park SK, Adar SD, O’Neill MS, et al. Long-termexposure to air pollution and type 2 diabetesmel-litus in amultiethnic cohort. Am J Epidemiol 2015;181:327–33624. Pope CA 3rd, Turner MC, Burnett RT, et al.Relationships between fine particulate air pollu-tion, cardiometabolic disorders, and cardiovascu-lar mortality. Circ Res 2015;116:108–11525. Brook RD, Cakmak S, Turner MC, et al. Long-term fine particulate matter exposure and mor-tality from diabetes in Canada. Diabetes Care2013;36:3313–332026. Thiering E, Cyrys J, Kratzsch J, et al. Long-termexposure to traffic-related air pollution and insulinresistance in children: results fromtheGINIplus andLISAplus birth cohorts. Diabetologia 2013;56:1696–170427. Coogan PF, White LF, Jerrett M, et al. Airpollution and incidence of hypertension and di-abetes mellitus in black women living in Los An-geles. Circulation 2012;125:767–77228. Coogan PF, White LF, Yu J, et al. Long termexposure to NO2 and diabetes incidence in theBlack Women’s Health Study. Environ Res 2016;148:360–36629. Kramer U, Herder C, Sugiri D, et al. Traffic-related air pollution and incident type 2 diabetes:results from the SALIA cohort study. EnvironHealth Perspect 2010;118:1273–127930. Schneider A, Neas L, Herbst MC, et al. Endo-thelial dysfunction: associations with exposure toambient fine particles in diabetic individuals. En-viron Health Perspect 2008;116:1666–167431. O’Donnell MJ, Fang J, Mittleman MA, KapralMK,Wellenius GA; Investigators of the Registry ofCanadian StrokeNetwork. Fine particulate air pol-lution (PM2.5) and the risk of acute ischemicstroke. Epidemiology 2011;22:422–43132. Brook RD, Xu X, Bard RL, et al. Reduced met-abolic insulin sensitivity following sub-acute expo-sures to low levels of ambient fine particulatematter air pollution. Sci Total Environ 2013;448:66–7133. Sun Q, Cornelis MC, Townsend MK, et al. As-sociation of urinary concentrations of bisphenol Aand phthalate metabolites with risk of type 2 di-abetes: a prospective investigation in the Nurses’Health Study (NHS) and NHSII cohorts. EnvironHealth Perspect 2014;122:616–62334. Bi Y, Wang W, Xu M, et al. Diabetes geneticrisk score modifies effect of bisphenol A exposureon deterioration in glucose metabolism. J Clin En-docrinol Metab 2016;101:143–15035. Hu J, Yang S,Wang Y, et al. Serumbisphenol Aand progression of type 2 diabetic nephropathy: a6-year prospective study. Acta Diabetol 2015;52:1135–114136. Watkins DJ, Peterson KE, Ferguson KK, et al.Relating phthalate and BPA exposure to metabo-lism in peripubescence: the role of exposure tim-ing, sex, and puberty. J Clin Endocrinol Metab2016;101:79–8837. Sjodin A, Jones RS, Caudill SP, Wong LY,TurnerWE, Calafat AM. Polybrominated diphenylethers, polychlorinated biphenyls, and persistentpesticides in serum from the National Health andNutrition Examination Survey: 2003-2008. Envi-ron Sci Technol 2014;48:753–76038. James RA, Hertz-Picciotto I, Willman E,Keller JA, Charles MJ. Determinants of serum

polychlorinated biphenyls and organochlorinepesticides measured in women from the ChildHealth and Development Study Cohort, 1963-1967. Environ Health Perspect 2002;110:617–62439. Schafer KS, Kegley SE. Persistent toxic chem-icals in the US food supply. J Epidemiol Commu-nity Health 2002;56:813–81740. Herrick RF, McClean MD, Meeker JD, BaxterLK, Weymouth GA. An unrecognized source ofPCB contamination in schools and other buildings.Environ Health Perspect 2004;112:1051–105341. Rudel RA, Seryak LM, Brody JG. PCB-containingwood floor finish is a likely source of elevated PCBs inresidents’ blood, household air and dust: a casestudy of exposure. Environ Health 2008;7:242. WeintraubM, BirnbaumLS. Catfish consump-tion as a contributor to elevated PCB levels in anon-Hispanic black subpopulation. Environ Res2008;107:412–41743. Aminov Z, Haase R, Olson JR, Pavuk M,Carpenter DO; Anniston Environmental HealthResearch Consortium. Racial differences in levelsof serum lipids and effects of exposure to persis-tent organic pollutants on lipid levels in residentsof Anniston, Alabama. Environ Int 2014;73:216–22344. Pavuk M, Olson JR, Wattigney WA, et al.; An-niston Environmental Health Research Con-sortium. Predictors of serum polychlorinatedbiphenyl concentrations in Anniston residents.Sci Total Environ 2014;496:624–63445. Choi AL, Levy JI, DockeryDW, et al. Does livingnear a Superfund site contribute to higher poly-chlorinated biphenyl (PCB) exposure? EnvironHealth Perspect 2006;114:1092–109846. Kouznetsova M, Huang X, Ma J, Lessner L,Carpenter DO. Increased rate of hospitalizationfor diabetes and residential proximity of hazard-ous waste sites. Environ Health Perspect 2007;115:75–7947. Silverstone AE, RosenbaumPF,Weinstock RS,et al. Polychlorinated biphenyl (PCB) exposureand diabetes: results from the Anniston Commu-nity Health Survey. Environ Health Perspect 2012;120:727–73248. BradmanAS, Schwartz JM, Fenster L, Barr DB,Holland NT, Eskenazi B. Factors predicting organ-ochlorine pesticide levels in pregnant Latinawomen living in a United States agriculturalarea. J Expo Sci Environ Epidemiol 2007;17:388–39949. Wang RY, Jain RB, Wolkin AF, Rubin CH,Needham LL. Serum concentrations of selectedpersistent organic pollutants in a sample of preg-nant females and changes in their concentrationsduring gestation. Environ Health Perspect 2009;117:1244–124950. Muennig P, Song X, Payne-Sturges DC, GeeGC. Blood and urine levels of long half-life toxi-cants by nativity among immigrants to the UnitedStates. Sci Total Environ 2011;412-413:109–11351. Canales AI. Inclusion and segregation: the in-corporation of Latin American immigrants intothe U.S. labor market. Lat Am Perspect 2007;34:73–8252. Smith D. Worldwide trends in DDT levels inhuman breast milk. Int J Epidemiol 1999;28:179–18853. Schweitzer L, Zhou J. Neighborhood air qual-ity, respiratory health, and vulnerable populationsin compact and sprawled regions. J Am Plann As-soc 2010;76:363–371

12 Environmental Inequality and Diabetes Risk Diabetes Care

54. Clark LP, Millet DB,Marshall JD. National pat-terns in environmental injustice and inequality:outdoor NO2 air pollution in the United States.PLoS One 2014;9:e9443155. Beckerman B, Jerrett M, Brook JR, Verma DK,ArainMA, FinkelsteinMM.Correlation of nitrogendioxide with other traffic pollutants near a majorexpressway. Atmos Environ 2008;42:275–29056. Rao X, Patel P, Puett R, Rajagopalan S. Airpollution as a risk factor for type 2 diabetes. Tox-icol Sci 2015;143:231–24157. RanjitN, Siefert K, PadmanabhanV.Bisphenol-A and disparities in birth outcomes: a review anddirections for future research. J Perinatol 2010;30:2–958. Vandenberg LN, Hunt PA, Myers JP, Vom SaalFS. Human exposures to bisphenol A: mismatchesbetween data and assumptions. Rev EnvironHealth 2013;28:37–5859. Larson NI, Story MT, Nelson MC. Neighbor-hood environments: disparities in access tohealthy foods in the U.S. Am J Prev Med 2009;36:74–8160. von Goetz N, Wormuth M, Scheringer M,Hungerbuhler K. Bisphenol A: how the most rele-vant exposure sources contribute to total con-sumer exposure. Risk Anal 2010;30:473–48761. Nelson JW, Scammell MK, Hatch EE, WebsterTF. Social disparities in exposures to bisphenol Aand polyfluoroalkyl chemicals: a cross-sectionalstudy within NHANES 2003-2006. Environ Health2012;11:1062. Lakind JS, Goodman M, Mattison DR.Bisphenol A and indicators of obesity, glucosemetabolism/type 2 diabetes and cardiovascu-lar disease: a systematic review of epidemio-logic research. Crit Rev Toxicol 2014;44:121–15063. Hays SM, Aylward LL, Blount BC. Variation inurinary flow rates according to demographic char-acteristics and body mass index in NHANES: po-tential confounding of associations betweenhealth outcomes and urinary biomarker concen-trations. Environ Health Perspect 2015;123:293–30064. Thayer KA, Heindel JJ, Bucher JR, Gallo MA.Role of environmental chemicals in diabetes andobesity: a National Toxicology Programworkshopreview. Environ Health Perspect 2012;120:779–78965. Ranciere F, Lyons JG, Loh VH, et al. BisphenolA and the risk of cardiometabolic disorders: a sys-tematic review with meta-analysis of the epide-miological evidence. Environ Health 2015;14:4666. Serrano SE, Braun J, Trasande L, Dills R,Sathyanarayana S. Phthalates and diet: a reviewof the food monitoring and epidemiology data.Environ Health 2014;13:4367. Ait Bamai Y, Araki A, Kawai T, et al. Associa-tions of phthalate concentrations in floor dust and

multi-surface dust with the interior materials inJapanese dwellings. Sci Total Environ 2014;468-469:147–15768. Campaign for Healthier Solutions. A Day Lateand a Dollar Short: Discount Retailers are FallingBehindon Safer Chemicals. Available from: http://ej4all.org/assets/media/documents/Report_ADayLateAndADollarShort.pdf. Accessed69. Parlett LE, Calafat AM, Swan SH. Women’sexposure to phthalates in relation to use of per-sonal care products. J Expo Sci Environ Epidemiol2013;23:197–20670. Zota AR, Calafat AM, Woodruff TJ. Temporaltrends in phthalate exposures: findings from theNational Health and Nutrition Examination Sur-vey, 2001-2010. Environ Health Perspect 2014;122:235–24171. Branch F, Woodruff TJ, Mitro SD, Zota AR.Vaginal douching and racial/ethnic disparitiesin phthalates exposures among reproductive-aged women: National Health and NutritionExamination Survey 2001-2004. Environ Health2015;14:5772. Trasande L, Spanier AJ, Sathyanarayana S,Attina TM, Blustein J. Urinary phthalates and in-creased insulin resistance in adolescents. Pediat-rics 2013;132:e646–e65573. Huang T, Saxena AR, Isganaitis E, James-ToddT. Gender and racial/ethnic differences in the as-sociations of urinary phthalate metabolites withmarkers of diabetes risk: National Health and Nu-trition Examination Survey 2001-2008. EnvironHealth 2014;13:674. KramerMR, Hogue CR. Is segregation bad foryour health? Epidemiol Rev 2009;31:178–19475. Gordon-Larsen P, Nelson MC, Page P, PopkinBM. Inequality in the built environment underlieskey health disparities in physical activity and obe-sity. Pediatrics 2006;117:417–42476. Massey DS, Denton NA. American Apartheid:Segregation and the Making of the Underclass.Cambridge, MA, Harvard University Press, 199877. Ard K. By all measures: an examination of therelationship between segregation and health riskfrom air pollution. Popul Environ 2016;38:1–2078. Ard K. Trends in exposure to industrial airtoxins for different racial and socioeconomicgroups: a spatial and temporal examination ofenvironmental inequality in the U.S. from 1995 to2004. Soc Sci Res 2015;53:375–39079. Chi G, Parisi D. Highway Expansion effects onurban racial redistribution in the post–Civil Rights pe-riod. Public Works Manag Policy 2011;16:40–5880. Trasande L, Lampa E, Lind L, Lind PM. Popu-lation attributable risks and costs of diabetogenicchemical exposures in the elderly. J EpidemiolCommunity Health 2017;71:111–11481. Zota AR, Phillips CA, Mitro SD. Recent fastfood consumption and bisphenol A and phthalates

exposures among the U.S. population in NHANES,2003-2010. Environ Health Perspect 2016;124:1521–152882. Makris KC, Andra SS, Jia A, et al. Associationbetween water consumption from polycarbonatecontainers and bisphenol A intake during harshenvironmental conditions in summer. EnvironSci Technol 2013;47:3333–334383. Carwile JL, Ye X, Zhou X, Calafat AM, MichelsKB. Canned soup consumption and urinary bi-sphenol A: a randomized crossover trial. JAMA2011;306:2218–222084. Harley KG, Kogut K, Madrigal DS, et al. Re-ducing phthalate, paraben, and phenol exposurefrom personal care products in adolescent girls:findings from the HERMOSA Intervention Study.Environ Health Perspect 2016;124:1600–160785. Rudel RA, Gray JM, Engel CL, et al. Food pack-aging and bisphenol A and bis(2-ethyhexyl)phthalate exposure: findings from a dietary inter-vention. Environ Health Perspect 2011;119:914–92086. Chen C-Y, Chou Y-Y, Lin S-J, Lee C-C. Develop-ing an intervention strategy to reduce phthalateexposure in Taiwanese girls. Sci Total Environ2015;517:125–13187. Sathyanarayana S, AlcedoG, Saelens BE, et al.Unexpected results in a randomized dietary trialto reduce phthalate and bisphenol A exposures.J Expo Sci Environ Epidemiol 2013;23:378–38488. Geusau A, Tschachler E, Meixner M, et al.Olestra increases faecal excretion of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Lancet 1999;354:1266–126789. Jandacek RJ, Heubi JE, Buckley DD, et al. Re-duction of the body burden of PCBs and DDE bydietary intervention in a randomized trial. J NutrBiochem 2014;25:483–48890. Redgrave TG, Wallace P, Jandacek RJ, Tso P.Treatment with a dietary fat substitute decreasedArochlor 1254 contamination in an obese diabeticmale. J Nutr Biochem 2005;16:383–38491. ArguinH, SanchezM, BrayGA, et al. Impact ofadopting a vegan diet or an olestra supplementa-tion on plasma organochlorine concentrations:results from two pilot studies. Br J Nutr 2010;103:1433–144192. Guo W, Huen K, Park JS, et al. Vitamin C in-tervention may lower the levels of persistent or-ganic pollutants in blood of healthy womendapilot study. Food Chem Toxicol 2016;92:197–20493. Hofe CR, Feng L, Zephyr D, Stromberg AJ,Hennig B, Gaetke LM. Fruit and vegetable intake,as reflected by serum carotenoid concentra-tions, predicts reduced probability of poly-chlorinated biphenyl-associated risk for type 2diabetes: National Health and Nutrition Exam-ination Survey 2003-2004. Nutr Res 2014;34:285–293

care.diabetesjournals.org Ruiz and Associates 13