Integrated Science Assessment for Lead Appendix 9: Effects on Other Organ Systems and Mortality External Review Draft March 2023


United States
Environmental Protection
Jf lkAgency

EPA/600/R-23/061
March 2023
www.epa.gov/isa

Integrated Science
Assessment for Lead

Appendix 9: Effects on Other
Organ Systems and Mortality

External Review Draft

March 2023

Health and Environmental Effects Assessment Division
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency

-------
DISCLAIMER

1	This document is an external review draft for peer review purposes only. This information is

2	distributed solely for the purpose of predissemination peer review under applicable information quality

3	guidelines. It has not been formally disseminated by the Environmental Protection Agency. It does not

4	represent and should not be construed to represent any agency determination or policy. Mention of trade

5	names or commercial products does not constitute endorsement or recommendation for use.

6

External Review Draft

9-ii

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

DOCUMENT GUIDE

This Document Guide is intended to orient readers to the organization of the Lead (Pb) Integrated
Science Assessment (ISA) in its entirety and to the sub-section of the ISA at hand (indicated in bold). The
ISA consists of the Front Matter (list of authors, contributors, reviewers, and acronyms), Executive
Summary, Integrated Synthesis, and 12 appendices, which can all be found at
https://cfpub. epa.gov/ncea/isa/recordisplay. cfm?deid=357282.

Front Matter

Executive Summary

Integrative Synthesis

Appendix 1. Lead Source to Concentration

Appendix 2. Exposure, Toxicokinetics, and Biomarkers

Appendix 3. Nervous System Effects

Appendix 4. Cardiovascular Effects

Appendix 5. Renal Effects

Appendix 6. Immune System Effects

Appendix 7. Hematological Effects

Appendix 8. Reproductive and Developmental Effects

Appendix 9. Effects on Other Organ Systems and Mortality

Appendix 10. Cancer

Appendix 11. Effects of Lead in Terrestrial and Aquatic Ecosystems
Appendix 12. Process for Developing the Pb Integrated Science Assessment

External Review Draft

9-iii

DRAFT: Do not cite or quote

-------
CONTENTS

LIST OF TABLES	9-vi

LIST OF FIGURES	9-vii

ACRONYMS AND ABBREVIATIONS	9-viii

APPENDIX 9	EFFECTS ON OTHER ORGAN SYSTEMS AND MORTALITY

	9-1

9.1	Effects on the Hepatic System	9-2

9.1.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-2

9.1.2	Scope	9-2

9.1.3	Epidemiologic Studies on the Hepatic System	9-4

9.1.4	Toxicological Studies on the Hepatic System	9-8

9.1.5	Biological Plausibility	9-9

9.1.6	Summary and Causality Determination	9-11

9.2	Metabolic Effects	9-16

9.2.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-16

9.2.2	Scope	9-16

9.2.3	Epidemiologic Studies on Metabolic Effects	9-18

9.2.4	Toxicological Studies on Metabolic Effects	9-23

9.2.5	Summary and Causality Determination	9-24

9.3	Effects on the Gastrointestinal System	9-25

9.3.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-25

9.3.2	Scope	9-25

9.3.3	Epidemiologic Studies on the Gastrointestinal System	9-27

9.3.4	Toxicological Studies on the Gastrointestinal System 	9-27

9.3.5	Summary and Causality Determination	9-28

9.4	Effects on the Endocrine System	9-28

9.4.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-28

9.4.2	Scope	9-29

9.4.3	Epidemiologic Studies on the Endocrine System	9-30

9.4.4	Toxicological Studies on the Endocrine System	9-32

9.4.5	Summary and Causality Determination	9-33

9.5	Effects on the Musculoskeletal System	9-34

9.5.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-34

9.5.2	Scope	9-36

9.5.3	Epidemiologic Studies on the Musculoskeletal System	9-37

9.5.4	Toxicological Studies on the Musculoskeletal System	9-41

9.5.5	Biological Plausibility	9-41

9.5.6	Summary and Causality Determination	9-45

9.6	Effects on Ocular Health	9-50

9.6.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-50

9.6.2	Scope	9-50

9.6.3	Epidemiologic Studies on Ocular Health	9-52

9.6.4	Toxicological Studies on Ocular Health 	9-53

9.6.5	Summary and Causality Determination	9-53

9.7	Effects on the Respiratory System	9-55

9.7.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-55

External Review Draft

9-iv

DRAFT: Do not cite or quote

-------
9.7.2	Scope	9-55

9.7.3	Epidemiologic Studies on the Respiratory System	9-57

9.7.4	Toxicological Studies on the Respiratory System	9-59

9.7.5	Summary and Causality Determination	9-60

9.8	Mortality	9-61

9.8.1	Introduction, Summary of the 2013 ISA, and Scope of the Current Review	9-61

9.8.2	Scope	9-62

9.8.3	Total (non-Accidental) Mortality	9-63

9.8.4	Cause-Specific Mortality	9-67

9.8.5	Biological Plausibility	9-68

9.8.6	Summary and Causality Determination	9-69

9.9	Evidence Inventories - Data Tables to Summarize Study Details	9-73

9.10	References	9-147

1

External Review Draft

9-v

DRAFT: Do not cite or quote

-------
LIST OF TABLES

Table 9-1

Table 9-2

Table 9-3
Table 9-4
Table 9-5
Table 9-6
Table 9-7
Table 9-8
Table 9-9
Table 9-10
Table 9-11
Table 9-12
Table 9-13
Table 9-14
Table 9-15
Table 9-16
Table 9-17

Evidence that is suggestive of, but not sufficient to infer, a causal relationship between Pb
exposure and hepatic effects.	9-14

Summary of evidence for a likely to be causal relationship between Pb exposure and
musculoskeletal effects.

Animal toxicological studies of exposure to Pb and gastrointestinal effects.
Epidemiologic studies of exposure to Pb and endocrine effects.	

Animal toxicological studies of exposure to Pb and endocrine effects._
Epidemiologic studies of exposure to Pb and musculoskeletal effects.

Animal toxicological studies of exposure to Pb and musculoskeletal effects.
Epidemiologic studies of exposure to Pb and ocular effects.	

Animal toxicological studies of Pb exposure and ocular effects.
Epidemiologic studies of Pb exposure and respiratory effects._

Animal toxicological studies of exposure to Pb and respiratory effects.
Epidemiologic studies of Pb exposure and total mortality.	

9-48

Summary of evidence for a causal relationship between Pb exposure and total mortality. 	9-72

Epidemiologic studies of exposure to Pb and hepatic effects.	9-73

Animal toxicological studies of exposure to Pb and hepatic effects.	9-81

Epidemiologic studies of exposure to Pb and metabolic effects. 	9-85

Animal toxicological studies of exposure to Pb and metabolic effects.	9-99

9-101
9-102
9-111
9-113
9-124
9-125
9-131
9-132
9-137
9-139

External Review Draft

9-vi

DRAFT: Do not cite or quote

-------
LIST	OF FIGURES

Figure 9-1	Potential biological pathways for hepatic effects following exposure to Pb.	9-11

Figure 9-2	Potential biological pathways for musculoskeletal effects following exposure to Pb.	9-44

Figure 9-3	Effect estimates for associations of blood Pb with all-cause mortality.	9-64

Figure 9-4	Dose-response relationship between blood Pb levels and all-cause mortality. 	9-65

External Review Draft

9-vii

DRAFT: Do not cite or quote

-------
ACRONYMS AND ABBREVIATIONS

AAS	atomic absorption spectrometry

AD	Alzheimer's disease

ALAD	S-aminolevulenic acid dehydratase

ALP	alkaline phosphatase

ALT	alanine aminotransferase

AMD	age-related macular degeneration

AOPP	advanced oxidation protein products

AQCD	Air Quality Criteria Document

ARCA	Automobile Racing Clube of America

AST	aspartate aminotransferase

AV/TV	adipocyte volume/total volume

BLL	blood lead (Pb) level

BMD	bone mineral density

BMI	body mass index

BMP	bone morphogenic protein

BV/TV	bone volume to total volume

C2C	serum cleavage neoepitope of type II
collagen

Ca2+	calcium ions

CAT	catalase

C-R	concentration-response

CAR	Cortisol awakening response

Cd	cadmium

CHEER	Children's Health and Environment
Research

CHF	congestive heart failure

CI	confidence interval

CK18	cytokeratin 18

COMP	cartilage oligomeric matrix protein

CPU	carboxypropeptide of type II collagen

CRP	C-reactive protein

CVD	cardiovascular disease

CYP	Cytochrome P450

d	day(s)

DBP	diastolic blood pressure

DMFT	decayed, missing, and filled teeth

DXA	Dual-energy X-ray absorptiometry

ECRHS	European Community Respiratory
Health Survey

EGF	epidermal growth factor

eGFR	estimated glomerular filtration rate

ELEMENT	Early Life Exposures in Mexico to
Environmental Toxicants

ER	endoplasmic reticulum

ERSD	end-stage renal disease

F#	filial generation

FBG	fasting blood glucose

FEV1	forced expiratory volume in one second

FIB-4	fibrosis-4

FT3	free triiodothyronine

FT4	free thyroxine

FVC	forced vital capacity

GADA	glutamic acid decarboxylase antibodies

GD	gestational day

GDM	gestational

GFAAS	graphite furnace atomic absorption
spectrometry

GFR	glomerular filtration rate

GGT	gamma-glutamyl transferase

GH	growth hormone

GI	gastrointestinal

GM	geometric mean

GPx	glutathione peroxidase

GSH	glutathione

GSH-PX	glutathione peroxidase

Hb	hemoglobin

HDL	high-density lipoprotein

HDL-C	high-density lipoprotein cholesterol

HF	hepatic fibrosis

HOMA- p	HOMA of P-cell function

HOMA-IR	Homeostatic Model Assessment for
Insulin Resistance

HR	hazard ratio

HS	hepatic steatosis

ICP-MS	inductively coupled plasma mass
spectrometry

IHC	immunohistochemistry

IHD	ischemic heart disease

i.p.	intraperitoneal

IOP	intraocular pressure

ISA	Integrated Science Assessment

KARE	Korean Association Resource

KNHANES	Korean National Health and Nutrition

Examination Survey

K-XRF	K-Shell X-Ray Fluorescence

LDL	low-density lipoprotein

LDL-C	low-density lipoprotein cholesterol

LOD	limit of detection

mo	month(s)

MDA	malondialdehyde

MetS	metabolic syndrome

External Review Draft

9-viii

DRAFT: Do not cite or quote

-------
METS	Modeling the Epidemiologic Transition

Study

MI	myocardial infarction

microCT	micro-computed tomography

mRNA	messenger ribonucleic acid

NAAQS	National Ambient Air Quality

Standards

NAFLD	nonalcoholic fatty liver disease

NANC	noncholinergic

NAS	Normative Aging Study

NASCAR	National Association for Stock Car

Auto Racing

NHANES	National Health and Nutrition

Examination Survey

NF -kB	nuclear factor kappa B

NP	nanoparticle

OA	osteoarthritis

OLD	obstructive lung disease

OLF	obstructive lung function

OR	odds ratio

Pb	lead

PbO	lead oxide

PCNA	proliferating cell nuclear antigen

PCR	polymerase chain reaction

PD	Potential difference

PECOS	Population, Exposure, Comparison,

Outcome, and Study

PIR	poverty-income-ratio

PM	particulate matter

PND	postnatal day

PROGRESS Programming Research in Obesity,

Growth Environment and Social Stress

PTE!	parathyroid hormone

PTElrP	parathyroid hormone-related protein

qRT-PCR	real-time quantitative reverse

transcription-polymerase chain reaction

RBC	red blood cell

RR	risk ratio

RT-PCR	reverse transcription-polymerase chain
reaction

SBP	systolic blood pressue

SBEE1C	Shiwha and Banwol Environmental

Elealth Cohort

SD	standard deviation

SE	standard error

SES	socioeconomic status

SNP	single nucleotide polymorphisms

SOD	superoxide dismutase

SPECT	single photon emission computed
tomography

SSBI	sugar sweetened beverage intake

T-SOD	total superoxide dismutase

T	tertile

TB	total bilirubin

TBARS	thiobarbituric acid reactive substance

TC	total cholesterol

TEM	transmission electron microscopy

Tg	thyroglobulin

TGAb	thyroglobulin antibodies

TGF-pi	transforming growth factor-beta 1

TNF	tumor necrosis factor

TSE1	thyroid stimulating hormone

TPOAb	thyroid peroxidade antibody

Q	quartile

wk	week(s)

yr	year(s)

External Review Draft

9-2

DRAFT: Do not cite or quote

-------
APPENDIX 9 EFFECTS ON OTHER ORGAN

SYSTEMS AND MORTALITY

Summary of Causality Determinations for Pb Exposure and Effects on Other Organ Systems

This appendix characterizes the scientific evidence that supports causality determinations for
lead (Pb) exposure and hepatic effects, metabolic effects, gastrointestinal effects, endocrine system
effects, effects on bone and teeth, effects on ocular health, and respiratory effects. The types of studies
evaluated within this appendix are consistent with the overall scope of the ISA as detailed in the
Process Appendix (see Section 12.4). In assessing the overall evidence, strengths and limitations of
individual studies were evaluated based on scientific considerations detailed in the Table 12-5 of the
Process Appendix (Section 12.6.1). More details on the causal framework used to reach these
conclusions are included in the Preamble to the ISA (U.S. EPA, 2015). The evidence presented
throughout this appendix supports the following causality conclusions:

Outcome Group

Causality Determination

Hepatic Effects

Suggestive

Metabolic Effects

Inadequate

Gastrointestinal Effects

Inadequate

Endocrine System Effects

Inadequate

Musculoskeletal Effects

Likely to be Causal

Effects on Ocular Health

Inadequate

Respiratory Effects

Inadequate

Mortality

Likely to be Causal

The Executive Summary, Integrated Synthesis, and all other appendices of this Pb ISA can be found
at https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=357282.

External Review Draft

9-1

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

9.1

Effects on the Hepatic System

9.1.1	Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

The 2013 Lead Integrated Science Assessment (2013 Pb ISA) concluded that "because of the
insufficient quality of studies, the available evidence was inadequate to determine if there is a causal
relationship between Pb exposure and hepatic effects" (U.S. EPA. 2013). Epidemiologic evidence from a
limited number of occupational studies demonstrated impaired liver function in Pb-exposed workers.
However, the internal validity and generalizability of these studies was limited by cross-sectional study
designs, lack of consideration for potential confounders, and notably higher blood Pb levels (BLLs)
(>29 (ig/dL) than the general population. Similarly, toxicological studies observed changes in liver
function enzymes and other markers of liver health in animals exposed to Pb, but the use of bolus
injections as a common route of exposure and high BLLs (>30 (ig/dL) introduced uncertainty regarding
their relevance to human exposures.

9.1.2	Scope

The scope of this section is defined by Population, Exposure, Comparison, Outcome, and Study
Design (PECOS) statements. The PECOS statement defines the objectives of the review and establishes
study inclusion criteria, thereby facilitating identification of the most relevant literature to inform the Pb
ISA.1 In order to identify the most relevant literature, the body of evidence from the 2013 Pb ISA was
considered in the development of the PECOS statements for this Appendix. Specifically, well-established
areas of research; gaps in the literature; and inherent uncertainties in specific populations, exposure
metrics, comparison groups, and study designs identified in the 2013 Pb ISA inform the scope of this
Appendix. The 2013 Pb ISA used different inclusion criteria than the current ISA, and the studies
referenced therein often do not meet the current PECOS criteria (e.g., due to higher or unreported
biomarker levels). Studies included in the 2013 Pb ISA, including many that do not meet the current
PECOS criteria, are discussed in this appendix to establish the state of the evidence prior to this
assessment. Except for supporting evidence used to demonstrate the biological plausibility of Pb-

1 The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

External Review Draft

9-2

DRAFT: Do not cite or quote

-------
1

associated effects on the hepatic system, recent studies were only included if they satisfied all the
components of the following discipline-specific PECOS statements:

Epidemiologic Studies:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Exposure: Exposure to Pb1 as indicated by biological measurements of Pb in the body - with a
specific focus on Pb in blood, bone, and teeth; validated environmental indicators of Pb
exposure;2 or intervention groups in randomized trials and quasi-experimental studies.

Comparison: Populations, population subgroups, or individuals with relatively higher versus
lower levels of the exposure metric (e.g., per unit or log unit increase in the exposure metric,
or categorical comparisons between different exposure metric quantiles).

Outcome: Effects on the hepatic system.

Study Design: Epidemiologic studies consisting of longitudinal and retrospective cohort studies,
case-control studies, cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest, randomized trials and quasi-experimental studies examining
interventions to reduce exposures.

Experimental Studies:

Population: Laboratory nonhuman mammalian animal species (e.g., mouse, rat, guinea pig,
minipig, rabbit, cat, dog) of any lifestage (including preconception, in utero, lactation,
peripubertal, and adult stages).

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.3'4

1 Recent studies of occupational exposure to Pb were considered insofar as they addressed a topic area of particular
relevance to the National Ambient Air Quality Standards (NAAQS) review (e.g., longitudinal studies designed to
examine recent versus historical Pb exposure).

2 Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM25) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure (Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with blood Pb levels (BLLs) are lacking.

3 Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

4 This level represents an order of magnitude above the upper end of the distribution of U.S. young children's BLL.
The 95th percentile of the 2011-2016 National Health and Nutrition Examination Survey (NHANES) distribution of
BLL in children (1-5 years; n = 2,321) is 2.66 (ig/dL (Eganetal.. 20211 and the proportion of individuals with BLL
that exceed this concentration varies depending on factors including (but not limited to) housing age, geographic
region, and a child's age, sex, and nutritional status.

External Review Draft

9-3

DRAFT: Do not cite or quote

-------
1

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

Outcomes: Effects on the hepatic system.

Study design: Controlled exposure studies of animals in vivo.

9.1.3 Epidemiologic Studies on the Hepatic System

Epidemiologic evidence evaluated in the 2013 Pb ISA (U.S. EPA. 2013) was limited to a small
number of occupational studies that demonstrated impaired liver function in Pb-exposed workers.
However, the internal validity and generalizability of these studies was limited by cross-sectional study
designs, lack of consideration for potential confounders, and notably higher BLLs (>29 (ig/dL) than the
general population. Recent epidemiologic studies of the hepatic system generally examine one of three
groups of endpoints: (1) direct evaluation of liver injury (e.g., nonalcoholic fatty liver disease [NAFLD]
and hepatic fibrosis); (2) serum biomarkers of liver function (e.g., alanine aminotransferase [ALT],
aspartate aminotransferase [AST], alkaline phosphatase [ALP], and gamma-glutamyl transferase [GGT]);
and (3) serum lipids (e.g., fatty acids, lipids, and cholesterol). Results from recent studies provide
inconsistent evidence of an association between BLLs and direct or indirect measures of liver damage.
Recent studies evaluating hepatic effects are generally limited to cross-sectional analyses, which are
unable to establish temporality between exposure and outcome. Additionally, with BLL, it is difficult to
characterize the specific timing, duration, frequency, and level of Pb exposure that contributed to
associations observed with liver function. This uncertainty may apply particularly to assessments of
BLLs, which in nonoccupationally-exposed adults, reflect both current exposures and cumulative Pb
stores in bone that are mobilized during bone remodeling. Measures of central tendency for Pb biomarker
levels used in each study, along with other study-specific details, including study population
characteristics and select effect estimates, are highlighted in Table 9-4. An overview of the recent
evidence is provided below.

9.1.3.1 Direct Evaluation of Liver Injury

A limited number of recent cross-sectional studies examined the association between BLLs and
liver injury, including NAFLD and fibrosis (Chung et al.. 2020; Reiaetal.. 2020; Werder et al.. 2020;
Zhai et al.. 2017). These studies, which use a variety of diagnostic tools, provide inconsistent evidence of
an association between BLLs and NAFLD and fibrosis. Liver biopsy is the gold standard for evaluating
NAFLD and liver fibrosis, but it is an invasive and cost prohibitive procedure. Therefore, epidemiologic
studies often rely on alternative measurement techniques, including imaging, biomarkers, and biomarker-
based prediction models. Imaging - either ultrasonic or magnetic resonance - generally has greater
sensitivity and specificity than reliance on biomarkers.

External Review Draft

9-4

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

A recent cross-sectional study of adults in the Yangtze River Delta in China examined the
relationship between BLLs and NAFLD measured by ultrasound (Zhai et al.. 2017). In addition to using
ultrasonic imaging, this study included a large number of participants (n = 2,011). In sex-stratified
models, Zhai et al. (2017) reported increases in the odds of NAFLD associated with increasing BLL
quartiles after adjusting for a range of demographic and hepatic and metabolic health factors. The
observed associations were stronger in magnitude among men (odds ratio [OR] = 2.168 [95% CI: 0.989,
4.751] quartile 4 versus quartile 1) compared with women (OR= 1.613 [95% CI: 1.082, 2.405] quartile 4
versus quartile 1); however, the effect estimates in men were much less precise due to a smaller sample of
men in the study population. Given the imprecise estimates for men (i.e., wide 95% CIs), it is difficult to
draw conclusions on sex-specific comparisons.

Results from other recent cross-sectional studies are inconsistent. In a small exploratory analysis
of oil spill response workers with low BLLs (mean = 1.82 (.ig/dL). Werder et al. (2020) evaluated the
association between BLLs and cytokeratin 18 (CK18), a serologic biomarker of hepatocyte death that has
been used as a marker for NAFLD. The authors observed an association between BLLs and caspase-
cleaved fragment CK18 (CK18 M30), but not whole protein CK18 (CK18 M65). Notably, CK18 M65 has
performed better as a measure of NAFLD than CK18 M30 (Lee et al.. 2020). adding further ambiguity to
the observed results. Additionally, Werder et al. (2020) examined a range of heavy metals and markers of
inflammation and did not adjust for multiple testing, which increases the likelihood of chance findings
and may explain the inconsistent results. In addition to this weak evidence of an association between
BLLs and markers of NAFLD, (Chung et al.. 2020) analyzed data from the Korean National Health and
Nutrition Examination Survey (KNHANES) and reported null or negative sex-specific associations
between BLLs and scores on the Hepatic Steatosis Index, a validated biomarker-based prediction model
of NAFLD. The authors also observed negative associations between BLLs and Fibrosis 4 Index, a
similarly validated model for fibrosis. This larger analysis (n = 4,420) reported similar mean BLLs
(1.81 (ig/dL) as those reported in Werder et al. (2020).

In addition to studies examining NAFLD and fibrosis separately, (Reia et al.. 2020) used a
biomarker-based index to estimate fibrosis level in National Health and Nutrition Examination Survey
(NHANES) participants with NAFLD. In this case, fibrosis level was used as an indicator of NAFLD
severity. Reia et al. (2020) reported large, but imprecise associations between BLL quartiles and
advanced liver fibrosis. For example, the authors noted that participants in the highest quartile of BLLs
(>1.62 (ig/dL) had a 493% increase in the odds of advanced liver fibrosis (95% CI: 188%, 1,124%)
compared to participants in the lowest quartile (<0.64 (ig/dL). Despite having a large sample size, the
authors only examined severe liver fibrosis, which likely resulted in a small number of cases (total cases
not reported), which would have decreased the statistical power of the study. Limited statistical power
resulting from a small sample size simultaneously reduces the likelihood of detecting a true effect and the
likelihood that an observed result reflects a true effect.

External Review Draft

9-5

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

9.1.3.2

Serum Biomarkers of Liver Function

Serum biomarkers can be used as indirect evidence of liver damage. For example, elevated levels
of ALT or AST can indicate the presence of necrosis in the liver, and elevated levels of bilirubin, ALP, or
GGT can be associated with cholestasis. However, changes in serum biomarker levels are also related to
effects on other biological systems. Elevated GGT can also occur with chronic heart failure, and elevated
ALP can be used to detect bone disorders. Therefore, studies evaluating these biomarkers in combination
are more likely to provide evidence of abnormal liver function relative to studies evaluating a single
biomarker.

There have been a limited number of recent epidemiologic studies that evaluated serum
biomarkers of liver function, including a longitudinal study (Pollack ct al.. 2015) and a few cross-
sectional analyses (Chen et al.. 2019; Obcng-Gvasi. 2019; Christensen et al.. 2013). Recent studies, which
adjust for a wide range of potential confounders, provide some evidence of an association between BLLs
and serum biomarkers, but results are not entirely consistent, and the implications of some associations
are unclear. Specifically, a small prospective cohort study of premenopausal women evaluated the percent
change in AST, ALT, ALP, and bilirubin over the course of an 8-week follow-up (Pollack et al.. 2015).
The authors reported imprecise increases in AST (5.02% [95% CI: -1.36%, 11.41%]), ALT (6.39% [95%
CI: 3.07%, 9.72%]), and ALP (2.14% [95% CI: -5.05%, 9.33%]) per 1 (ig/dL increase in BLLs measured
at baseline (mean = 1.03 (ig/dL), but no change in bilirubin (-0.20% [95% CI: -7.50%, 7.11%]). The
clinical relevance of these findings is uncertain given the majority of the study population fell well within
the normal ranges of each of the biomarkers. A recent cross-sectional study of adults living near an e-
waste facility in China better addresses clinical relevance by examining the association between BLLs
and abnormal liver function, defined as having two or more transaminases (AST, ALT, GGT) elevated
above the normal range, or having one transaminase at least twice as high as the upper bound of the
normal range (Chen et al.. 2019). In this study, which had notably higher median BLLs (5.1 to 8.7 (ig/dL
across study locations), a 1 (ig/dL increase in BLL was associated with a large, but imprecise increase in
the odds of abnormal liver function (OR = 1.94 [95% CI: 1.00, 3.73]).

Results from recent large cross-sectional NHANES analyses examining a single serum biomarker
of liver function were inconsistent. In an analysis of 2003-2004 NHANES participant's ages 12 years and
older, Christensen et al. (2013) reported null associations between increasing BLL quartiles and ALT
levels. An analysis restricted to adult participants of more recent NHANES survey cycles (2011-2016)
observed an increase in the odds of GGT levels above the study population median (18 U/L) associated
with a 1 (ig/dL increase in BLLs (OR = 1.94 [95% CI: 1.652, 2.28] for young adults and 1.34 [95% CI:
1.14, 1.58] for middle-aged adults) (Obcng-Gvasi. 2019). Similar to the Pollack et al. (2015) study, the
median GGT levels in this study were within the normal range, making it difficult to interpret the clinical
relevance of the results.

External Review Draft

9-6

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

9.1.3.3

Serum Lipids

Many fatty acids, lipids, and cholesterol are synthesized and eliminated in the liver; the
relationships among them and their relevance to other aspects of human health, including metabolic
effects (Section 9.2) and cardiovascular effects (Appendix 4). are complex. Although increases or
decreases in serum or liver cholesterol levels may be associated with liver damage, it can be challenging
to determine whether the changes are a consequence of said damage or a contributing factor in disease
progression (Arguello et al.. 2015; Chrostek ct al.. 2014). Recent epidemiologic studies of serum lipids
have been conducted in populations of adults and children and include a mix of prospective cohorts and
cross-sectional designs. Recent studies also account for a range of potential confounders, including
demographics and socioeconomic status (SES) factors, medical history, and medication use. Associations
between BLLs and serum lipids have been largely inconsistent across both lifestages.

In a recent study including a subset of the Veterans Affairs Normative Aging Study (NAS) cohort
with healthy older adults, Peters et al. (2012) examined the associations between BLLs at baseline and
serum lipid levels after three to four years of follow-up. The authors reported increased odds of clinically
elevated total cholesterol associated with an increase in BLLs (OR = 1.08 [95%: 0.99, 1.19] per 1 (ig/dL
increase in BLL). Associations with clinical cut points for other serum lipids were either null (elevated
triglycerides and low-density lipoprotein [LDL] cholesterol) or negative (low high-density lipoprotein
[HDL] cholesterol). Cross-sectional studies of adult populations, including analyses of nationally
representative health survey data (Xu et al.. 2021; Lee and Kim. 2016) and a small analysis of adults of
African descent (Ettinger et al.. 2014). are also inconsistent. Results across these studies (see Table 9-2)
provide no discernable pattern of associations between BLLs and triglycerides, LDL cholesterol, or HDL
cholesterol. BLL measures of central tendency were low across the evaluated studies (<5 (ig/dL) and do
not appear to explain the inconsistencies.

Results from studies in children are similarly inconsistent. Two recent studies of serum lipids
analyzed data from separate birth cohorts in Mexico - the Early Life Exposures in Mexico to
Environmental Toxicants (ELEMENT) study (Liu et al.. 2020) and the Programming Research in
Obesity, Growth Environment and Social Stress (PROGRESS) birth study (Kupsco et al.. 2019). In
children ages 4 to 6, Kupsco et al. (2019) reported null associations between prenatal BLLs and serum
triglycerides and non-HDL cholesterol. In contrast, in an analysis including older children and teens, Liu
et al. (2020) observed an increase in triglyceride Z-scores in children with prenatal BLLs >5 (ig/dL
compared to those with BLLs less than 5 (ig/dL (0.58 [95% CI: -0.05, 1.20]). The authors observed
negative associations between prenatal BLLs and cholesterol Z-scores (total, LDL, and HDL). A large
cross-sectional analysis of NHANES participants ages 12 to 19 noted a 2.3% (95% CI: 0.3%, 4.2%)
increase in LDL cholesterol and a 0.6% (95% CI: -0.1%, 1.3%) increase in total cholesterol per 1 (ig/dL
increase in BLL (Xu et al.. 2017). The authors observed null (total cholesterol and HDL cholesterol) or
negative (triglycerides) associations between BLLs and other serum lipids.

External Review Draft

9-7

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

9.1.4

Toxicological Studies on the Hepatic System

As described in the 2013 Pb ISA, evidence from toxicological studies indicates exposure to Pb
can result in altered liver function and hepatic oxidative stress (U.S. EPA. 2013). A few studies reported
Pb-induced decreases in cytochrome P450 (CYP) enzymes (Phase I xenobiotic metabolism), as well as
Pb-induced decreases in serum protein and albumin levels and increased AST, ALT, ALP, and GGT
activities (indicators of decreased liver function), increased oxidative stress, and decreased antioxidant
status. A number of recent studies have corroborated findings of Pb exposure and decreased liver function
(Barkaoui et al.. 2020; Dumkova et al.. 2020b; Gao et al.. 2020; Andielkovic et al.. 2019; Laamech et al..
2017; Long et al.. 2016; Liu et al.. 2013; Berrahal et al.. 2011). While impaired lipid metabolism was
reported in the 2013 Pb ISA, results from recent studies of cholesterol have been inconsistent. Laamech et
al. (2017) found an increase in total cholesterol in mice given Pb acetate in their drinking water (BLL:
18 (ig/dL). Conversely, Dumkova et al. (2020a) found lower levels of total cholesterol in rats that were
given Pb oxide nanoparticles by inhalation (BLLs: 3.1-8.5 (.ig/dL); however, the latter group did report an
increase in lipid droplets by liver histology [BLLs: 3.1-17.8 (ig/dL; (Dumkova et al.. 2020a; Dumkova et
al.. 2020b; Dumkova et al.. 2017)1. Observation of Pb-associated increases in hepatic oxidative stress, as
indicated by a decrease in glutathione (GSH) levels and catalase (CAT), superoxide dismutase (SOD),
and glutathione peroxidase (GPx) activities has been found in additional recent studies of oral Pb
exposure [drinking water: 21.4-29.0 (ig/dL (Barkaoui et al.. 2020; Andielkovic et al.. 2019; Long et al..
2016); oral gavage: 18.5-30.2 (ig/dL (Gao et al.. 2020; Laamech et al.. 2017; Li et al.. 2017)1.

Since the 2013 Pb ISA, several recent studies have reported perturbations related to oxidative
stress in addition to the endpoints noted above. For example, Andielkovic et al. (2019) found changes in
multiple parameters of oxidative stress in liver and kidney tissue in male rats, indicative of an oxidative
stress response to Pb exposure (BLL: 29.0 (ig/dL). Long et al. (2016) also reported several markers of
oxidative damage and response, in mouse liver tissue. They showed in addition, consistent with an
oxidative damage response, attenuation of such response after administration of proanthocyanidins, which
are naturally occurring antioxidant compounds. The same authors reported changes in several markers
that are consistent with a generalized endoplasmic reticulum (ER) response in the liver to environmental
stressors. Likewise, Liu et al. (2013) showed Pb responsiveness of ER stress markers, and the antagonistic
effect of quercetin (a natural flavonoid) on this response. Barkaoui et al. (2020) reported finding
alleviation of Pb-induced oxidative effects from administration of antioxidative, phenolic compounds
extracted from Plantigo albicans.

Cell death by apoptosis may be a downstream result of the molecular sequelae of Pb exposure
described in the preceding paragraph. Indeed, such a result has been reported in mouse livers, both
phenotypically and via molecular markers (Dumkova et al.. 2017; Long et al.. 2016).

External Review Draft

9-8

DRAFT: Do not cite or quote

-------
1

9.1.5

Biological Plausibility

This section describes biological pathways that potentially underlie effects of Pb on the liver and
hepatic function. Figure 9-1 depicts the proposed pathways as a continuum of upstream events, connected
by arrows, which may lead to downstream events observed in epidemiologic studies. This discussion of
how exposure to Pb may lead to hepatic effects contributes to an understanding of the biological
plausibility of epidemiologic results evaluated above. Note that the structure of the biological plausibility
sections and the role of biological plausibility in contributing to the weight-of-evidence analysis used in
the current Pb ISA are discussed in Section IS.4.2.

The hepatic effects of Pb exposure have been studied in many experimental models. The pathway
proposed, outlined in Figure 9-1, involves the induction of oxidative stress and inflammation leading to
downstream cellular loss and metabolic changes that could plausibly be responsible for the development
of health effects in the liver. Oxidative stress control and inflammation are highly regulated processes and
are tightly linked. As discussed above and in both the 2013 Pb ISA and 2006 Pb Air Quality Criteria
Document (AQCD), inflammatory signaling and marker of oxidative stress have been found in the livers
of animals exposed to Pb (see Section 9.1.3 and (U.S. EPA. 2013. 2006). Hepatic inflammation and
oxidative stress co-occur thus it is difficult to determine if one process precedes the other, thus, they are
grouped in the same grey box in Figure 9-1.

Regulation of inflammation and oxidative stress involve widespread gene expression changes that
could plausibly alter the expression of metabolizing enzymes and proteins necessary for cholesterol
synthesis and maintaining lipid homeostasis which could lead to fat accumulation and subsequent fatty
liver disease. As discussed in the 2013 Pb ISA, Pb treatment can cause elevated cholesterol levels through
changes in cholesterol synthesis pathways in the liver. Pb can also alter the expression and activity of
CYP enzymes that are important in the response to xenobiotics as well as metabolism of cholesterol-
derived steroid hormones. A recent study in knockout mice showed that mice deficient in the II-1
inflammatory mediators were protected from the hypercholesterolemia in response to Pb compared to
wild type mice (Koiima et al.. 2012). Knockout mice also did not experience the messenger ribonucleic
acid (mRNA) upregulation cholesterol synthesizing enzymes HMGR and Cyp51 or the downregulation of
bile acid synthesizing enzyme Cyp7al. These data support the necessity of inflammation to the regulation
of cholesterol metabolism and are the basis for the solid line from inflammation to the box containing
CYP activity and altered cholesterol synthesis in Figure 9-1.

Excessive damage from oxidative stress and inflammatory responses could lead to cell death
which, in excess, could lead to changes in hepatocyte structure and ultimately decrease liver function. As
discussed above and in the 2013 Pb ISA and 2006 Pb AQCD, many animal studies have shown that Pb
exposure of varying durations and developmental stages results in liver injury, which is most commonly
measured as increased activity of liver enzymes (e.g., AST, ALT, ALP) in the blood serum or plasma.
Increases of liver enzyme activity have been seen in the serum of humans occupationally exposed to Pb

External Review Draft

9-9

DRAFT: Do not cite or quote

-------
1

(Mazumdar and Goswami. 2014; U.S. EPA. 2013). As mentioned above, elevated liver enzymes in the
blood can serve as an indirect markers of liver damage. Previous research has shown that exposure to Pb
in animal models can lead to upregulation of cell death pathways (U.S. EPA. 2013) and more recent
studies provide additional support (Almasmoum et al.. 2019; Abu-Khudir et al.. 2017; Hasanein et al..
2016; Long et al.. 2016; Mabrouk et al.. 2016; Liu et al.. 2013; Pal et al.. 2013; Liu et al.. 2012. 201IV
Studies have shown that treatment with antioxidants, like vitamin E (Almasmoum et al.. 2019). vitamin C
(Upadhvav et al.. 2009). or therapeutic compounds that have anti-inflammatory and antioxidant properties
(Abu-Khudir et al.. 2017; Hasanein et al.. 2016; Long et al.. 2016; Mabrouk et al.. 2016; Liu et al.. 2013;
Pal et al.. 2013; Liu et al.. 2012) can prevent the Pb-induced upregulation of apoptotic pathways and
concomitantly reduced both markers of oxidative damage and serum markers of liver injury. Interestingly,
some therapeutic compounds reduce the liver Pb burden suggesting that the reduction in oxidative stress
may be caused by toxicokinetic changes that reduce the liver Pb exposure concentration (Liu et al.. 2013;
Liu etal.. 2011). however, some studies have seen that antioxidant treatment can reduce oxidative stress
even while live Pb levels remain elevated suggesting that oxidative stress is directly related to
downstream liver damage (Almasmoum et al.. 2019; Long et al.. 2016; Mabrouk et al.. 2016; Reckziegel
et al.. 2016). Together these data provide support for the solid line from the box containing inflammation
and oxidative stress to cell death.

Excessive cell loss can result in changes to liver architecture and trigger repair processes that can
lead to liver scarring, both of which can lead to loss of liver function. The 2013 Pb ISA discussed studies
that showed that Pb treatment led to noticeable histologic changes including signs of increased fibrotic
liver changes (U.S. EPA. 2013). More recent work supports this with evidence that liver histologic
changes are accompanied by increased markers of apoptosis and necrosis (Long et al.. 2016; Mabrouk et
al.. 2016). A study also showed that 4 months of Pb exposure in rats increased wound repair signaling
pathways which corresponded to increased deposition of extracellular matrix proteins in the liver (Perez
Aguilar et al.. 2014). Sufficient damage to the liver can reduce liver function which can be measured as a
reduced level of protein in the blood. Recent studies have shown decreases in serum proteins following
Pb exposure that coincide with molecular or histological signs of liver damage (Almasmoum et al.. 2019;
El-Tantawv. 2016; Hasanein et al.. 2016). Similar evidence is seen in the 2013 Pb ISA. Together, it is
plausible that widespread cell death in the liver can lead to changes in hepatocyte structure that leads to
liver damage and resulting decline in liver function.

The proposed pathway leading from Pb exposure to hepatic health effects begins with the
induction of inflammation and increase in oxidative stress. This results in both changes in metabolizing
enzymes and cholesterol synthesis that could be responsible for fatty accumulation in the liver.
Widespread oxidative damage results in cell loss which could disrupt the normal liver structure and
contribute to loss of liver function. Together, the evidence supports a plausible pathway from Pb exposure
to the hepatic effects seen in epidemiologic and animal tox studies.

External Review Draft

9-10

DRAFT: Do not cite or quote

-------
1

Altered cholesterol synthesis

¦HUH

Altered CYP activity

Fatty liver disease

CYP = cytochrome P450.

Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to Pb exposure, and
the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway used in an experimental study involving Pb exposure. Dotted arrows denote a possible
relationship between effects. Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect
individual boxes, groupings of boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted
from left to right and color coded (white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population
level or a key clinical effect). Here, population-level effects generally reflect results of epidemiologic studies. When there are gaps in
the evidence, there are complementary gaps in the figure and the accompanying text below. The structure of the biological
plausibility sections and the role of biological plausibility in contributing to the weight-of-evidence analysis used in the 2022 Pb ISA
are discussed in IS.7.2.

Figure 9-1 Potential biological pathways for hepatic effects following
exposure to Pb.

9.1.6 Summary and Causality Determination

The 2013 Pb ISA (U.S. EPA. 2013) concluded that the available evidence was "inadequate to
determine if there is a causal relationship between Pb exposure and hepatic effects." A limited number of
occupational epidemiologic studies evaluated potential associations between increased BLLs and
decreases in serum protein and albumin levels and increased liver function enzymes, oxidative stress, and
antioxidant status. The implications of the occupational epidemiologic evidence were limited because of
the cross-sectional design of the studies, the high BLLs examined (means >29 ug/dL). and the lack of
consideration for potential confounding by factors such as age, diet, BMI, smoking, or other occupational
exposures. Similar changes in liver function enzymes were found in mature animals exposed to high
levels of Pb during adulthood, and animals exposed during gestation and lactation. Pb exposure was also
shown to impair lipid metabolism in animals, as evidenced by increased hepatic cholesterogenesis, and
altered triglyceride and phospholipid levels (Shanna et al.. 2010; Ademuviwa et al.. 2009; Khotimchenko
and Kolenchenko. 2007). Multiple toxicological studies observed Pb-associated increases in hepatic

External Review Draft

9-11

DRAFT: Do not cite or quote

-------
1

oxidative stress, generally indicated by an increase in lipid peroxidation along with a decrease in GSH
levels and CAT, SOD, and GPx activities (Pandva et al.. 2010; Sharma et al.. 2010; Yu et al.. 2008;
Adegbesan and Adenuga. 2007; Jurczuk et al.. 2007; Khotimchenko and Kolenchenko. 2007; Jurczuk et
al.. 2006). However, the relevance of the toxicological evidence was uncertain, as many studies
administered Pb as bolus doses. Additionally, few toxicological studies reported the resulting BLLs and
those studies that did provide this evidence had BLLs of limited relevance to environmentally exposed
humans (>30 (ig/dL). Thus, despite some evidence of Pb-induced hepatic effects, uncertainties related to
the relevance of the available studies limited the causal conclusions that could be drawn in the 2013 Pb
ISA.

Recent toxicological studies include more relevant routes of exposure (i.e., drinking water, oral
gavage, and inhalation) and exposures resulting in lower BLLs than those available for the previous ISA
(BLL range: 3.6-30.2 (.ig/dL). These studies provide consistent evidence of Pb-induced increases in AST,
ALT, ALP, and GGT activities, which are indicative of reduced liver function (Barkaoui et al.. 2020;
Dumkova et al.. 2020b; Gao et al.. 2020; Andielkovic et al.. 2019; Laamech et al.. 2017; Long et al..
2016; Liu et al.. 2013; Berrahal et al.. 2011). Additionally, recent studies provide consistent evidence of
Pb-associated increases in hepatic oxidative stress, as indicated by decreases in GSH levels and CAT,
SOD, and GPx activities (Barkaoui et al.. 2020; Gao et al.. 2020; Andielkovic et al.. 2019; Laamech et al..
2017; Li et al.. 2017; Long et al.. 2016). While impaired lipid metabolism was reported in the 2013 Pb
ISA, a limited number of recent studies of cholesterol have reported contrasting results, one indicating
Pb-induced increases in total cholesterol (Laamech et al.. 2017) and the other reporting decrements in
total cholesterol (Dumkova et al.. 2020a).

In contrast to toxicological evidence, recent epidemiologic studies evaluating the relationship
between BLLs and hepatic effects are generally inconsistent or inconclusive. Similar to studies evaluated
in the 2013 Pb ISA, most recent studies implement cross-sectional designs, although they include more
robust adjustment for potential confounders and populations with much lower mean BLLs. Still, these
studies do not address potentially large differences in past versus current exposures. There is therefore
uncertainty as to the specific timing, duration, frequency, and level of Pb exposure that contributed to any
observed associations. The strongest evidence for direct liver injury comes from a large cross-sectional
analysis of adults in China that reported a positive association between BLLs and NAFLD prevalence
measured by ultrasound (Zhai et al.. 2017). Other cross-sectional analyses used biomarkers or biomarker
indices to assess NAFLD, which are less accurate than ultrasonic imaging and may introduce non-
differential misclassification. Non-differential misclassification of a dichotomous outcome is likely to
bias results toward the null. The available biomarker studies of NAFLD did not provide convincing
evidence that BLLs are associated with NAFLD prevalence (Chung et al.. 2020; Reia et al.. 2020; Werder
et al.. 2020). Results from studies that examined serum biomarkers of general liver function (e.g., AST,
ALT, ALP, GGT, and bilirubin) provided some evidence that BLLs are associated with increased
biomarker levels (Chen et al.. 2019; Obeng-Gvasi. 2019; Pollack et al.. 2015). but the inferences that can
be drawn from two of these studies is limited due to study populations that had biomarkers well within

External Review Draft

9-12

DRAFT: Do not cite or quote

-------
1

normal ranges (Chen etal.. 2019; Obeng-Gvasi. 2019). There are also a few recent studies that examined
serum lipids in adults or children and the results are inconsistent. Across studies, contrasting associations
were observed between BLLs and specific lipids, with no discernable pattern of associations between
BLLs and triglycerides, LDL cholesterol, HDL cholesterol, or total cholesterol.

Overall, recent toxicological studies build upon evidence from the 2013 Pb ISA and provide
largely consistent evidence that indicates exposure to Pb can result in altered liver function and hepatic
oxidative stress. Compared to the 2013 Pb ISA, recent toxicological studies include routes of exposure
and BLLs that are more relevant to humans. Results from a limited number of recent epidemiologic
studies examining liver enzymes are generally coherent with the toxicological evidence, indicating Pb-
associated increases in enzymes that are consistent with altered liver function. However, due to the
reported liver enzyme levels in the epidemiologic studies, there is uncertainty as to whether the observed
changes in enzymes are indicative of liver injury. Additionally, epidemiologic studies of direct liver
injury provide inconsistent evidence of an association with BLLs. Thus, based on the strength of the
toxicological evidence and some remaining inconsistencies and uncertainties in the epidemiologic
evidence, the collective evidence is suggestive of, but not sufficient to infer, a causal relationship
between Pb exposure and hepatic effects. The key evidence, as it relates to the causal framework, is
summarized in Table 9-1.

External Review Draft

9-13

DRAFT: Do not cite or quote

-------
Table 9-1 Evidence that is suggestive of, but not sufficient to infer, a causal relationship between Pb exposure
and hepatic effects.

Rationale for Causality
Determination3

Key Evidence13

References'3

Pb Biomarker Levels Associated with Effects0

Consistent evidence from
animal toxicological
studies at relevant BLLs

Toxicological studies provide largely
consistent evidence that indicates exposure
to Pb can result in:

Altered liver function

Increases in hepatic oxidative stress, as
indicated by decreases in GSH levels and
CAT, SOD, and GPx activities

Berrahal etal. (2011)
Liu etal. (2013)

Long etal. (2016)
Andielkovic et al. (2019)
Gao et al. (2020)
Dumkova et al. (2020b)
Laamech et al. (2017)
Barkaoui et al. (2020)

Li etal. (2017)

Long etal. (2016)
Andielkovic et al. (2019)
Barkaoui et al. (2020)
Gao et al. (2020)
Laamech et al. (2017)

Range of mean BLLs across studies:
18.0 to 29.0 [jg/dL

Range of mean BLLs across studies:
3.6 to 30.2 [jg/dL

Limited or inconsistent
evidence from
epidemiologic studies at
relevant BLLs

Inconsistent evidence of associations
between BLLs and NAFLD

Some evidence that BLLs are associated
with increased levels of serum biomarkers
of liver function, but limited inference due to
study populations that had biomarkers well
within normal ranges

See Section 9.1.3.1

Pollack et al. (2015)
Chen etal. (2019)
Obeng-Gvasi (2019)

Range of mean BLLs across studies:
1.0 to 5.29 [jg/dL

Range of mean BLLs across studies:
1.0 to 8.7 [jg/dL

External Review Draft

9-14

DRAFT: Do not cite or quote

-------
Rationale for Causality
Determination3

Key Evidence13

References'3

Pb Biomarker Levels Associated with Effects0

Biological Plausibility

The proposed pathway leading from Pb
exposure to hepatic health effects begins
with the induction of inflammation and
increase in oxidative stress. This results in
both changes in metabolizing enzymes and
cholesterol synthesis that could be
responsible for fatty accumulation in the
liver. Widespread oxidative damage results
in cell loss which could disrupt the normal
liver structure and contribute to loss of liver
function.

See Section 9.1.4

BLLs = blood lead levels; CAT = catalase; GSH = glutathione; GPx = glutathione peroxidase; NAFLD = nonalcoholic fatty liver disease; Pb = lead; SOD = superoxide dismutase.
aBased on aspects considered in judgments of causality and weight-of-evidence in causal framework in Table I and Table II of the Preamble to the ISAs (U.S. EPA. 2015).
bDescribes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and, where applicable, to uncertainties or
inconsistencies. References to earlier sections indicate where the full body of evidence is described.

°Describes the Pb biomarker levels at which the evidence is substantiated.

1

External Review Draft

9-15

DRAFT: Do not cite or quote

-------
1

9.2

Metabolic Effects

9.2.1 Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

The 2013 Pb ISA (U.S. EPA. 2013) did not have a separate discussion of potential metabolic
effects of exposure to Pb. However, evidence relevant to metabolic effects was provided by a small
number of studies that examined glucose and insulin homeostasis, lipids, cholesterol, and liver health
endpoints. These studies provided evidence for modes of action and were discussed across a few sections
of the 2013 Pb ISA (U.S. EPA. 2013). including Section 4.4 (Cardiovascular Effects), Section 4.5 (Renal
Effects), and Section 4.9.1 (Effects on the Hepatic System). There was no causality determination for
metabolic effects in the 2013 Pb ISA (U.S. EPA. 2013).

The metabolic effects reviewed in this section include diabetes mellitus and insulin resistance
(Section 9.2.3.1), metabolic syndrome and its components (Section 9.2.3.2), and effects on body weight
measures (Section 9.2.3.3). Other metabolic indicators, such as changes in liver function, serum lipids,
and neuroendocrine signaling, are discussed in other sections of this appendix (Sections 9.2 and 9.4).

9.2.2 Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria
than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological

1 The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

External Review Draft

9-16

DRAFT: Do not cite or quote

-------
1

plausibility of Pb-associated metabolic effects, recent studies were only included if they satisfied all of the
components of the following discipline-specific PECOS statements:

Epidemiologic Studies:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Outcome: Metabolic effects.

Experimental Studies:

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.3'4

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

1 Recent studies of occupational exposure to Pb were considered insofar as they addressed a topic area of particular
relevance to the NAAQS review (e.g., longitudinal studies designed to examine recent versus historical Pb
exposure).

2 Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM2 5) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

3 Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

4 This level represents an order of magnitude above the upper end of the distribution of U.S. young children's BLL.
The 95th percentile of the 2011-2016 NHANES distribution of BLL in children (1-5 years; n= 2,321) is 2.66 (ig/dL
(Eganetal.. 20211 and the proportion of individuals with BLL that exceed this concentration varies depending on
factors including (but not limited to) housing age, geographic region, and a child's age, sex, and nutritional status.

External Review Draft

9-17

DRAFT: Do not cite or quote

-------
1

Outcomes: Metabolic effects.

Study design: Controlled exposure studies of animals in vivo.

9.2.3 Epidemiologic Studies on Metabolic Effects

9.2.3.1 Diabetes Mellitus and Insulin Resistance

Diabetes mellitus is a chronic condition characterized by an inability to regulate glucose in the
blood by producing or responding to insulin. A number of epidemiologic studies evaluated in the 2013 Pb
ISA (U.S. EPA. 2013) examined diabetes as a potential at-risk factor that could modify the relationship
between Pb exposure and other health outcomes, but none examined the direct relationship between Pb
exposure and diabetes incidence or prevalence. Recent studies have examined this relationship,
commonly categorizing diabetes mellitus status as meeting one or more of the following criteria: (1)
elevated fasting blood glucose (FBG), (2) self-reported use of insulin or oral medications for diabetes, or
(3) self-reported physician diagnosis with diabetes. There are three primary types of diabetes: type I, type
II, and gestational (GDM). Some of the evaluated studies distinguished between types of diabetes
mellitus, while others did not. Most studies were cross-sectional in design, meaning temporality between
exposure and outcome could not be established.

Recent epidemiologic studies examining the relationship between Pb exposure and diabetes
mellitus, or insulin resistance have reported mostly null findings across lifestages. In adult populations, a
limited number of case-control and cross-sectional studies examining diabetes prevalence reported null or
inverse associations between BLLs and diabetes mellitus or levels of insulin resistance. Results from
recent studies examining insulin resistance in adolescents and gestational diabetes in pregnant women are
also mostly null. Measures of central tendency for Pb biomarker levels used in each study, along with
other study-specific details, including study population characteristics and select effect estimates, are
highlighted in Table 9-6. An overview of the recent evidence is provided below.

Studies in Adults

In a recent cross-sectional analysis of blood Pb and diabetes using data from the 2009 and 2010
cycles of the KNHANES, Moon (2013) observed a negative trend in diabetes prevalence across blood Pb
quartiles. Compared to the lowest blood Pb quartile (geometric mean (GM): 1.43 |ig/dL). the largest
reductions in the odds of diabetes were observed in the highest exposure quartile (GM: 4.08 (ig/dL;
OR = 0.745 [95% CI: 0.516, 1.077]) and in the second highest quartile (GM: 2.74 (ig/dL) (OR = 0.759
[95% CI: 0.531, 1.086]). Similarly, in sex-stratified analyses of subjects without diabetes, Moon (2013)
reported slight reductions in the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR),

External Review Draft

9-18

DRAFT: Do not cite or quote

-------
1

HOMA of (3-cell function (HOMA-J3), and fasting insulin per log unit increase in blood Pb. The observed
results were comparable in men and in women.

Two recent cross-sectional case-control studies originating from the Nord-Trondelag Health
Study (HUNT3) evaluated differences in blood Pb measurements between subjects with and without type
II diabetes and reported results that are also consistent with a null or negative association (Hansen et al..
2017; Simic et al.. 2017). Specifically, Hansen et al. (2017) identified 128 cases of previously
undiagnosed, screening-detected type II diabetes and 755 age- and sex-matched controls. The authors
observed a slight, but notably imprecise increase in odds of screening-detected type II diabetes for blood
Pb quartile 4 compared to quartile 1 (OR= 1.12 [95% CI: 0.58, 2.16]). As indicated by the wide
confidence intervals, the increase in odds is difficult to distinguish from chance. In a parallel analysis,
Simic et al. (2017) identified 267 cases of self-reported type II diabetes and 609 frequency-matched
controls from the same HUNT3 cohort. Consistent with results from Moon (2013). (Simic et al.. 2017)
observed a substantial reduction in diabetes prevalence for BLLs in the highest quartile compared to the
lowest (OR = 0.24 [95% CI: 0.13, 0.47]). The observation of a negative association for Pb and type II
diabetes by Simic et al. (2017) but not Hansen et al. (2017) may be related to differences in exposure
contrast between identified cases and controls. Hansen et al. (2017) reported median BLLs of 1.99 (ig/dL
for controls and 1.94 (ig/dL for cases, while Simic et al. (2017) reported median BLLs of 2.02 (ig/dL for
controls and 1.64 (ig/dL for cases. Additionally, the differences could be due to an effect of diabetes
treatment on BLLs, which highlights an uncertainty of these cross-sectional analyses.

Studies in Adolescents

A recent study assessed the relationship between exposure to Pb in utero and insulin resistance in
adolescence (Liu et al.. 2020). Pregnant mothers were enrolled in the ELEMENT project from 1997-1999
and 2001-2003 and their children were followed until 2015. There was a null association between first
trimester maternal blood Pb >5 (ig/dL and HOMA-IR in adolescence. In combined and sex-stratified
analyses, associations were null.

Studies in Pregnant Women

A number of recent studies have investigated the relationship between Pb exposure and GDM.
These studies, most of which have reported null associations between BLLs and GDM, are discussed in
more detail in Section 8.4.1.1.2 of the Reproductive and Developmental Effects Appendix.

9.2.3.2 Metabolic Syndrome and its Components

Metabolic syndrome (MetS) describes a set of cardiometabolic conditions that increase a person's
risk for cardiovascular diseases. Components of MetS include elevated blood pressure, low HDL

External Review Draft

9-19

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

cholesterol, elevated blood tryglicerides, elevated FBG, and a high waist circumference, also referred to
as abdominal obesity. A MetS diagnosis is commonly defined as meeting three or more of the following
criteria: (1) elevated blood pressure (systolic blood pressure >130 mmHg or diastolic blood pressure
>85 mmHg or current use of blood pressure medication); (2) low HDL cholesterol (<40 mg/dL in women
or <50 mg/dL in men); (3) elevated serum triglycerides (>150 mmHg) or current use of anti-dyslipidemia
medication; (4) elevated FBG (>100 (.ig/dL): (5) abdominal obesity (waist circumference >90 cm in men
or >85 cm in women). None of the studies evaluated in the 2013 Pb ISA (U.S. EPA. 2013) examined the
relationship between Pb exposure and MetS. Recent evidence for the effects of Pb exposure on MetS and
its components is inconsistent. Measures of central tendency for Pb biomarker levels used in each study,
along with other study-specific details, including study population characteristics and select effect
estimates, are highlighted in Table 9-3. An overview of the recent evidence is provided below.

9.2.3.3 Metabolic Syndrome

A number of recent large, population-based cross-sectional studies have analyzed the relationship
between BLLs and MetS prevalence and provide inconsistent evidence of an association. Across studies,
mean and/or median BLLs were below 5 (ig/dL, including some below 2 (ig/dL. Studies analyzing data
from overlapping cycles of the KNHANES observed increased MetS prevalence in participants with
higher BLLs (Moon. 2014; Rhee et al.. 2013). Specifically, Rhee et al. (2013) reported that 2008
KNHANES participants with BLLS in the highest exposure quartile (3.07-19.43 |ig/L) were 2.57 (95%
CI: 1.46, 4.51) times more likely to have MetS than subjects in the lowest quartile (0.42-1.73 |ig/L). The
authors noted a consistent concentration-response trend across quartiles. In an analysis incorporating
more KNHANES cycles (2007-2012), Moon (2014) observed a smaller increase in the odds of MetS for
subjects in the second highest exposure quartile (GM 2.51 (ig/dL) (OR = 1.21 [95% CI: 0.90, 1.62])
compared to the lowest (GM 1.23 (ig/dL) but did not observe a clear dose-response trend across quartiles.

In contrast to KNHANES studies, other analyses of data from a variety of large population-based
surveys noted negative associations between BLLs and MetS (Wen et al.. 2020; Bulka et al.. 2019; Shim
et al.. 2019). Bulka et al. (2019) used data from two cycles (2011-2014) of the NHANES to perform a
cross-sectional analysis of blood Pb and MetS prevalence. The authors observed reduced odds of MetS
with increasing blood Pb quartile, with the largest reduction observed in subjects in the highest quartile of
lead exposure (1.64-15.98 (ig/dL) compared to the lowest quartile (0.18-0.70 (ig/dL) (OR= 0.81 [95%
CI: 0.64, 1.03]). Shim et al. (2019) and Wen et al. (2020) similarly reported reduced odds of MetS
associated with increased BLLs in the Korean National Environmental Health Survey II (KNHANES II)
and a survey of adults in Taiwan, respectively.

External Review Draft

9-20

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

Components of Metabolic Syndrome

In addition to cross-sectional studies evaluating MetS prevalence, several recent studies have
assessed the potential effects of Pb on the individual components of MetS (abdominal obesity [often
measured by waist circumference], low HDL cholesterol, elevated triglycerides, and elevated FBG;
studies evaluating blood pressure and hypertension are discussed in Section 4.3). Similar to studies that
evaluated MetS prevalence, most of these studies analyzed cross-sectional data from nationally
representative health surveys. In general, results from recent studies were inconsistent across individual
MetS components, with the exception of blood pressure and serum triglycerides.

Waist Circumference

Recent KNHANES analyses of BLLs and waist circumference were inconsistent (Lee and Kim.
2016. 2013; Rhee et al.. 2013). In an analysis of KNHANES participants from 2005-2010, Lee and Kim
(2013) observed no apparent association between BLLs and waist circumference. The same authors
evaluated more recent KNHANES cycles (2007-2012) and observed slightly increased odds of waist
circumference >85 cm in the second ( >2.199-3.011 jxg/d) and third (>3.011 (ig/dL) blood Pb tertiles
compared to the first tertile (<2.199 (.ig/dL). but slightly decreased odds per twofold continuous increase
in blood Pb (Lee and Kim. 2016). In contrast, in an analysis of 2008 KNHANES participants, Rhee et al.
(2013) found a modest but positive association between blood Pb and abdominal circumference as a
continuous variable.

Results from two recent NHANES analyses were similarly inconsistent (Bulka et al.. 2019; Wang
et al.. 2018c). Wang et al. (2018c) used data from NHANES cycles between 2003 and 2014 and observed
a 0.8% (95% CI: 0.6, 1.0%) reduction in waist circumference per 1-SD increase in logio-transformed
blood Pb (fig/dL). While the large sample size of this analysis leads to precise 95% CIs, the relevance of a
notably small decrement in waist circumference is unclear. In contrast, a study including two NHANES
cycles that overlapped with the Wang et al. (2018c) study (2011-2014) reported negative associations
between BLLs and probability of abdominal obesity (Bulka etal.. 2019).

HDL Cholesterol and Serum Triglycerides

The previously discussed KNHANES analyses also assessed HDL cholesterol and serum
triglycerides. These studies do not provide evidence that BLLs are associated with increased odds of low
HDL cholesterol (Lee and Kim. 2016. 2013; Rhee et al.. 2013). The same studies did provide consistent
evidence of higher serum triglycerides in association with higher BLLs, although these studies were
notably conducted in overlapping populations (i.e. non-independent samples). Lee and Kim (2013) and
Lee and Kim (2016) observed slight increases in odds of high serum triglycerides (>150 (ig/dL) with
higher BLLs (analyzed as a continuous variable and as tertiles). Similarly, Rhee et al. (2013) reported a
modest positive association between serum triglycerides and log-transformed BLLs.

External Review Draft

9-21

DRAFT: Do not cite or quote

-------
1

In addition to studies that examined HDL cholesterol and serum triglycerides in conjunction with
MetS, a few other recent studies also evaluated these measures as part of a broader lipids profile. As
discussed in Section 9.1.3.3, these studies were inconsistent for HDL cholesterol and triglycerides,
including a prospective cohort study of older Veterans participating in the NAS that reported null
associations between BLLs at baseline and HDL cholesterol and triglyceride levels after three to four
years of follow-up (Peters et al.. 2012).

Elevated Fasting Glucose

The majority of recent population-based cross-sectional studies of MetS components did not
observe associations between BLLs and FBG. Specifically, KNHANES analyses (Lee and Kim. 2016;
Rhee et al.. 2013) and a recent NHANES analysis (Bulka et al.. 2019) reported null associations between
BLLs and FBG. In contrast, in an analysis of earlier KNHANES cycles, Lee and Kim (2013) reported
blood Pb to be positively associated with elevated FBG (>100 (.ig/dL). with the odds of elevated FBG
increasing with each doubling of BLLs (OR= 1.118 [95% CI: 0.953, 1.311]). In addition to large cross-
sectional studies, a smaller cross-sectional analysis of adults of African descent across five countries of
varying social and economic development in Africa also examined the relationship between BLLs and
elevated FBG (Ettinger et al.. 2014). Ettinger et al. (2014) reported a large increase in the odds of elevated
FBG (>100 mg/dL) in subjects with a blood Pb exposure level above the median (1.66 (ig/dL) compared
to those below it (OR = 4.99 [95% CI: 1.97, 12.69]). However, the small sample size (n = 150) in this
study reduces statistical power, as well as the likelihood that an observed result reflects a true effect.

9.2.3.4 Body Weight Measures in Adults

A few epidemiologic studies evaluated in the 2013 Pb ISA (U.S. EPA. 2013) examined obesity as
a potential risk factor that could modify the relationship between Pb exposure and other health outcomes,
but none examined the direct relationship between Pb exposure and body weight measures in adults.
Recent studies have examined this relationship, commonly assessing body weight using body mass index
(BMI), a measure of body fat that is calculated as a person's weight divided by the square of their height.
For adults, overweight is defined as having a BMI of 25 kg/m2 or greater and obesity is defined as having
a BMI of 30 kg/m2 or greater. Studies examining Pb and body weight measures in children and
adolescents are discussed in the Reproductive and Developmental Effects Appendix of this ISA (Section
8.5.1.1).

A limited number of recent studies have examined the relationship between Pb exposure and
body weight measures in adults. Overall, the current evidence for the effects of Pb exposure on body
weight measures is inconsistent, although small sample sizes limit the interpretation of a few of the
studies. Additionally, recent studies are cross-sectional, which reduces confidence in their results because
temporality between exposure and outcome cannot be established. Measures of central tendency for Pb

External Review Draft

9-22

DRAFT: Do not cite or quote

-------
1

biomarker levels used in each study, along with other study-specific details, including study population
characteristics and select effect estimates, are highlighted in Table 9-3. An overview of the recent
evidence is provided below.

Recent studies examining Pb exposure and body weight measures in adults utilize cross-sectional
study designs. In an analysis of a large population-based survey of Chinese citizens, Wang et al. (2018a)
observed small but precise increases in BMI ((3 = 0.24 kg/m2 [95% CI: 0.08, 0.40 kg/m2]) and odds of
being overweight or obese (OR =1.13 [95% CI: 1.02, 1.25]) per natural log unit increase in blood Pb
(|ig/L). In order to account for potential reverse causality, the authors used Mendelian randomization to
assess the relationship between BLLs and genetic variants associated with increased BMI. Because the
genetic variants precede exposure, the variants are expected to be associated with BLLs if BMI is a
potential causal factor of increased BLLs. Wang et al. (2018a) reported null associations between BLLs
and an aggregate measure of single nucleotide polymorphisms constructed to represent susceptibility to
high BMIs.

Other recent studies were less informative due to small sample sizes. In a cross-sectional analysis
of adults of African descent across five countries of varying social and economic development in Africa,
Ettinger et al. (2014) compared the prevalence of being overweight (BMI >25) or being obese (BMI >30)
among subjects above versus below the median blood Pb exposure level (1.66 (ig/dL). Among subjects
with above median blood Pb, Ettinger et al. (2014) observed slightly reduced odds of being overweight
(OR = 0.88 [95% CI: 0.31, 2.51]), but increased odds of being obese (OR = 2.70 [95% CI: 0.75, 9.75]).
The observed associations, however, were notably imprecise due to the small sample size (n = 150). In
contrast, another small cross-sectional study of 145 adult men living in China observed a null association
between BLLs and BMI (Guo et al.. 2019). As is the case in both of these studies, limited statistical
power resulting from a small sample size simultaneously reduces the likelihood of detecting a true effect
and the likelihood that an observed result reflects a true effect, which might explain the incongruous
results.

9.2.4 Toxicological Studies on Metabolic Effects

The 2013 Pb ISA did not have a section devoted to toxicological studies related to the effect of Pb
on metabolism. However, as discussed in the Section 9.1.4, a few studies evaluated in the 2013 Pb ISA
demonstrated that Pb exposure can impair lipid metabolism in animals, as evidenced by increased hepatic
cholesterogenesis, and altered triglyceride and phospholipid levels (Sharma et al.. 2010; Ademuviwa et
al.. 2009; Khotimchenko and Kolenchenko. 2007). The relevance of the toxicological evidence is
uncertain, as many studies administered Pb as bolus doses and/or results were observed in animals with
high BLLs. In subsequent years, there have been a few PECOS-relevant publications on Pb exposure and
metabolic effects. In general, these studies cover disparate endpoints, but provide some evidence of Pb-
induced changes in metabolic activity in rodents.

External Review Draft

9-23

DRAFT: Do not cite or quote

-------
1

In a lifetime study using mice, Faulk et al. (2014) assessed perinatal Pb exposures via Pb acetate
in drinking water from conception to weaning. Average maternal BLLs for exposed groups ranged from
4.1 to 32 (ig/dL. The study findings included sex-specific increases in energy expenditure, food intake,
body weight, total body fat, activity, and insulin response. In addition, a study in weanling rats that
focused on neuropathology found that lead exposure decreased cholesterol levels in brain tissue (Zhou et
al.. 2018). The latter study, which also used Pb acetate in drinking water, reported BLLs ranging from
14.7 to 28.9 (ig/dL. Finally, in an investigation of the effects of vitamin D metabolism in rats, Rahman et
al. (2018) reported that Pb interferes with vitamin D metabolism by affecting the expression of its
metabolizing enzymes.

9.2.5 Summary and Causality Determination

There was no causality determination for metabolic effects in the 2013 Pb ISA (U.S. EPA. 2013).
The number of studies examining Pb exposure and metabolic effects has expanded substantially since the
2013 Pb ISA (U.S. EPA. 2013). highlighted by a number of recent epidemiologic studies, as well as a few
animal toxicological studies currently available for review. The focus of this causality determination is on
altered glucose resistance, diabetes mellitus, MetS, and obesity. Notably, there is significant overlap
between components of metabolic health and the cardiovascular and hepatic systems. While blood
pressure and serum lipids are important components of MetS, they are also discussed in detail in the
cardiovascular effects appendix (Appendix 4) and hepatic effects section (Section 9.1), and contribute to
the causality determinations therein. For the metabolic effects causality determination, these endpoints are
considered to the extent that they contribute to a diagnosis of MetS.

There is some evidence from a limited number of animal toxicological studies that exposure to Pb
resulting in BLLs relevant to humans alters cholesterol metabolism (Zhou et al.. 2018) and leads to
increases in body weight, body fat, and insulin response (Faulk et al.. 2014). In contrast, recent
epidemiologic studies are inconsistent across a range of metabolic outcomes and thus not coherent with
the limited toxicological evidence. A limited number of cross-sectional studies examining diabetes
prevalence and insulin resistance in adults reported null (Hansen et al.. 2017; Simic et al.. 2017) and
negative (Moon. 2013) associations with BLLs. Further, results from analyses of MetS in large national
surveys in the United States and Korea were largely inconsistent. Many of these same studies also provide
generally inconsistent evidence of associations between BLLs and individual components of MetS,
though there is substantial epidemiologic and toxicological evidence that exposure to Pb leads to
increased blood pressure and hypertension (Section 4.3). While a limited number of KNHANES analyses
demonstrate consistent associations between BLLs and serum triglycerides (Lee and Kim. 2016. 2013;
Rhee et al.. 2013). these studies include overlapping study populations and therefore do not provide
independent evidence of associations. Additionally, a recent prospective cohort study of older adults
observed null associations between BLLs at baseline and serum triglyceride levels measured three to four
years later (Peters et al.. 2012). Despite observed associations between BLLs and some of the individual

External Review Draft

9-24

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

components of MetS, the available evidence examining the cluster of components does not consistently
associate BLLs with MetS. Collectively, given the insufficient quantity of toxicological studies and
inconsistency in epidemiologic results, the evidence is inadequate to infer the presence or absence of a
causal relationship between Pb exposure and metabolic effects.

9.3 Effects on the Gastrointestinal System

9.3.1 Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

The 2013 Pb ISA concluded that "because of the insufficient quantity and quality of studies, the
available evidence was inadequate to determine if there is a causal relationship between Pb exposure and
gastrointestinal effects" (U.S. EPA. 2013). There were very few studies evaluated in the 2013 Pb ISA that
examined Pb exposure and gastrointestinal (GI) effects in humans or animals. Epidemiologic evidence of
an association between Pb exposure and GI effects was limited to a small number of occupational studies
of prevalent symptoms in Pb-exposed workers. The internal validity and generalizability of these studies
was limited by cross-sectional study designs, lack of consideration for potential confounders, and notably
higher BLLs (>40 (ig/dL) than those experienced by the general population. In addition to the
epidemiologic evidence, there were a limited number of toxicological studies that provide evidence of Pb-
induced effects on mechanisms underlying GI damage and impaired function.

9.3.2 Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria

1 The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

External Review Draft

9-25

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological
plausibility of Pb-associated effects on the gastrointestinal system, recent studies were only included if
they satisfied all of the components of the following discipline-specific PECOS statements:

Epidemiologic Studies

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Exposure: Exposure to Pb1 as indicated by biological measurements of Pb in the body - with a
specific focus on Pb in blood, bone, and teeth; validated environmental indicators of Pb
exposure2; or intervention groups in randomized trials and quasi-experimental studies.

Comparison: Populations, population subgroups, or individuals with relatively higher versus
lower levels of the exposure metric (e.g., per unit or log unit increase in the exposure metric,
or categorical comparisons between different exposure metric quantiles).

Outcome: Effects on the gastrointestinal system.

Study Design: Epidemiologic studies consisting of longitudinal and retrospective cohort studies,
case-control studies, cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest, randomized trials, and quasi-experimental studies examining
interventions to reduce exposures.

Experimental Studies

Population: Laboratory nonhuman mammalian animal species (i.e., mouse, rat, Guinea pig,

minipig, rabbit, cat, dog; whole organism) of any lifestage (including preconception, in utero,
lactation, peripubertal, and adult stages);

1	Recent studies of occupational exposure to Pb were considered insofar as they addressed a topic area of particular
relevance to the NAAQS review (e.g., longitudinal studies designed to examine recent versus historical Pb
exposure).

2	Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM25) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

External Review Draft

9-26

DRAFT: Do not cite or quote

-------
1

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.1,2

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

Outcomes: Effects on the gastrointestinal system.

Study design: Controlled exposure studies of animals in vivo.

9.3.3 Epidemiologic Studies on the Gastrointestinal System

The epidemiologic evidence evaluated in the 2013 Pb ISA was limited to a small number of
occupational cohort studies of prevalent GI symptoms in Pb-exposed workers (U.S. EPA. 2013). As noted
in Section 9.3.1, these studies had a number of limitations, including cross-sectional study designs, lack
of consideration for potential confounders, and notably higher BLLs (>40 (ig/dL) than those experienced
by the general population. There are no recent PECOS-relevant epidemiologic studies that evaluate
potential associations between exposure to Pb and effects on the gastrointestinal system. A limited
number of studies reported associations between BLLs and gut microbiota diversity, as discussed in the
Immune System Effects Appendix (Section 6.6). However, these studies do not inform the relationship
between Pb exposure and specific GI health effects.

9.3.4 Toxicological Studies on the Gastrointestinal System

In the 2013 Pb ISA (U.S. EPA. 2013). specific attention was drawn to a pair of rat studies; one
reporting frequency-dependent inhibition of electric field-stimulated relaxations to nonadrenergic
noncholinergic (NANC) nerve stimulation in rat gastric fundus (possibly due to the modulated release of
NO), and the other focusing on Pb-induced oxidative stress in the gastric mucosa, wherein an increase in
gastric mucosal damage induced by the acidified ethanol was observed. Neither of these studies reported
BLLs. Neither of the two pertinent studies since the 2013 Pb ISA directly addresses these findings
lYReddv et al.. 2018; Kosik-Bogacka et al.. 2011); see below].

In a chronic exposure study with rats, Kosik-Bogacka et al. (2011) confirmed an inhibitory effect
of Pb on electrophysiological parameters, among other findings. These findings were strengthened by

1 Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

2 This level represents an order of magnitude above the upper end of the distribution of U.S. young children's BLL.
The 95th percentile of the 2011-2016 NHANES distribution of BLL in children (1-5 years; n= 2,321) is 2.66 (ig/dL
(Eganetal.. 20211 and the proportion of individuals with BLL that exceed this concentration varies depending on
factors including (but not limited to) housing age, geographic region, and a child's age, sex, and nutritional status.

External Review Draft

9-27

DRAFT: Do not cite or quote

-------
1

results showing the ability of L-ascorbic acid to (at least partially) abrogate the effects of Pb exposure.
Mean BLLs in this study were reported at 7 (ig/dL.

In a 2018 microbiome study, Reddv et al. (2018) found that Pb-exposed rats had decreased 8-
aminolevulenic acid dehydratase (ALAD) activity and intestinal lactobacillus levels, irrespective of the
dietary iron supplementation. Withdrawal of Pb exposure increased lactobacilli, whereas re-exposure to
Pb decreased lactobacilli population. BLLs were reported in the range of 19 to 48 (ig/dL.

9.3.5 Summary and Causality Determination

The 2013 Pb ISA concluded that evidence was "inadequate" to determine a causal relationship
between Pb exposure and GI effects (U.S. EPA. 2013). This causality determination was based on an
insufficient quantity and quality of studies in the cumulative body of evidence. A limited number of
occupational cohort studies indicated associations between BLLs and prevalent symptoms, such as
stomach pain, gastritis, constipation, and intestinal paralysis. However, the implications of these findings
are limited by the cross-sectional study designs, high BLLs associated with effects (mostly >40 (.ig/dL).
and limited consideration of potential confounding by factors such as age, smoking, alcohol use, nutrition,
or other occupational exposures. Toxicological evidence indicates that Pb is absorbed primarily in the
duodenum by active transport and diffusion, although variability is observed by Pb compound, age of
intake, and nutritional factors. There was some coherence between the evidence in Pb-exposed workers
and observations in animals that Pb induces damage to the intestinal mucosal epithelium, decreases
duodenum contractility and motility, reduces absorption of calcium ions (Ca2+), inhibits NANC
relaxations in the gastric fundus, and induces oxidative stress (lipid peroxidation, decreased SOD and
CAT) in the gastric mucosa.

Recent studies are limited in number, and while some provide potential biological plausibility for
Pb-induced GI effects, none directly inform the relationship between Pb exposure and GI effects. Given
the insufficient quantity and quality of studies, the evidence remains inadequate to infer the presence or
absence of a causal relationship between Pb exposure and gastrointestinal effects.

9.4 Effects on the Endocrine System

9.4.1 Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

The 2013 Pb ISA (U.S. EPA. 2013) evaluated a limited number of studies examining the
relationship between exposure to Pb and effects on the endocrine system. Epidemiologic and
toxicological evidence related to male and female sex hormones, which was generally inconsistent, is

External Review Draft

9-28

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

discussed in more detail in Appendix 8 (Sections 8.6.1.1 and 8.7.2). In addition to studies on sex
hormones, results from a small number of epidemiologic and toxicological studies on Pb-associated
endocrine effects such as changes in thyroid hormones, Cortisol, corticosterone, and vitamin D levels were
also inconsistent. Further, epidemiologic studies were mostly cross-sectional and included limited
consideration for potential confounders. As a whole, the limited quantity, quality, and consistency of the
available evidence was "inadequate to determine if there is a causal relationship between Pb exposure and
endocrine effects related to thyroid hormones, Cortisol, and vitamin D."

9.4.2 Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria
than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological
plausibility of Pb-associated effects on the gastrointestinal system, recent studies were only included if
they satisfied all of the components of the following discipline-specific PECOS statements:

Epidemiologic Studies:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

1 The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

External Review Draft

9-29

DRAFT: Do not cite or quote

-------
1

Exposure: Exposure to Pb1 as indicated by biological measurements of Pb in the body - with a
specific focus on Pb in blood, bone, and teeth; validated environmental indicators of Pb
exposure2; or intervention groups in randomized trials and quasi-experimental studies.

Outcome: Effects on the endocrine system.

Experimental Studies:

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.3'4

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

Outcomes: Effects on the endocrine system.

Study design: Controlled exposure studies of animals in vivo.

9.4.3 Epidemiologic Studies on the Endocrine System

A limited number of epidemiologic studies evaluated in the 2013 Pb ISA (U.S. EPA. 2013)
reported associations between exposure to Pb and endocrine effects related to changes in thyroid

2 Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM2 5) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

3 Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

External Review Draft

9-30

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

hormones, Cortisol, and vitamin D levels. However, most studies were cross-sectional in design, and
many did not consider potential confounding factors. Further, while some studies did find associations
between Pb exposure and endocrine effects, the results for specific hormones were not consistent.

A limited number of recent epidemiologic studies of Pb exposure and endocrine effects also
implement cross-sectional analyses but included more robust adjustment for potential confounding
factors, including use of thyroid medication. The majority of recent studies are large NHANES analyses
that provide generally consistent evidence of null associations between Pb exposure and endocrine effects
of thyroid hormone and Cortisol levels. However, given that these studies examined overlapping study
populations, the generally consistent results across these studies should not be considered independent
evidence of a null association. Most recent studies evaluated potential associations between Pb exposure
and thyroid hormone levels, including triiodothyronine (T3), thyroxine (T4), and thyroid stimulating
hormone (TSH). There were a few studies that looked at associations between Pb exposure and Cortisol
levels and no recent PECOS-relevant studies that looked at Pb exposure and vitamin D levels. Measures
of study-specific BLLs and endocrine effect estimates are highlighted in Table 9-9. An overview of recent
evidence is provided below.

The most consistent evidence from recent studies indicates null associations between BLLs and
TSH, T3, and free T4 (FT4) levels in adults. A few recent NHANES analyses, which included nationally
representative study populations of adults over 20 years old, reported null associations between BLL and
TSH levels in adults (Krieg. 2019; Chen et al.. 2013; Mendv et al.. 2013; Christensen. 2012). Recent
NHANES analyses also provide generally consistent evidence of null associations between BLLs and
FT4 levels (Luo and Hendrvx. 2014; Chen et al.. 2013; Mendv etal.. 2013) as well as between blood Pb
and T3 levels (Nie et al.. 2017; Luo and Hendrvx. 2014; Chen et al.. 2013; Mendv et al.. 2013;
Christensen. 2012).

Recent NHANES studies evaluating a potential association between BLLs and total T4 levels
were less consistent. While some recent studies reported null associations between BLLs and T4 levels in
adults (Luo and Hendrvx. 2014; Chen et al.. 2013). others observed negative associations (Kricg. 2019:
Mendv et al.. 2013; Christensen. 2012). For example, Mendv et al. (2013) noted a 0.162 (ig/dL (95% CI:
-0.321, -0.004 (ig/dL) decrease in T4 per 1 (ig/dL increase in BLL. Additionally, while Luo and Hendrvx
(2014) noted a null association between BLLs and T4 levels in the total population, the authors observed
a significant negative association between blood Pb and T4 levels among men after stratifying by sex.
Krieg (2019) also found a negative association between blood Pb and T4 levels, reporting a 38.91% (95%
CI: -51.25, -23.44) decrease in T4 per 1 (ig/dL increase in blood lead level.

A limited number of NHANES analyses evaluated potential associations between blood Pb and
free T3 (FT3) levels (Luo and Hendrvx. 2014; Chen et al.. 2013; Mendv et al.. 2013). In an analysis of
adults, Mendv et al. (2013) reported a null association between blood Pb and FT3 levels in the general
adult population. This is consistent with the findings of Chen et al. (2013). who reported a null
association between BLLs and FT3 levels in both adolescents (12-19 years old) and adults (>20 years

External Review Draft

9-31

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

old). Both studies performed analyses on the 2007-2008 continuous NHANES cycle. Luo and Hendrvx
(2014) evaluated 2007-2010 data, reporting a positive association between blood Pb and FT3 in the
general adult population. The authors reported a modest 0.04 (ig/dL (95% CI: 0.01, 0.08) increase in FT3
per 1 (ig/dL increase in blood Pb in adults in the highest tertile of blood Pb when compared to the lowest.
After stratifying by sex, males were also found to have a positive association with a 0.05 (ig/dL (95% CI:
0.01, 0.09) increase in FT3 per 1 (ig/dL increase of blood Pb in the highest tertile compared to the lowest.

In addition to the NHANES analyses discussed above, another recent cross-sectional study
examined the relationship between BLLs and thyroid hormone levels in a small study of pregnant women
(n = 291) from the Yugoslavia Prospective Study of Environmental Lead Exposure Cohort (Kahn ct al..
2014). Kahn et al. (2014) reported a null association between BLL and TSH levels and a negative
association between blood Pb and FT4 levels.

Two recent cross-sectional studies examined associations between BLLs and Cortisol levels
(Ngueta et al.. 2018; Souza-Talarico et al.. 2017). In a small study of older adults (n = 65) in Montreal,
Canada, Ngueta et al. (2018) reported null associations between BLLs and both diurnal and stress-reactive
Cortisol secretion. In contrast, another small study of non-occupationally exposed Brazilian older adults
(n = 126), Souza-Talarico et al. (2017) reported positive associations between BLLs and both Cortisol
awakening response (CAR) and overall Cortisol concentration. The authors reported a 0.791 (ig/dL (95%
CI: 0.672, 1.073 (ig/dL) increase in CAR per 1 (ig/dL increase in BLL. However, it is worth noting that
participants showed an elevated basal circadian level of salivary Cortisol independent of Pb exposure,
suggesting this population has more repeated exposure to stressful events. Furthermore, while all
participants were older postmenopausal adults, sex was unevenly represented with n = 105 (83%) of the
participants being women.

9.4.4 Toxicological Studies on the Endocrine System

The 2013 Pb ISA summarized a few toxicological studies that reported on effects of Pb exposure
on the endocrine system. Specifically, T3 and T4 levels were found to be elevated in cows that were
grazing on land near Pb/operational Zn smelters when compared with cows grazing in unpolluted areas
(Swarup et al.. 2007). However, when regression analyses were done to evaluate potential associations
between BLLs and plasma Cortisol levels in these same cows, no association was observed. Another study
conducted in Wistar rats reported that 21 days of intraperitoneal (i.p.) injections with 8.0 mg/kg Pb led to
increased corticosterone levels and adrenal weights [BLLs not reported; (Biswas and Ghosh. 2006)1.

Some recent studies have also investigated the effects of Pb on the endocrine system (Table 9-5). The
only studies that investigated adrenal gland weight were conducted in Sprague Dawley rats that were
dosed from postnatal day (PND) 4 to 28 and reported no effect of Pb treatment on the weight of the
adrenal glands [BLLs 3.27-12.5 (ig/dL; (Amos-Kroohs et al.. 2016; Graham etal.. 2011)1. Findings
concerning corticosterone levels in recent studies are equivocal. Some studies reported increased

External Review Draft

9-32

DRAFT: Do not cite or quote

-------
1

corticosterone in rats exposed to Pb. Specifically, one study that dosed Long-Evans rats starting prior to
conception until 304 days of age reported increases in corticosterone levels in female rats at 2 months of
age but reported no changes in males at any time point [BLLs 11.3 (ig/dL on PND 61 in females; (Rossi-
George ct al.. 2011)1. Another study measured corticosterone levels in Sprague Dawley rats at different
intervals following a shallow water stressor. This study reported that treatment with Pb from PND 4 to 28
increased corticosterone levels in male and female rats 0, 30, and 60 minutes following the stressor on
PND 11,0 and 30 minutes following the stressor on PND 19, and 0 and 30 minutes following the stressor
on PND 29 [BLLs 3.2-12.5 (ig/dL on PND 29; (Graham et al.. 2011)1. A single study reported decreases
in corticosterone in F3 female C57 BL/6 mice whose F1 sires were exposed to Pb from gestational day
(GD) -61 to PND 21 [BLLs 0.4 (ig/dL on PND 6-7; (Sobolcw ski et al.. 2020)1. Contrasting these studies
are those that did not report any effects of Pb exposure on corticosterone levels. Interestingly, these
studies used similar dosing paradigms to those that reported effects with one study dosing C57 BL/6 mice
starting preconceptionally through adulthood [ending on PND 365; (Corv-Slechta et al.. 2013)1 and the
other study dosing Sprague Dawley rats from PND 4 to 28 (Amos-Kroohs et al.. 2016). and neither study
reported alterations of corticosterone levels in either sex.

9.4.5 Summary and Causality Determination

The 2013 Pb ISA concluded that the evidence was inadequate to determine if there is a causal
relationship between Pb exposure and endocrine effects related to changes in levels of thyroid hormones,
cortisol/corticosterone, and vitamin D. This causality determination was based on an insufficient quantity
and quality of studies that provided inconsistent or inconclusive evidence for Pb-related endocrine effects.
Epidemiologic evidence presented in the 2013 Pb ISA regarding the effects of Pb on Cortisol levels
consisted of a single study showing a positive association between prenatal Pb exposure and salivary
Cortisol levels in children following an acute stressor (Gump et al.. 2008). The few epidemiologic studies
investigating associations between Pb and thyroid hormone levels presented in the 2013 Pb ISA reported
inconsistent associations. Toxicological evidence in the 2013 Pb ISA regarding the effects of Pb on the
endocrine system in animals was sparse. Biswas and Ghosh (2006) reported that Pb exposure increased
corticosterone levels and adrenal gland weights in Wistar rats. A single study evaluating thyroid hormone
levels in animals summarized in the 2013 Pb ISA reported no clear associations between Pb exposure and
thyroid hormone levels in cattle with environmental exposure to Pb (Swarup et al.. 2007).

Recent epidemiologic and toxicological evidence evaluating the effects of Pb exposure on the
endocrine system continues to be limited and inconsistent. The most recent epidemiologic studies
measured associations between BLLs and thyroid hormone levels. Results from these studies were mostly
null, though there was some inconsistent evidence of an inverse association between BLLs and T4 levels
in three studies (Kricg. 2019; Mendv et al.. 2013; Christensen. 2012). and a single study noted sex-
specific associations between BLLs and T4 and FT4 levels (Luo and Hendrvx. 2014). While the results
are generally consistent, the analyses include overlapping study populations, so they should not be

External Review Draft

9-33

DRAFT: Do not cite or quote

-------
1

interpreted as independent evidence of a null association. Additionally, consistent with the studies
evaluated in the 2013 Pb ISA, recent studies are cross-sectional in design, which introduces uncertainty
about the temporality between exposure and outcome. No recent toxicological studies investigating the
effects of Pb on thyroid hormone levels were available. Only a few recent epidemiologic studies
examined Pb effects on Cortisol levels (Ngucta et al.. 2018; Souza-Talarico et al.. 2017). Interns of
associations of Pb exposure and Cortisol levels in humans, evidence was limited and inconsistent. Only
two studies were available that measured Pb exposure with Cortisol outcomes (Ngucta et al.. 2018; Souza-
Talarico et al.. 2017). and both had small sample sizes. Multiple toxicological studies reported on the
effects of Pb exposure on corticosterone levels in animals, but results are equivocal. One study reported
decreases (Sobolewski et al.. 2020). two studies reported increases (Graham etal.. 2011; Rossi-George et
al.. 2011). and two studies reported no effect (Amos-Kroohs et al.. 2016; Corv-Slechta et al.. 2013) on
corticosterone levels in Pb-intoxicated animals. In terms of the effects of Pb on adrenal gland weights in
animals, only two recent studies investigated the effects of Pb on adrenal gland weight. These studies
reported no effects of Pb on adrenal gland weight in Sprague Dawley rats (Amos-Kroohs et al.. 2016;
Graham et al.. 2011). contrasting with the only study that investigated adrenal gland weights in the 2013
Pb ISA (Biswas and Ghosh. 2006) which reported increased adrenal gland weights. This contrast may be
due to variability in route of exposure used in the experimental design leading to differences in BLLs
between the animals in Biswas and Ghosh (2006). and the more recent studies. Specifically, Biswas and
Ghosh (2006) dosed animals with 8 mg/kg/d of Pb via i.p. injection, whereas the most recent publications
dosed animals with either 1 or 10 mg/kg/d of Pb b via oral gavage (Amos-Kroohs et al.. 2016) or
indirectly dosed animals via Pb in the milk from their dams which were dosed via oral gavage (Graham et
al.. 2011). No recent PECOS-relevant epidemiologic or toxicological studies were identified that
measured vitamin D levels.

In conclusion, recent epidemiologic and toxicological studies continue to provide limited and
inconsistent evidence for endocrine system effects associated with Pb exposure. Due to the insufficient
quantity and quality of the studies available for review and the inconsistent results across those studies,
the evidence remains inadequate to infer the presence or absence of a causal relationship between Pb
exposure and endocrine effects related to changes in thyroid hormones, cortisol/corticosterone, and
vitamin D levels.

9.5 Effects on the Musculoskeletal System

9.5.1 Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

The 2013 Pb ISA evaluated the effects of Pb exposure on bone and teeth (U.S. EPA. 2013). In
order to be more inclusive of other health effects related to bone and teeth, this ISA expands the

External Review Draft

9-34

DRAFT: Do not cite or quote

-------
1

considered health outcomes to include effects on the entire musculoskeletal system. The musculoskeletal
system consists of the bones, teeth, muscles, joints, cartilage, and other connective tissues that support the
body, allow for movement, and protect vital organs. Primary effects on the musculoskeletal system
include increases in osteoporosis, increased frequencies of falls and fractures, changes in bone cell
function as a result of replacement of bone calcium with Pb, and depression in early bone growth. Other
effects include tooth loss and periodontitis. Mechanistic evidence from toxicological studies includes
effects on cell proliferation, procollagen type I production, intracellular protein, and osteocalcin in human
dental pulp cell cultures.

A small body of epidemiologic studies evaluated in the 2013 Pb ISA (U.S. EPA. 2013) provided
consistent evidence of associations between Pb biomarker levels and various effects on bone and teeth,
including an increase in osteoporosis, increased frequencies of falls and fractures, tooth loss, and
periodontitis. The results from these studies, adjusting for potential confounding by age and SES-related
factors, were supported by strong toxicological evidence evaluated in the 2013 Pb ISA and the 2006 Pb
AQCD (U.S. EPA. 2006). which reported effects in bone and teeth in animals following Pb exposure.
Exposure of animals to Pb during gestation and the immediate postnatal period was reported to
significantly depress early bone growth with the effects showing concentration-dependent trends.

Systemic effects of Pb exposure included disruption in bone mineralization during growth, alteration in
bone cell differentiation and function due to alterations in plasma levels of growth hormones and
calcitropic hormones such as l,25-[OH]2D3 and impact on Ca2+- binding proteins and increases in Ca2+
and phosphorus concentrations in the bloodstream. Bone cell cultures exposed to Pb had altered vitamin
D-stimulated production of osteocalcin accompanied by inhibited secretion of bone-related proteins such
as osteonectin and collagen. In addition, Pb exposure caused suppression in bone cell proliferation most
likely due to interference from factors such as growth hormone (GH), epidermal growth factor (EGF),
transforming growth factor-beta 1 (TGF-(31), and parathyroid hormone-related protein (PTHrP).

As in bone, Pb exposure was found to easily substitute for Ca2+ in the teeth and was taken up and
incorporated into developing teeth in experimental animals. Since teeth do not undergo remodeling like
bones do during growth, most of the Pb in the teeth remains in a state of permanent storage. Pb has also
been shown to decrease cell proliferation, procollagen type I production, intracellular protein, and
osteocalcin in human dental pulp cell cultures. Adult rats exposed to Pb have exhibited an inhibition of
the posteruptive enamel proteinases, delayed teeth eruption times, as well as a decrease in microhardness
of surface enamel. Further discussion of these processes and effects, including corresponding references,
can be found in sections 5.8.7 through 5.8.13 of the 2006 AQCD (U.S. EPA. 2006).

In considering the weight of the evidence, the 2013 Pb ISA (U.S. EPA. 2013) concluded that "a
causal relationship is likely to exist between Pb exposure and effects on bone and teeth."

External Review Draft

9-35

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

9.5.2

Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria, thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria
than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological
plausibility of Pb-associated effects on the musculoskeletal system, recent studies were only included if
they satisfied all of the components of the following discipline-specific PECOS statements:

Epidemiologic Studies:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Exposure: Exposure to Pb2 as indicated by biological measurements of Pb in the body - with a
specific focus on Pb in blood, bone, and teeth; validated environmental indicators of Pb
exposure3; or intervention groups in randomized trials and quasi-experimental studies.

Comparison: Populations, population subgroups, or individuals with relatively higher versus
lower levels of the exposure metric (e.g., per unit or log unit increase in the exposure metric,
or categorical comparisons between different exposure metric quantiles).

Outcome: Effects on the musculoskeletal system.

1	The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

2	Recent studies of occupational exposure to Pb were considered insofar as they addressed a topic area of particular
relevance to the NAAQS review (e.g., longitudinal studies designed to examine recent versus historical Pb
exposure).

3	Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM2 5) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

External Review Draft

9-36

DRAFT: Do not cite or quote

-------
1

Experimental Studies:

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.1,2

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

Outcomes: Effects on the musculoskeletal system.

Study design: Controlled exposure studies of animals in vivo.

9.5.3 Epidemiologic Studies on the Musculoskeletal System

A limited number of cross-sectional epidemiologic studies evaluated in the 2013 Pb ISA (U.S.
EPA. 2013) provided consistent evidence of associations between Pb biomarker levels and osteoporosis
and tooth loss after adjusting for potential confounding by age and SES-related factors. Uncertainties in
the evidence base included limited consideration of potential confounding by nutritional factors, a lack of
temporality between exposure and outcome, and uncertainty in the level, timing, frequency, and duration
of Pb exposure that contributed to the observed associations. Recent epidemiologic studies of the
musculoskeletal system generally examine one of three groups of endpoints: (1) bone mineral density
(BMD); (2) joint degeneration; and (3) oral health. Results from recent studies, which adjust for a range
of potential confounders, provide generally consistent evidence of an association between BLLs and
osteoporosis, osteoarthritis, dental caries, and periodontal disease. Recent studies evaluating
musculoskeletal effects are largely cross-sectional analyses, which are unable to establish temporality
between exposure and outcome. Additionally, with BLLs, it is difficult to characterize the specific timing,
duration, frequency, and level of Pb exposure that contributed to the observed associations. This
uncertainty may apply particularly to assessments of BLLs, which in nonoccupationally-exposed adults,
reflect both current exposures and cumulative Pb stores in bone that are mobilized during bone

1 Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

External Review Draft

9-37

DRAFT: Do not cite or quote

-------
1

remodeling. Measures of central tendency for Pb biomarker levels used in each study, along with other
study-specific details, including study population characteristics and select effect estimates, are
highlighted in Table 9-11. An overview of the recent evidence is provided below.

9.5.3.1 Bone Mineral Density

A number of recent cross-sectional studies provide generally consistent evidence of an
association between exposure to Pb and BMD in adults. In these studies, BMD (g/cm2) was measured via
X-ray absorptiometry or ultrasound and often converted to a standardized score (i.e., z- and Z-scores)1.
Osteoporosis and osteopenia are characterized by varying degrees of BMD decrements that can
compromise bone microarchitecture. A z-score below -1 often corresponds to osteopenia, whereas a z-
score below -2.0 to -2.5 is categorized as osteoporosis. There are significant sex and age differences in
the incidence of osteoporosis and osteopenia, with postmenopausal women being at greatest risk for
declines in BMD. Because osteoporosis and osteopenia are more common in women, many of the recent
epidemiologic studies evaluating the relationship between BLLs and BMD are either conducted in study
populations comprised of older women or stratified by sex. Importantly, the cross-sectional nature of the
studies does not rule out the possibility that the association is driven by increased BLLs due to higher
bone turnover in individuals with osteoporosis. Additionally, although most analyses include study
populations with mean BLLs <3 (ig/dL, study participants were born prior to the phase-out of leaded
gasoline and therefore likely had much higher past Pb exposures, making it difficult to characterize the
specific timing, duration, frequency, and level of Pb exposure that contributed to the observed
associations.

A few recent analyses of data from large, nationally representative health surveys provide
generally consistent evidence of an association between BLLs and BMD in women (Wang et al.. 2019;
Cho et al.. 2012; Lee and Kim. 2012). In an analysis of 2008 KNHANES data, Cho et al. (2012) observed
increased odds of osteoporosis associated with increasing BLL quartiles in postmenopausal women. The
authors noted associations at low levels (e.g., Q2 [1.83 to <2.32 |ig/dL| versus quartile 1 [<1.83 |ig/dL|)
that were similar in magnitude to comparisons between the higher quartiles and the first quartile,
suggesting a potentially non-linear association. In a similar study, Lee and Kim (2012) analyzed data
from the same KNHANES cycle but expanded the age range to include premenopausal women. The
authors reported that increases in BLLs were associated with decreased BMD at several bone sites.
Additionally, Pb-related BMD decrements were consistently higher in postmenopausal women compared
to premenopausal women. For example, a 1 (ig/dL increase in BLLs was associated with a -0.28 g/cm2

1 Standardized scores are used to analyze BMD data as deviations from average BMD in matched healthy
populations. Underlying populations vary by study.

External Review Draft

9-38

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

(95% CI: -0.45, -0.11 g/cm2) decrease in femoral BMD in postmenopausal women compared to a
-0.15 g/cm2 (95% CI: -0.33, 0.03 g/cm2) decrease in premenopausal women.

In contrast to KNHANES analyses, an analysis of more recent NHANES cycles (2013-2014)
observed null associations between BLLs and BMD in postmenopausal women (Wang et al.. 2019).
Notably, the authors did not control for hormone therapy, which could impact BLLs due to changes in
bone turnover rates. Wang et al. (2019) did note that a 1 (ig/dL increase in BLLs was associated with
small decrements in femoral (-0.06 g/cm2 [95% CI: -0.08, -0.03 g/cm2]) and spinal (-0.05 g/cm2 [95%
CI: -0.08, -0.02 g/cm2]) BMD in premenopausal women, as well as increases in 10-year fracture risk
scores in the total population (including adult men and women). The findings in premenopausal women
are somewhat consistent with a recent cross-sectional analysis of premenopausal women in western New
York that observed a 0.02 (-0.02, 0.05) g/cm2 decrease spinal BMD associated with a 1 (ig/dL increase in
BLLs (Pollack et al.. 2013). However, in contrast to the results from Wang et al. (2019). Pollack et al.
(2013) reported null associations between BLLs and total hip and wrist BMD in premenopausal women.

In a smaller cross-sectional analysis of adults from two communities in southwestern China,
including one with a history of Pb mining and smelting, Li et al. (2020b) observed some evidence of sex-
specific differences in Pb-associated BMD levels. Specifically, female study participants with BLLs
>3.4 (ig/dL had increased odds of osteoporosis compared to female study participants with BLLs
<3.4 (ig/dL (OR= 1.33 [95% CI: 0.61, 2.88]); whereas an inverse association was reported formen
(OR = 0.60 [95% CI: 0.24, 1.49]). However, given the imprecise effect estimates (i.e., wide 95% CIs), it
is difficult to draw firm conclusions on these sex-specific comparisons.

Other recent studies evaluated the relationship between Pb exposure and BMD in analyses
combining men and women. The inferences that can be drawn from these studies are limited due to
established sex-specific differences in osteoporosis incidence. In an analysis of 2008-2011 KNHANES
cycles, Lim et al. (2016) observed increased odds of osteoporosis or osteopenia across BLL quartiles,
with the largest increase in odds noted in quartile 4 (>2.93 (ig/dL) compared to quartile 1 (<1.66 (ig/dL;
OR = 1.49 [95% CI: 1.12, 1.98]). In a much smaller study of Korean adults, Lee and Park (2018)
similarly reported a decrease in BMD t-scores associated with a 1 (ig/dL increase in BLLs that was
greater in magnitude in participants with a history of smoking (-0.472 [95% CI-0.85, -0.094]) compared
to non-smokers (-0.148 [95% CI: -0.369, 0.073]). The authors also examined over 344,396 single
nucleotide polymorphisms (SNPs) mapped to gene-coding regions to assess potential interactions
between BLLs and genetic variations. The observed interactions were inconsistent after adjustment for
multiple testing, but many implicated genes and pathways involved in angiogenesis, bone mass, and
nuclear receptor signaling, provide areas of interest for exploring possible mechanisms that may underlie
the observed relationship between BLLs and osteoporosis.

External Review Draft

9-39

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

9.5.3.2

Osteoarthritis

A few recent cross-sectional studies examined the association between BLLs and osteoarthritis
(OA) in adults. In an analysis of multiple KNHANES cycles (2010-2012), Park and Choi (2019) reported
that an increase in natural log BLLs was associated with an increase in the odds of radiographic and
symptomatic knee OA (radiographic osteoarthritis [rOA] and symptomatic osteoarthritis [sxOA]) in
postmenopausal women (OR= 1.77 [95% CI 1.17, 2.67] and 1.50 [95% CI: 0.90, 2.53], respectively).
There is some evidence that the association is mediated by BMI, but there is evidence of a direct
association as well (i.e., adjusted for BMI). The authors noted null associations between BLLs and back
OA.

In a cross-sectional analysis of African American and white adults, Nelson etal. (201 lb) also
observed associations between BLL and rOA and sxOA in the knee. In a similar study, the same group
noted associations between BLLs and some biomarkers of joint tissue metabolism, including NTX-I,
which is responsible for bone turnover; CTX-II, which is associated with prevalence of rOA in the knee;
COMP (cartilage oligomeric matrix protein), which is a cartilage biomarker; and CPU (carboxypropeptide
of type II collagen), which is linked with collagen synthesis (Nelson et al.. 201 la). Notably, the authors
examined a wide range of biomarkers and stratified their models by sex, increasing the likelihood of
multiple testing bias.

Although all of the studies examining OA had low median BLLs (<2.5 (.ig/dL). study participants
were born prior to the phase-out of leaded gasoline and therefore likely had much higher past Pb
exposures, making it difficult to characterize the specific timing, duration, frequency, and level of Pb
exposure that contributed to the observed associations. Additionally, similar to studies of osteoporosis,
the cross-sectional nature of the studies does not rule out the possibility that the association is driven by
cartilage turnover resulting in increased Pb in blood.

9.5.3.3 Oral Health

Recent epidemiologic studies of Pb exposure and oral health are split into two major categories:
(1) periodontal disease in adults and (2) dental caries in children.

A limited number of recent studies of periodontal disease in adults examined overlapping
KNHANES cycles from 2008 to 2010 (Han et al.. 2013; Kim and Lee. 2013; Won et al.. 2013). These
studies, all of which defined periodontal disease according to the World Health Organization's
Community Periodontal Index, provided consistent evidence of an association between BLLs and the
prevalence of periodontitis. All of the studies included extensive adjustment for potential confounders,
including oral hygiene. Given that these studies examined largely overlapping study populations, the
observed results should not be considered independent evidence of an association. Kim and Lee (2013)
noted associations that were stronger in magnitude in men (OR = 1.85 [95% CI: 1.26, 2.71] per doubling

External Review Draft

9-40

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

of BLL) compared to women (OR = 1.30 [95% CI: 0.88, 1.91] per doubling of BLL), and that
associations were slightly attenuated, but still positive after adjustment for blood mercury (Hg) and
cadmium (Cd; 1.69 [95% CI: 1.15, 2.50] and 1.24 [95% CI: 0.83, 1.85], respectively). In analyses that
stratified by smoking, effect estimates were imprecise (i.e., wide 95% CIs), but comparable in magnitude
for smokers and non-smokers (Han et al.. 2013; Won et al.. 2013).

Recent epidemiologic studies of dental caries in children included more diverse study
populations. A prospective analysis of mother-child pairs that recruited from hospitals serving low- to
moderate-income populations in Mexico examined the relationship between Pb biomarkers at different
developmental windows and incidence of decayed, missing, and filled teeth (DMFT) in adolescence [10
to 18 years old; Wu et al. (2019)1. The authors reported a 12 to 17% increase in risk of DMFT associated
with a natural log increase in prenatal and early childhood BLLs. No associations were observed with
concurrent BLLs or postnatal maternal bone Pb. Prenatal (mean: 5.24 to 6.36 (ig/dL) and early childhood
(mean: 15.18 to 15.48 (ig/dL) BLLs were notably higher than concurrent levels (mean: 3.60-3.34 (.ig/dL).
which is consistent with age-specific patterns of Pb kinetics (Sections 2.2 and 2.4). Wu et al. (2019)
additionally stratified their models by sugar sweetened beverage intake (SSBI) and observed stronger
associations between prenatal and early childhood BLLs and DMFT score in children with high SSBI. In
recent cross-sectional studies with lower BLLs (see Table 9-6), BLLs in young children were associated
with increased prevalence of dental caries in deciduous teeth (Kim et al.. 2017; Wiener et al.. 2015). but
not permanent teeth (Kim et al.. 2017).

9.5.4	Toxicological Studies on the Musculoskeletal System

The 2013 Pb ISA (U.S. EPA. 2013) evaluated a number of toxicological studies that
demonstrated changes in bone cell function as a result of replacement of bone calcium with Pb depression
in early bone growth. Studies also reported Pb-induced effects on cell proliferation, procollagen type I
production, intracellular protein, and osteocalcin in human dental pulp cell cultures. Earlier work,
summarized in the 2006 Pb AQCD (U.S. EPA. 2006). reported concentration-dependent depression of
early bone growth after gestational exposure of animals to Pb. Recent evidence is limited. In a study of
lifetime Pb exposure in mice, Beier et al. (2016) reported a reduction in osteoclast activity and a
subsequent disruption in bone accrual in Pb-exposed mice. In another publication, the same group
reported no other musculoskeletal effects resulting from Pb exposure alone (Beier etal.. 2017).

9.5.5	Biological Plausibility

This section describes biological pathways that potentially underlie musculoskeletal effects of Pb.
Figure 9-2 depicts the proposed pathways as a continuum of upstream events, connected by arrows, which
may lead to downstream events observed in epidemiologic studies. This discussion of how exposure to Pb

External Review Draft

9-41

DRAFT: Do not cite or quote

-------
1

may lead to musculoskeletal effects contributes to an understanding of the biological plausibility of
epidemiologic results evaluated above. Note that the structure of the biological plausibility sections and
the role of biological plausibility in contributing to the weight-of-evidence analysis used in the current Pb
ISA are discussed in Section IS.4.2.

The proposed pathway, outlined in Figure 9-2, involves both direct and indirect effects of Pb that
could plausibly result in the weakening of bones and increased risk of fractures as well as the dental
effects that are measured in epidemiologic studies. Skeletal bone development and biomechanical
strength is controlled by the balance between osteoblasts, the cells responsible for the production of bone
matrix, and osteoclasts, the cells responsible for bone resorption. Dysregulation of this balance can lead to
bone loss and decreased mineralization. Pb can directly replace Ca2+ in the bone matrix as well as exert
direct effects on bone cells to alter bone development. Pb can also alter bone growth and differentiation
signals that can further disrupt the balance of bone formation and resorption.

As discussed in the 2013 Pb ISA, Pb suppresses the differentiation of osteoblasts and promotes
osteoclast function which could result in delayed bone development and reduced bone mechanical
integrity. Recent literature supports this hypothesis as studies have continued to show that animals treated
with Pb have decreased bone mineralization (Li et al.. 2020a; Sheng et al.. 2020; Qi et al.. 2019;

Olchowik et al.. 2014). bone weight (Alvarez-Lloret et al.. 2017; de Figueiredo et al.. 2014). and reduced
trabecular bone (Li et al.. 2020a; Sheng et al.. 2020; Alvarez-Lloret et al.. 2017; Beier et al.. 2017). Many
of these studies show concurrent changes in osteoblastic and osteoclastic markers that support an overall
shift to increased bone resorption. For example, recent in vivo studies have seen reductions markers of
osteoblast differentiation (Qi etal.. 2019; Zhang et al.. 2019; Beier et al.. 2017). reductions of proteins
that suppress osteoclast activity (Li et al.. 2020a; Sheng et al.. 2020; Oi et al.. 2019; Kupraszewicz and
Brzoska. 2013). and increases of markers of osteoclast activity (Li et al.. 2020a; Qi et al.. 2019; Zhang et
al.. 2019; Kupraszewicz and Brzoska. 2013) suggesting that bone changes result from dysregulation of
the balance between bone formation and bone resorptive processes. The mechanism behind the reduced
osteoblastic activity is not fully understood but both direct and indirect mechanisms have been proposed.

Support for a direct action of Pb on osteoblast function comes from in vitro studies showing that
Pb treatment of primary osteoblasts leads to reduction in mineral deposition (Beier etal.. 2015; Abbas et
al.. 2013; Ma et al.. 2012). Previously reviewed data also implicated changes in TGF(3, bone morphogenic
protein (BMP), nuclear factor kappa B (NF-kB), and activator protein-1 signaling (U.S. EPA. 2013).
Recent studies suggest that Pb-induced suppression of Wnt signaling and upregulation of the protein
sclerostin may also be involved (Sun et al.. 2019; Beier et al.. 2017; Beier et al.. 2015). Similar studies of
dental pulp cultures showed that in vitro treatment with Pb resulted in decreased cell proliferation and
reduced extracellular matrix deposition. This could explain the increased incidence of dental carries in
epidemiology studies.

Indirect mechanisms of Pb treatment have also been discussed in the 2013 Pb ISA. The
replacement of Pb for Ca2+ in cells can lead to Ca2+ release. The 2013 Pb ISA and 2006 AQCD discussed

External Review Draft

9-42

DRAFT: Do not cite or quote

-------
1

studies that found that Pb treatment leads to increased systemic Ca2+ levels in the blood stream (U.S.
EPA. 2013. 2006). Calcium is a cellular signaling molecule involved in mitochondrial function and cell
death and thus changes in calcium signaling could have effects on cells elsewhere in the body. Bone
growth can be affected by systemic signaling of hormones and vitamins that regulate osteoblast formation
as well as storage and release of Ca2+ including parathyroid hormone (PTH), GH, BMP, and vitamin D.
As discussed previously in the 2006 Pb AQCD and 2013 Pb ISA, Pb exposure can alter these pro-
osteoblastic signals which are thought to be involved in the reduction of bone growth and mineralization
seen following Pb exposure. Recent studies show similar alterations in calcitropic and osteoplastic signals
that could be responsible for reduced bone formation (Zhang et al.. 2019; Kupraszewicz and Brzoska.
2013). Together, these data provide plausible indirect pathway by which Pb exposure can regulate skeletal
bone homeostasis.

The pathway for development of osteoarthritis is less well studied. Osteoarthritis results from
erosion of cartilage and articular bone in the joints. Chondrocytes are responsible for matrix deposition
and joint maintenance. Signaling though TGF(3 is thought to be important in proper joint maintenance. A
recent study showed that Pb treatment in rats induced cartilage loss which was associated with loss of
extracellular matrix proteins (Holz et al.. 2012). In the same study, in vitro treatment of chondrocytes
from rat or chicks resulted in reduced markers of TGF(3 signaling and increased markers of matrix
degradation. These data suggest that Pb-induced osteoarthritis could be a result of Pb effects of
chondrocytes and subsequent cartilage degradation.

Teeth do not undergo the same bone turnover processes as skeletal bone and thus Pb incorporated
into the teeth is permanently sequestered. As discussed in the 2013 Pb ISA, dental effects of Pb are
thought to arise from the effects of Pb on enamel producing cells in combination with the incorporation of
Pb into areas of mineralization (U.S. EPA. 2013). Previously evaluated studies showed decrease cell
proliferation, procollagen type I production, intracellular protein, and osteocalcin in human dental pulp
cell cultures (U.S. EPA. 2013). A recent study supports the link between Pb exposure and dental effects
by showing reduced molar diameter and increased dental cracks in the offspring of rats treated with Pb
during either gestation or lactation (Chen et al.. 2012). Together Pb-induced dental effects could result
from effects on dental pulp cells resulting in reduced matrix proteins.

The toxicologic data support Pb-induced alterations in multiple aspects of bone, teeth, and joint
maintenance. For skeletal bones, shift in the balance between bone building osteoblasts and bone
resorbing osteoclasts could be responsible for delayed bone growth and increased bone degeneration seen
in epidemiologic studies. In teeth and joints, Pb appears to suppress the synthesis of cellular matrix
proteins important for joint maintenance and enamel formation which could plausibly contribute to the
osteoarthritic and dental effects seen in some epidemiology studies.

External Review Draft

9-43

DRAFT: Do not cite or quote

-------
Osteoporosis

Exposure

Depressed cell growth
and mineralization

r ^

Altered
osteoblast/osteoclast
balance

Increased
falls/fractures

Osteoarthritis

Depressed protein
synthesis

Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to Pb exposure, and the arrows indicate a proposed relationship
between those effects. Solid arrows denote evidence of essentiality as provided, for example, by an inhibitor of the pathway used in an experimental study involving Pb exposure.
Dotted arrows denote a possible relationship between effects. Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes,
groupings of boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color coded (white, exposure; green, initial effect;
blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here, population-level effects generally reflect results of epidemiologic studies. When there are
gaps in the evidence, there are complementary gaps in the figure and the accompanying text below, iS.7.2 discusses the structure of the biological plausibility sections and the role of
biological plausibility in contributing to the weight-of-evidence analysis used in the 2022 Pb ISA.

Figure 9-2

Potential biological pathways for musculoskeletal effects following exposure to Pb.

External Review Draft

9-44

DRAFT: Do not cite or quote

-------
1

9.5.6

Summary and Causality Determination

The 2013 Pb ISA concluded that evidence was "sufficient to conclude that a causal relationship
is likely to exist between Pb exposure and effects on bone and teeth" (U.S. EPA. 2013). This causality
determination was based on a small body of epidemiologic evidence showing associations between Pb
biomarker levels and effects on bones after adjusting for potential confounding by age and SES-related
factors, as well as strong toxicological evidence that reported effects on bone in animals following Pb
exposure. Specifically, a few epidemiologic studies indicated an association between higher Pb biomarker
levels and lower bone density in adults. A prospective study of older women provided evidence that
higher BLLs (>4 (ig/dL versus <3 (ig/dL) were associated with greater risk of falls and osteoporosis-
related fractures, as well as lower bone density measured after 2-4 years (Khalil et al.. 2009). This finding
was supported by cross-sectional associations between higher BLLs and lower BMD (Campbell and
Auinger. 2007) and biochemical biomarkers of higher bone turnover (Nelson etal.. 2011a; Machida et al..
2009) in adults. In evaluating the cross-sectional epidemiologic evidence, it is difficult to determine
whether an increase in BLLs results from lower bone density or from higher bone turnover, and whether
either of these effects lead to a greater release of Pb from bone into the bloodstream. Exposure of animals
to Pb during gestation and the immediate postnatal period was reported to significantly depress early bone
growth with the effects showing concentration-dependent trends. Systemic effects of Pb exposure
included disruption in bone mineralization during growth, alteration in bone cell differentiation and
function due to alterations in plasma levels of growth hormones and calcitropic hormones such as 1,25-
[OH]2D3 and impact on Ca2+- binding proteins and increases in Ca2+ and phosphorus concentrations in
the bloodstream. Bone cell cultures exposed to Pb had altered vitamin D-stimulated production of
osteocalcin accompanied by inhibited secretion of bone-related proteins such as osteonectin and collagen.
In addition, Pb exposure caused suppression in bone cell proliferation most likely due to interference from
factors such as GH, EGF, transforming growth factor-beta 1 (TGF-(31), and PTHrP.

In addition to effects on bone, epidemiologic and toxicological studies evaluated in the 2013 ISA
provided evidence of Pb-related effects on teeth. A limited number of epidemiologic studies reported
associations between increased BLLs and increased dental caries in children (Moss et al.. 1999) and
periodontitis in adults (Saraiva et al.. 2007). Additionally, higher patella and tibia Pb levels were
associated with tooth loss in men participating in the NAS (Aroraetal.. 2009). This epidemiologic
evidence was based on cross-sectional study design analyses, which precludes conclusions about the
directionality of effects. However, these findings are supported by toxicological evidence in animals for
Pb-induced increases in Pb uptake into teeth; and decreases in cell proliferation, procollagen type I
production, intracellular protein, and osteocalcin in cells exposed to Pb in vitro. Despite evidence for
associations between Pb exposure and effects in bone and teeth at relatively low concurrent BLLs, these
outcomes were most often examined in older adults that have been exposed to higher levels of Pb earlier
in life. Therefore, uncertainty still remains concerning the Pb exposure level, timing, frequency, and
duration that contribute to the observed associations.

External Review Draft

9-45

DRAFT: Do not cite or quote

-------
1

Recent cross-sectional epidemiologic studies continue to support associations between Pb
exposure and effects on bone. The majority of recent studies of osteoporosis or osteopenia were
conducted in female populations or included models stratified by sex to account for sex-specific
difference in osteoporosis and osteopenia incidence. The evaluated studies provide generally consistent
evidence of a positive association between low BLLs (mean/median ranges cross studies: 1.03 to
3.4 (ig/dL) and osteoporosis or osteopenia in women (Li et al.. 2020b; Wang et al.. 2019; Pollack et al..
2013; Cho et al.. 2012; Lee and Kim. 2012). Other studies also observed positive associations in models
including men and women (Lee and Park. 2018; Lim et al.. 2016). but the inferences that can be drawn
from these studies are limited due to the previously noted sex differences in BMD. A few recent cross-
sectional studies also reported associations between low BLLs and symptomatic and radiographic OA in
the knee (Park and Choi. 2019; Nelson etal.. 2011b). These findings were supported by another study
demonstrating associations between BLLs and some biomarkers of joint tissue metabolism, which could
either lead to OA or be indicative of prevalent OA (Nelson et al.. 201 la). These studies of OA represent
an emerging area of research for an endpoint that was not discussed in the 2013 Pb ISA. Recent
epidemiologic evidence is prone to similar uncertainties and limitations identified in the previous ISA.
Notably, the cross-sectional design of these studies does not establish temporality between the exposure
and outcome. This may be particularly relevant for health outcomes that correlate with bone turnover
rates that could lead to higher BLLs. Additionally, although a number of recent studies have been
conducted in adult populations with low BLLs, uncertainty regarding past exposures continues to limit the
characterization of the Pb exposure levels, timing, frequency, and duration that contribute to the observed
associations.

The recent toxicological evidence base for effects on bones is smaller, but consistent with
findings from the 2013 Pb ISA and coherent with recent epidemiologic evidence. Notably, a recent study
reported a reduction in osteoclast activity and a disruption in bone accrual in Pb-exposed animals (Beier
et al.. 2016). This finding, along with similar evidence from previous ISAs and AQCDs, provides support
for a temporal relationship between Pb exposure and effects on bone accrual and bone density that cannot
be established by the available cross-sectional epidemiologic evidence.

In addition to studies of Pb exposure and effects on bone, recent epidemiologic studies have also
explored the relationship between BLLs and effects on teeth. Recent studies in adults focused on the
prevalence of periodontitis, whereas studies in children examined the prevalence or incidence of dental
caries. A group of studies examining overlapping KNHANES cycles observed positive associations
between low BLLs and periodontitis prevalence in adults (Han et al.. 2013; Kim and Lee. 2013; Won et
al.. 2013). including some evidence of a stronger association in men, and persistent associations in models
adjusting for Hg and Cd (Kim and Lee. 2013). Given the largely overlapping study populations, the
observed results should not be interpreted as independent evidence of an association. Additionally, the use
of BLLs in adult populations with higher past exposures limits the ability to characterize the Pb exposure
levels, timing, frequency, and duration that contribute to the observed associations. In a prospective birth
cohort study of low- to moderate-income mother-child pairs, increases in prenatal and early childhood

External Review Draft

9-46

DRAFT: Do not cite or quote

-------
1

BLLs were associated with increased risk of dental caries in adolescence (Wu et al.. 2019). The authors
also observed a null association with concurrent BLLs, which suggests that there may be critical windows
of exposure earlier in life. These findings were supported by a few cross-sectional studies that reported
associations between BLLs in early childhood and increased prevalence of dental caries in deciduous
teeth (Kim et al.. 2017; Wiener etal.. 2015). No recent toxicological studies have examined the effects of
Pb exposure on teeth, but as described earlier, previous and recent mechanistic evidence provides
biological plausibility for the observed epidemiologic associations.

In summary, there is an expanded epidemiologic evidence base that continues to demonstrate
associations between BLLs and various musculoskeletal effects after adjusting for potential confounding.
However, the recent epidemiologic evidence does not thoroughly address uncertainties identified in the
previous ISA, including unclear temporality of exposure and outcome resulting from mostly cross-
sectional study designs, and a lack of studies that adequately characterize the Pb exposure levels, timing,
frequency, and duration that contribute to the observed associations. Although there are not many recent
toxicological studies that meet PECOS relevance, the evaluated studies are consistent with a large
evidence base from the previous ISA and AQCD, which provides support for the observed epidemiologic
associations. Overall, the collective evidence is sufficient to conclude that there is likely to be a causal
relationship between Pb exposure and musculoskeletal effects. The key evidence, as it relates to the
causal framework, is summarized in Table 9-2.

External Review Draft

9-47

DRAFT: Do not cite or quote

-------
Table 9-2 Summary of evidence for a likely to be causal relationship between Pb exposure and
musculoskeletal effects.

Rationale for Causality
Determination3

Key Evidence13

Key References'3

Pb Biomarker Levels Associated with
Effects0

Consistent evidence from
epidemiologic studies of
osteoporosis and osteopenia

Evidence from cross-sectional
epidemiologic studies supports
associations between Pb exposure and
osteoporosis or osteopenia in adult
female populations.

Cho et al. (2012)
Wang etal. (2019)
Lee and Kim (2012)
Pollack et al. (2013)
Li et al. (2020b)

Mean/median ranges cross studies:
1.03 to 3.4 [jg/dL

Supporting evidence from
toxicological studies with
relevant exposures
investigating effects on bone

Toxicological evidence is coherent with
epidemiologic evidence and provides
support for a temporal relationship
between Pb exposure and effects on
bone accrual and bone density

Beier et al. (2016)
(U.S. EPA. 2013)
(U.S. EPA. 2006)

Mean range of 20.8 to 49.9 [jg/dL

Consistent evidence from
epidemiologic studies of
dental caries in children

A prospective birth cohort study provides
evidence that increases in prenatal and
early childhood BLLs are associated with
increased risk of dental caries in
adolescence

Wu etal. (2019)

Mean (males, female):
15.48, 15.18 [jg/dL

Supporting cross-sectional evidence of
associations between early childhood
BLLs and dental caries in deciduous
teeth

Kim etal. (2017)
Wiener et al. (2015)

Geometric Mean: 1.53 [jg/dL
Mean NR (28.2% <2 pg/dL; 48.3% 2 to
<5 [jg/dL; 18.4% 5 to <10 pg/dL; 5.1%
>10 pg/dL)

External Review Draft

9-48

DRAFT: Do not cite or quote

-------
Rationale for Causality
Determination3

Key Evidence13

Key References'3

Pb Biomarker Levels Associated with
Effects0

Biological Plausibility	Pb can directly replace Ca2+ in the bone Section 9.5.4.

matrix as well as exert direct effects on
bone cells to alter bone development. Pb
can also alter bone growth and
differentiation signals that can further
disrupt the balance of bone formation
and resorption. Pb has also been shown
to decrease cell proliferation, procollagen
type I production, intracellular protein,
and osteocalcin in human dental pulp cell
cultures.

Preamble to the ISAs (U.S. EPA. 2015).
where applicable, to uncertainties or

BLLs = blood lead levels; Ca2+ = calcium ions; NR = not reported; Pb = lead.

"Based on aspects considered in judgments of causality and weight-of-evidence in causal framework in Table I and Table II of the
'Describes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and,
inconsistencies. References to earlier sections indicate where the full body of evidence is described.

°Describes the Pb biomarker levels at which the evidence is substantiated.

External Review Draft

9-49

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

9.6

Effects on Ocular Health

9.6.1	Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

This section of effects on ocular health focuses on impairments related to the structure of the eye,
including but not limited to cataracts, glaucoma, macular degeneration, and retinal stippling. Studies
examining effects on vision that are related to sensory processing in the central nervous system can be
found in Appendix 3 of this ISA (Sections 3.5.6.2 and 3.6.3.2). The 2013 Pb ISA concluded that because
the studies of effects on ocular health were of insufficient quantity and quality, the overall evidence was
"inadequate to determine a causal relationship between Pb exposure and ocular effects" (U.S. EPA.
2013). There were very few studies evaluated in the 2013 Pb ISA that examined Pb exposure and ocular
effects in humans or animals. Those studies that were reviewed examined disparate outcomes and the
epidemiologic studies lacked rigorous statistical analyses.

9.6.2	Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria
than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological
plausibility of Pb-associated effects on the ocular health, recent studies were only included if they
satisfied all of the components of the following discipline-specific PECOS statements:

1 The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

External Review Draft

9-50

DRAFT: Do not cite or quote

-------
1

Epidemiologic Studies:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Outcome: Effects on ocular health.

Experimental Studies:

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.3'4

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

Outcomes: Ocular effects.

Study design: Controlled exposure studies of animals in vivo.

2 Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM25) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

3 Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

External Review Draft

9-51

DRAFT: Do not cite or quote

-------
1

9.6.3

Epidemiologic Studies on Ocular Health

A limited number of epidemiologic studies evaluated in the 2013 Pb ISA (U.S. EPA. 2013) did
not provide evidence of an association between exposure to Pb and ocular health. A cross-sectional study
of macular degeneration reported higher concentrations of Pb in the retinal tissue of donors with macular
degeneration compared to those without (Erie et al.. 2009). However, the authors did not control for
confounders in this comparison of means. Another study measured BLLs in smokers and non-smokers
with cataracts, but the authors did not make comparisons between exposure to Pb and severity of cataracts
(Mosad et al.. 2010).

Recent studies provide inconsistent evidence of an association between exposure to Pb and ocular
effects. The majority of recent studies evaluating ocular health and Pb exposures are population-based
cross-sectional analyses, which are unable to establish temporality between exposure and outcome.
Additionally, because many of the observed ocular impairments generally occur in older adult populations
who likely had higher past than current Pb exposure, there is uncertainty regarding the Pb exposure level,
duration, frequency, and timing that may contribute to any observed associations. Measures of central
tendency for blood and/or bone Pb levels used in each study, along with other study-specific details,
including study population characteristics and select effect estimates, are highlighted in Table 9-13. An
overview of the recent evidence is provided below.

A limited number of recent studies have evaluated the relationship between levels of Pb in the
blood or bone and glaucoma. The strongest evidence for an association comes from a longitudinal
analysis of the Veterans Affairs NAS, a prospective cohort study of male Veterans (Wang et al.. 2018b).
Wang et al. (2018b) reported that increases in tibia and patella Pb were associated with 28% (95% CI:
-1%, 65%) and 42% (95% CI: 11%, 82%) increases in risk of primary open-angle glaucoma,
respectively. These results are supported by a recent KNHANES mediation analysis that evaluated
intraocular pressure, which is an important risk factor for glaucoma (Park and Choi. 2016). The authors
reported that a 1 (ig/dL increase in blood Pb was associated with a 0.09 mmHg (95% CI: 0.06,
0.12 mmHg) increase in intraocular pressure, after accounting for indirect effects of exposure to Pb
through increases in blood pressure. The estimated total effect (i.e., not controlling for mediation by blood
pressure) for a 1 (ig/dL increase in blood Pb was 0.11 mmHg (standard error not reported). In contrast,
two recent large cross-sectional studies of the KNHANES did not observe an association between BLLs
and glaucoma (Lee et al.. 2016; Lin et al.. 2015). However, potential associations with chronic age-related
diseases, such as glaucoma, may be better evaluated using measurements of Pb in bone, which has a much
longer half-life than in blood and is therefore a better indicator of cumulative exposure.

In addition to studies of glaucoma, there were also a few recent population-based cross-sectional
studies that examined the association between BLLs and age-related macular degeneration (AMD) in
older adults (Hwang et al.. 2015; Park et al.. 2015; Wu et al.. 2014). AMD is a common eye-disorder in
older adults that is caused by retinal damage, resulting in deteriorated central vision. Two recent studies
of the KNHANES provided evidence of an association between BLLs and AMD (Hwang et al.. 2015;

External Review Draft

9-52

DRAFT: Do not cite or quote

-------
1

Park et al.. 2015). Using data from the 2008-2011 cycles of KNHANES, Park et al. (2015) reported a
12% (95% CI: 2%, 23%) increase in the odds of early-stage AMD (i.e., damaged macula with no vision
loss) and a 25% (95% CI: 5%, 50%) increase in the odds of late-stage AMD (i.e., damaged macula with
vision loss) per 1 (ig/dL increase in blood Pb. In a similar study that analyzed one additional year of
KNHANES data (2008-2012), Hwang et al. (2015) similarly observed increasing odds of early-stage
AMD with increasing quintiles of Pb exposure. Notably, in analyses stratified by sex, the observed
associations in the total population appeared to be driven by a much stronger association in women. The
authors also reported associations for late-stage AMD, but the case numbers were so low for each quintile
that the reduced statistical power to detect an association made the results unreliable. In contrast to the
results from the KNHANES studies, Wu et al. (2014) reported null associations between BLLs and AMD
in an analysis of older adults in the 2005-2008 cycles of the U.S. NHANES.

Additional cross-sectional studies examined other ocular health effects for disparate outcomes,
including an NHANES analysis of cataract surgery in older adults (Wang et al.. 2016) and a KNHANES
study of dry eye disease (Jung and Lee. 2019). Both of these studies reported null associations between
BLLs and the ocular health outcome of interest.

9.6.4 Toxicological Studies on Ocular Health

The 2013 Pb ISA (U.S. EPA. 2013) made note of a limited number of animal studies finding Pb-
induced mouse retinal progenitor cell proliferation and neurogenesis, as well as increased opacity of rat
lens after Pb exposure.

Two recent toxicological studies were identified since the 2013 Pb ISA for inclusion in the
present Pb ISA. Perkins et al. (2012) described remodeling of rod and cone synaptic mitochondria in mice
after postnatal exposure to Pb acetate in drinking water (21 (ig/dL BLL at weaning). The observed Pb-
induced changes are consistent with deficits in range of vision. The effect of Pb on rod and cone
mitochondria was mediation by Bcl-xL, a protein that has been implicated in Pb-induced apoptosis. Using
adult rats exposed to Pb acetate in drinking water (1-20 (ig/dL BLL), Shen et al. (2016) found increased
blood-retinal permeability. The authors noted an association between long-term increased vascular
permeability with retinal dysfunction and degeneration.

9.6.5 Summary and Causality Determination

The 2013 Pb ISA concluded that evidence was "inadequate" to determine a causal relationship
between Pb exposure and ocular health effects (U.S. EPA. 2013). This causality determination was based
on an insufficient quantity and quality of studies in the cumulative body of evidence. Although a cross-
sectional epidemiologic study reported higher concentrations of Pb in the retinal tissue of donors with
macular degeneration compared to those without (Erie et al.. 2009). the study did not account for smoking

External Review Draft

9-53

DRAFT: Do not cite or quote

-------
1

status as potential confounder. Toxicological studies were limited in number, but reported Pb-induced
retinal progenitor cell proliferation, retinal electroretinograms, and lens opacity.

Since the completion of the 2013 Pb ISA, there has been an increase in the number of
epidemiologic studies that examine the relationship between Pb exposure and ocular health effects.

Recent epidemiologic studies provide inconsistent evidence of an association between Pb exposure and
ocular health effects. The strongest evidence comes from a prospective cohort study of male Veterans that
reported large, but imprecise associations between bone Pb levels and glaucoma (Wang et al.. 2018b).
These results are supported by a cross-sectional association between BLLs and intraocular pressure,
which is an important risk factor for glaucoma (Park and Choi. 2016). However, additional population-
based cross-sectional studies in the same population reported null associations between BLLs and
glaucoma (Lee et al.. 2016; Lin et al.. 2015). No recent experimental studies examined endpoints related
to glaucoma.

Findings from a limited number of population-based cross-sectional studies of Pb exposure and
AMD were inconsistent across populations - with null results observed in a U.S.-based study and a
positive association in a South Korean-based study. A recent toxicological study reported Pb-induced
increases in blood-retinal permeability, which may lead to increased risk of macular degeneration.

Although the evidence base has expanded since the completion of the previous assessment, the
limited number of studies and the inconsistent results do not provide sufficient information to draw a
conclusion regarding causality. Thus, the evidence remains inadequate to infer the presence or absence
of a causal relationship between exposure to Pb and ocular health effects.

External Review Draft

9-54

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

9.7

Effects on the Respiratory System

9.7.1	Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

The 2013 Pb ISA evaluated studies of respiratory effects related to inflammatory and atopic
diseases (like asthma) separately from effects on lung function, morphology, and respiratory symptoms.
Similarly, in this review, studies evaluating the effect of Pb on asthma are discussed with effects on the
immune system in Appendix 6. This section discusses the effects of Pb on the respiratory system in the
otherwise healthy lung. The 2013 Pb ISA concluded that there was "insufficient quantity and quality of
studies" related to the impacts of Pb on the non-asthmatic lung and the evidence was therefore
"inadequate to determine a causal relationship" (U.S. EPA. 2013). Epidemiologic studies in non-
asthmatics were lacking in number, consistency, and statistical rigor, despite observed associations
between BLLs and respiratory effects in children and asthmatics (Appendix 6). The few respiratory
toxicological studies described previously were in vivo and in vitro studies that administered concentrated
ambient particulate matter, of which Pb was a component. The ability to evaluate the independent effect
of Pb in these studies was limited due to the inability to account for confounding effects of copollutants
and the lack of characterization of Pb particles in the samples. Given the limitations of these studies, the
scope for this review was narrowed to remove toxicological studies that analyzed the health effects of Pb
containing mixtures but lacked a Pb alone treatment group.

9.7.2	Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria

1 The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

External Review Draft

9-55

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological
plausibility of Pb-associated effects on the immune system, recent studies were only included if they
satisfied all of the components of the following discipline-specific PECOS statements:

Epidemiologic Studies:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Exposure: Exposure to Pb1 as indicated by biological measurements of Pb in the body - with a
specific focus on Pb in blood, bone, and teeth; validated environmental indicators of Pb
exposure2; or intervention groups in randomized trials and quasi-experimental studies.

Comparison: Populations, population subgroups, or individuals with relatively higher versus
lower levels of the exposure metric (e.g., per unit or log unit increase in the exposure metric,
or categorical comparisons between different exposure metric quantiles).

Outcome: Effects on the respiratory system.

Study Design: Epidemiologic studies consisting of longitudinal and retrospective cohort studies,
case-control studies, cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest, randomized trials and quasi-experimental studies examining
interventions to reduce exposures.

Experimental Studies:

Population: Laboratory nonhuman mammalian animal species (e.g., mouse, rat, guinea pig,
minipig, rabbit, cat, dog) of any lifestage (including preconception, in utero, lactation,
peripubertal, and adult stages).

1	Recent studies of occupational exposure to Pb were considered insofar as they addressed a topic area of particular
relevance to the NAAQS review (e.g., longitudinal studies designed to examine recent versus historical Pb
exposure).

2	Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM25) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

External Review Draft

9-56

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

Exposure: Oral, inhalation, or intravenous routes administered to a whole animal (in vivo) that
results in a BLL of 30 (ig/dL or below.1,2

Comparators: A concurrent control group exposed to vehicle-only treatment or untreated
control.

Outcomes: Effects on the respiratory system.

Study design: Controlled exposure studies of animals in vivo.

9.7.3 Epidemiologic Studies on the Respiratory System

A limited number of epidemiologic studies evaluated in the 2013 Pb ISA did not provide strong
evidence of an association between BLLs and airway responses in asthma-free populations. Further, these
studies lacked rigorous statistical analysis and included limited consideration of potential confounders. In
panel and time-series epidemiologic studies considering ambient air Pb (measured in PM2 5 or PM10 air
samples), associations were reported between short-term increases in air Pb and decreases in lung
function and increases in respiratory symptoms and asthma hospitalizations in children but not adults.
Despite this evidence for respiratory effects related to air Pb concentrations, the limitations of air Pb
studies - including the limited data on the size distribution of Pb-PM, the uncertain relationships of Pb-
PM10 and Pb-PIVL 5 with BLLs, and the lack of adjustment for other correlated particulate matter (PM)
chemical components - precluded firm conclusions about ambient air Pb-associated respiratory effects.
Recent studies have examined lung function and respiratory symptoms in non-asthmatic children and
adults. While the majority of recent studies utilized cross-sectional designs that are unable to establish
temporality between exposure and outcome, most adjust for a wide range of potential confounders and
examine populations with lower BLLs. In general, recent evidence in children is inconsistent, though
there is some evidence from a prospective cohort study that BLLs are associated with accelerated lung
function decline in adults. Notably, because adult populations likely had higher past than current Pb
exposure, there is uncertainty regarding the Pb exposure level, duration, frequency, and timing that may
contribute to the observed association. Measures of central tendency for blood and/or serum Pb levels
used in each study, along with other study-specific details, including study population characteristics and
select effect estimates, are highlighted in Table 9-15. An overview of the recent evidence, delineated by
lifestage, is provided below.

1	Pb mixture studies are included if they employ an experimental arm that involves exposure to Pb alone.

2	This level represents an order of magnitude above the upper end of the distribution of U.S. young children's BLL.
The 95th percentile of the 2011-2016 NHANES distribution of BLL in children (1-5 years; n= 2,321) is 2.66 (ig/dL
(Eganetal.. 20211 and the proportion of individuals with BLL that exceed this concentration varies depending on
factors including (but not limited to) housing age, geographic region, and a child's age, sex, and nutritional status.

External Review Draft

9-57

DRAFT: Do not cite or quote

-------
1

9.7.3.1

Respiratory Effects in Children

A limited number of recent cross-sectional studies have examined the relationship between BLLs
and pulmonary function or respiratory symptoms in children. Studies conducted in different locations
reported inconsistent evidence of an association between BLLs and pulmonary function. In an analysis of
6- to 17-year-old children participating in the 2011-2012 NHANES survey cycle, Madrigal et al. (2018)
reported modest and imprecise increases in mean forced expiratory volume (FEV1) (41.9 mL [95% CI:
-46.9, 130.6 mL]) and forced vital capacity (FVC) (45.5 mL [95% CI: -49.2, 140.2 mL]) for children
with BLLs in the highest quartile (>0.86 (ig/dL) compared to children with BLLs in the first quartile
(<0.44 (ig/dL). Similar comparisons were null for FEV1 :FVC and forced expiratory flow (FEF)25%-75%.
Notably, while the study population had a very low median BLL (0.56 (.ig/dL). there were small exposure
contrasts between exposure quartiles, which may have limited the statistical power to detect an
association. In contrast with the NHANES analysis, smaller cross-sectional studies conducted in
preschool-aged children in China (Zeng et al.. 2017) and 10- to 15-year-old children in Poland (Little et
al.. 2017) observed limited evidence of associations between BLLs and decreased FVC (Little et al..
2017; Zeng et al.. 2017) or FEV1 (Zeng et al.. 2017). Both studies noted small and imprecise associations
and had small sample sizes. Limited statistical power resulting from a small sample size reduces the
likelihood of detecting a true effect and the likelihood that an observed result reflects a true effect, which
might explain the incongruous results. Additionally, the associations observed by Little et al. (2017) may
have been subject to unmeasured confounding (e.g., by age, SES factors, environmental tobacco smoke),
as the authors only adjusted their regression models for children's heights.

In addition to studies of pulmonary function, a single study examined respiratory symptoms in
children. (Zeng et al.. 2016) reported inconsistent associations between BLLs and respiratory symptoms
in preschool-aged children in China, including some living in a community near an e-waste facility. The
authors compared children with BLLs >5 (ig/dL to those with BLLs <5 (ig/dL and reported that those in
the higher exposure group had decreased odds of parental-reported wheeze and dyspnea, a slight increase
in the odds of parental-reported phlegm, and no perceptible change in parental-reported cough. Caution is
warranted in interpreting results of parental-reported symptoms in locations with known environmental
contamination due to potential over-reporting of symptoms.

9.7.3.2 Respiratory Effects in Adults

A limited number of recent studies have examined the relationship between blood or serum Pb
levels and respiratory effects in adults. There is evidence from a prospective cohort study that BLLs are
associated with accelerated lung function decline in adults, although a large, population-based cross-
sectional study reports conflicting results. All of the studies evaluated in this subsection reported low
levels of blood or serum Pb levels (mean and geometric mean levels <3 (.ig/dL).

External Review Draft

9-58

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

The most compelling evidence of an association between Pb exposure and lung function in adults
comes from a prospective cohort study of adults living adjacent to a large industrial complex in South
Korea (Pak et al.. 2012). The authors reported that BLLs were associated with accelerated lung function
decline, measured as the difference in spirometric measurements taken at baseline and after two-years of
follow-up. Specifically, Pak et al. (2012) noted accelerated decline in FVC (-177 mL [95% CI: -330,
-24]) and FEV1 (-107 mL [95% CI: -215, 1]) per 1 (ig/dL increase in BLL at baseline. Notably, because
adult populations likely had higher past than current Pb exposure, there is uncertainty regarding the Pb
exposure level, duration, frequency, and timing that may contribute to the observed association. In
contrast to results from Pak et al. (2012). a recent cross-sectional study of 2008-2012 KNHANES
participants with low BLLs observed null associations between BLLs and FVC and FEV1 in adults
(Leem et al.. 2015).

Leem et al. (2015) also examined obstructive lung function (FEV1/FVC <0.7) in the same
population and observed a null association with BLLs. In a similar recent analysis of a large population-
based health survey (NHANES), (Rokadia and Agarwal. 2013) reported a large, but imprecise increase in
the odds of obstructive lung function (94% [95%: 10%, 342%] per 1 (ig/dL increase in serum Pb levels)
that appears to be driven by an association in participants with moderate to severe obstructive lung
function (349% [95%: 70%, 715%] per 1 (ig/dL increase in serum Pb levels). The observed associations
were similar in analyses stratified by smoking status, although the associations in non-smokers were even
less precise due to a smaller number of cases.

9.7.4 Toxicological Studies on the Respiratory System

The 2013 ISA evaluated a limited number of studies investigating the effects of ambient
particulate mixtures of which Pb was a component. The effects directly attributable to Pb were not able to
be distinguished from other confounding mixture components. The PECOS criteria used in this ISA to
identify new respiratory toxicological studies focused on identifying studies that studied Pb exposure
alone. One study reviewed in the 2013 Pb ISA showed that injection of Pb acetate resulted in histologic
signs of damage and inflammation in the lung although uncertainty regarding the biological relevance of
Pb injection remained A few new experimental studies were identified that investigated the effect of
inhaled Pb and met our PECOS criteria (Table 9-9). The studies, all published by the same group,
assessed the localization and clearance of inhaled ultrafine (>100 nm in diameter) Pb particles and the
corresponding effect on lung (and secondary organ) tissue structure. These studies involved 2-11 weeks
of exposures (24 hours/day, 7 days/week) to inhaled Pb nanoparticles after which the investigators
analyzed lung histology and markers of lung damage. Exposure of female mice to roughly 106
particles/cm3 lead oxide (PbO) particles for 6 weeks led to a mean BLL of 132 ng/g (-13.922 (ig/dL) and
corresponded to histological signs of lung damage including alveolar septal wall thickening, emphysema,
perivascular infiltration of immune cells, and signs of thrombosis (Dumkova et al.. 2017). Exposure to a
higher concertation of PbO (2.23 x 106 particles/cm3) for 3 days, 2, 6, and 11 weeks led to BLLs ranging

External Review Draft

9-59

DRAFT: Do not cite or quote

-------
1

from 10.4 (ig/dL at 2 weeks up to 17.4 (ig/dL after 11 weeks of exposure. The BLL at 3 days was not
reported. Histological signs of cellular infiltration and alveolar septal wall thickening was observed after
6 and 11 weeks of PbO exposure along with signs of macrophage proliferation (PCNA-staining)
(Dumkova et al.. 2020b). These effects were not reported for the two-week exposure or an acute 3-day
exposure to PbO. Despite increased signs of lung inflammation, signs of fibrosis and apoptosis were not
observed. Interestingly, a 5-week recovery period with no PbO exposure following 6 weeks of PbO
exposure was able to reduce both the lung Pb concertation and partially recover the histopathological
signs of inflammation seen at 6 weeks of PbO (Dumkova et al.. 2020b).

In a separate experiment, a similar procedure as Dumkova et al. (2020b) was followed using more
soluble Pb(N03)2nanoparticles in place of PbO. Mice were exposed to Pb(NOs)2 particles for either 3
days, 2 weeks, 6 weeks, or 11 weeks and a separate recovery group that was exposed to Pb(NOs)2 for
6 weeks and then filtered air for 5 weeks (Dumkova et al.. 2020a). Similar to the results with PbO,
Pb(NOs)2 exposure showed an increase in histological signs of inflammation and lung damage.
Histological effects with Pb(NOs)2 particle exposure were seen starting at 2 weeks of exposure and did
not completely resolved in the recovery group. Exposure to Pb(NOs)2 reduced the number of lung
macrophages (CD68 positive stained cells) in the lung tissue which corresponded to an increase in
neutrophils (Myeloperoxidase positive cells) and mastocytes (Toluidine blue staining). Similar to the
findings with PbO, a 5-week recovery period with no Pb(NOs)2 exposure following 6 weeks of Pb(NOs)2
exposure was able to reduce both the lung Pb concertation and partially recover the histopathological
signs of inflammation. While macrophage number was partially restored after a 5-week recovery period,
the level of mastocytes remained elevated. Lung mRNA for inflammatory genes like IL-1B, IL-la, and
tumor necrosis factor-a were largely unchanged however RNA levels of NF-kB and IL6 were suppressed
after 3 days and 11 weeks of Pb(NOs)2 suggesting that Pb(NOs)2 dysregulates the inflammatory response
in the lung. While the data presented in these studies are mostly qualitative, it provides some preliminary
evidence of respiratory effects from inhalation of either Pb(NOs)2 or PbO nanoparticles.

9.7.5 Summary and Causality Determination

The effects of Pb on asthma incidence and host defense, which includes data related to host
response to lung infection, are analyzed in the context of allergic disease and immune suppression
(Section 6.7.1 and Section 6.7.2).

The 2013 Pb ISA determined that the evidence for respiratory effects was "inadequate to
determine a causal relationship between Pb exposure and respiratory effects in populations without
asthma." This determination was based on inconsistent findings among studies and the limited quantity
and quality of both epidemiologic and experimental toxicologic evidence of respiratory effects. While
there was some epidemiologic evidence of an association between short-term increases in ambient air Pb
and decreases in lung function, these studies were not informative to the causality determination due to

External Review Draft

9-60

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

notable uncertainties regarding the size distribution of ambient air Pb, the relationship between ambient
air Pb and BLLs, and the confounding effects of co-occurring pollutants.

Evidence evaluated in the 2013 Pb ISA showed inconsistent relationships between BLLs and
bronchial responsiveness and lung function. Results from recent epidemiologic studies of the effect of
blood Pb on lung function and respiratory symptoms in children remain inconsistent (Section 9.7.3.1). In
adults, a new prospective cohort study provides evidence of accelerated lung function decline in those
with higher BLLs (Pak et al.. 2012). however the relationship between lung function decrements and
BLLs is inconsistent in a few recent cross-sectional analyses (Section 9.7.3.2). This lack of consistency in
the epidemiologic literature is compounded by uncertainty related to exposure assessment and relative
lack of adjustment for correlated air pollutants. Toxicological data in the 2013 ISA was mostly limited to
studies of concentrated ambient PM of which Pb was a component within a mixture of pollutants, leaving
uncertainty for the role of Pb in the observed effects. New toxicological studies evaluating inhalation of
Pb particles are limited in number but do provide evidence of gross histologic signs of transient
inflammation and lung damage; however, these data are largely qualitative and the impact of these
changes on lung function are unknown. Uncertainty still remains about the relative size distribution of Pb
particles in ambient air and thus how well experimental generation of Pb particles reflects ambient
concentrations and particle size distribution. Given the lack of consistency across a small body of
epidemiologic evidence and uncertainty in the direct relevance of a limited number of toxicological
results to human lung function, the evidence is not sufficient to draw a conclusion regarding causality.
Thus, the cumulative body of evidence is inadequate to infer the presence or absence of a causal
relationship between Pb exposure and respiratory effects in populations without asthma.

9.8 Mortality

9.8.1 Introduction, Summary of the 2013 ISA, and Scope of the Current
Review

In the 2013 Pb ISA (U.S. EPA. 2013). the strongest evidence for Pb-associated mortality was
from studies examining cardiovascular mortality. The evidence did not provide strong support for Pb-
associated mortality other than through cardiovascular pathways, and very few studies examined total
(nonaccidental) mortality. For these reasons, the 2013 Pb ISA evaluated studies of all-cause mortality
together with studies examining cardiovascular mortality, and these studies were all included within the
cardiovascular disease chapter. Although this evidence contributed to the "causal relationship" between
Pb exposure and coronary heart disease, there were no distinct causality determinations for total or cause-
specific mortality. In this ISA, the strongest evidence for Pb-associated cause-specific mortality continues
to come from studies of cardiovascular mortality. However, additional studies examining total non-
accidental mortality have become available since the last ISA, and this section discusses and evaluates

External Review Draft

9-61

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

those studies. Studies that examine cardiovascular-related mortality or other cause-specific mortality are
discussed in detail within the appropriate outcome-specific appendices (e.g., cardiovascular disease
(CVD)-related mortality is discussed in Appendix 4) and are briefly summarized in this section.

9.8.2 Scope

The scope of this section is defined by PECOS statements. The PECOS statement defines the
objectives of the review and establishes study inclusion criteria thereby facilitating identification of the
most relevant literature to inform the Pb ISA.1 In order to identify the most relevant literature, the body of
evidence from the 2013 Pb ISA was considered in the development of the PECOS statements for this
Appendix. Specifically, well-established areas of research; gaps in the literature; and inherent
uncertainties in specific populations, exposure metrics, comparison groups, and study designs identified
in the 2013 Pb ISA inform the scope of this Appendix. The 2013 Pb ISA used different inclusion criteria
than the current ISA, and the studies referenced therein often do not meet the current PECOS criteria
(e.g., due to higher or unreported biomarker levels). Studies included in the 2013 Pb ISA, including many
that do not meet the current PECOS criteria, are discussed in this appendix to establish the state of the
evidence prior to this assessment. Except for supporting evidence used to demonstrate the biological
plausibility of Pb-associated effects on mortality, recent studies were only included if they satisfied all the
components of the following PECOS statements:

Population: Any human population, including specific populations or lifestages that might be at
increased risk of a health effect.

Exposure: Exposure to Pb2 as indicated by biological measurements of Pb in the body - with a
specific focus on Pb in blood, bone, and teeth; validated environmental indicators of Pb
exposure3; or intervention groups in randomized trials and quasi-experimental studies.

1	The following types of publications are generally considered to fall outside the scope and are not included in the
ISA: review articles (which typically present summaries or interpretations of existing studies rather than bringing
forward new information in the form of original research or new analyses), Pb poisoning studies or clinical reports
(e.g., involving accidental exposures to very high amounts of Pb described in clinical reports that may be extremely
unlikely to be experienced under ambient air exposure conditions), and risk or benefits analyses (e.g., that apply
concentration-response functions or effect estimates to exposure estimates for differing cases).

2	Recent studies of occupational exposure to Pb were considered insofar as they addressed a topic area of particular
relevance to the NAAQS review (e.g., longitudinal studies designed to examine recent versus historical Pb
exposure).

3	Studies that estimate Pb exposure by measuring Pb concentrations in particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |im3 (PMio) and particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im3 (PM2 5) ambient air samples are only considered for inclusion if they also
include a relevant biomarker of exposure. Given that size distribution data for Pb-PM are fairly limited, it is difficult
to assess the representativeness of these concentrations to population exposure [Section 2.5.3 (U.S. EPA. 201311.
Moreover, data illustrating the relationships of Pb-PMio and Pb-PM2 5 with BLLs are lacking.

External Review Draft

9-62

DRAFT: Do not cite or quote

-------
1

Outcome: Mortality.

9.8.3 Total (non-Accidental) Mortality

The 2013 Pb ISA (U.S. EPA. 2013) evaluated a small number of studies that examined the
association between biomarkers of Pb exposure and all-cause mortality. Overall, these studies reported
consistently positive associations between Pb biomarkers and all-cause mortality. Specifically, Lustberg
and Silbergeld (2002) indicated an increased risk of all-cause mortality when comparing the highest
tertiles of BLLs (20-29 (ig/dL) to the lowest (<10 (ig/dL). Lustberg and Silbergeld (2002) conducted this
analysis among NHANES II cohort, which had high BLLs (mean 14 (.ig/dL). Additionally, Schober et al.
(2006) and Menke et al. (2006) both evaluated the NHANES III cohort, which had an overall lower BLL
(mean: 2.6 |ig/dL). and still identified a positive association between BLLs and all-cause mortality
(Figure 9-1). Notably, both NHANES cohorts included adult study populations with higher past than
recent Pb exposures, making it difficult to characterize the specific timing, duration, frequency, and level
of Pb exposure that contributed to the observed associations. Recent evidence continues to support the
association between Pb biomarkers and all-cause mortality. Study-specific details, including biomarker
Pb levels, study population characteristics, confounders, and select results from these studies, are
highlighted in Figure 9-3 and Table 9-17. Studies in Figure 9-3 are standardized to be interpreted as the
risk of all-cause mortality associated with a 1 (ig/dL increase in BLL. Study details in Table 9-10 include
standardized results as well as results that could not be standardized based on the information provided in
each paper. An overview of the recent evidence is provided below.

External Review Draft

9-63

DRAFT: Do not cite or quote

-------
Reference	Stwty Population PMistrtoititMi

Menke et al, 2005 NHANES III Adults £20 Mean: 2.58

Pt> measurement -Years of
year	foBow-up

Lanphear et al, 2018 NHANES III Adults * 20

Schober et al, 2006 NHANES III Adults £40

Geometric Mean: 2.71
Geometric SE: 1.31

Median
T1 (2.6)

T2 (6.3)

T3 (11.8)

Median

van Bemmeletal, 2011 NHANES III Adults £40 <5ugML2.6

£ 5 ugML7.5

Duan etal, 2020*

NHANES Adults £ 20

Median flQRJ
1.49 (0.93,2.31)

1988-1994

1988-1994

1988-1994

1988-1994

1999-2014

12

19

7.1

all cause

all cause

all cause

7.5-7.8 all cause

all cause ALAD GG
all cause ALAD CG/GG

all cause

1.00	120	1.40

Hazard Ratio (95% CI) per 1 ug/dL increase In blood Pb

ALAD GG and ALAD CG/GG = variants of 5-aminolevulinic acid dehydratase, T1 = Tertile 1, T2 = Tertile 3, T4 = Tertile 4, NHANES = National Health and Nutrition Examination
Survey.

Note: Red text: Studies published since the 2013 Pb ISA; Black text: Studies included in the 2013 Pb ISA.

Effect estimates are standardized to a 1 |jg/dL increase in blood Pb. If the Pb biomarker is log-transformed, effect estimates are standardized to the specified unit increase for the
10th—90th percentile interval of the biomarker level. Effect estimates are assumed to be linear within the evaluated interval.

*Study estimated relative risk.

Figure 9-3 Effect estimates for associations of blood Pb with all-cause mortality.

External Review Draft

9-64

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

In a recent extended analysis of the NHANES III cohort, Lanphear et al. (2018) increased the
average follow-up time of the Menke et al. (2006) analysis by over 7 years (from 12 to -19 years),
resulting in a substantial increase in the number of total deaths observed (4,222 versus 1,661). Lanphear
et al. (2018) reported that a 1 (ig/dL increase in BLL was associated with a hazard ratio (HR) of 1.06
[95% CI: 1.03, 1.09]) for all-cause mortality. The authors also calculated the population attributable
fraction for both all-cause and cardiovascular mortality, to estimate the proportional reduction in mortality
that would be expected if BLLs in those >20 were reduced to 1 (ig/dL. Lanphear et al. (2018) estimated
that the population attributable fraction for all-cause mortality was 18% (95% CI: 10.9-26.1), while the
population attributable fraction for cardiovascular mortality was 28.7% (95% CI: 15.5, 39.5). Therefore,
given the proportion of all-cause mortality attributable to cardiovascular causes (both in this study [~38%]
and nationally [-33%; NHLBI, 2017, 3980932}]), while CVD mortality is likely strongly influencing a
large proportion of the all-cause mortality signal, it does not account for all of it. The authors also used a
five-knot restricted cubic spline analysis to assess potential non-linearities and observed a generally
sigmoidal concentration-response (C-R) relationship between BLLs and all-cause mortality, with some
attenuation of the C-R relationship below 2.5 (ig/dL (Figure 9-4). The general shape of the C-R
relationship is consistent with previous results from Menke et al. (2006).

4~\



0	2.5	5	7.5	10

Concentrations of lead in blood |ig/dL

Note: Restricted cubic spline (5 knots) (red line) and adjusted HRs (black line) with 95% CI's (hatched lines) for all-cause mortality.
Source: Adapted from Lanphear et al. (2018).

Figure 9-4 Dose-response relationship between blood Pb levels and all-
cause mortality.

Other recent studies also evaluated the relationship between blood Pb and total mortality using
NHANES data. Using NHANES III, van Bemmel et al. (2011) estimated an increased association
between BLLs and all-cause mortality (HR: 1.04 [95% CI: 0.98, 1.10]). In addition, van Bemmel et al.

External Review Draft

9-65

DRAFT: Do not cite or quote

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

(2011) also evaluated this relationship by polymorphisms in 5-aminolevulinic acid dehydratase (ALAD).
A critical mechanism of Pb toxicity is its ability to interact and inhibit key enzymes, such as ALAD, in
the heme biosynthesis pathway. This study evaluated associations between BLLs, and mortality stratified
by ALAD variant (ALADGG [more common genotype] or ALADCG/GG). However, there was little
difference between the estimates generated when stratified (ALADGG HR: 1.03 [95% CI:0.98, 1.08],
ALADCG/GG HR: 1.09 [95% CI:0.93, 1.28]), when comparing BLLs >5 (ig/dL to levels <5 (ig/dL.

Using more recent NHANES cycles (1999-2014), Duan et al. (2020) also reported a positive association
between blood Pb and all-cause mortality (RR: 1.39 [95% CI: 128, 1.51]). In a similar analysis using
recent KNHANES cycles (2007-2015), Bvun et al. (2020) evaluated the association between BLLs and
total (nonaccidental) mortality using KNHANES (2007-2015) baseline data, and mortality data linked
through 2018. Overall, there were positive associations between increasing tertiles of blood Pb exposure
and all-cause mortality. Compared to the first tertile of BLLs (<1.91 |ig/dL). the HR for all-cause
mortality was 2.02 (95% CI: 1.20, 3.40) for the second tertile (1.91-2.71 (ig/dL) and 1.91 (95% CI: 1.13,
3.23) for the third tertile (>2.71 (ig/dL).

In addition to studies using nationally representative survey data, a recent study by Hollingsworth
and Rudik (2021) implemented a quasi-experimental design to examine the effect of the phase out of
leaded gasoline in automotive racing on mortality rates in older adults. Comparing time periods prior to
and after the phaseout of leaded gasoline in professional racing series (i.e., the National Association for
Stock Car Auto Racing [NASCAR] and the Automobile Racing Club of America [ARCA]), the authors
used a difference-in-differences technique to estimate county-level changes in air Pb concentrations,
elevated BLL prevalence among children, and mortality rates in race counties and counties bordering race
counties relative to control counties. A detailed discussion of results for air Pb concentrations and BLLs is
presented in Section 2.4.1. In short, there were substantial declines in both air Pb concentrations and the
prevalence of children with elevated BLLs associated with the phaseout of leaded gasoline. The authors
also reported significant declines in mortality rates over this same period. Specifically, in the period
following de-leading of gasoline, there was an estimated decline in annual age-standardized all-cause
mortality rates of 91 deaths per 100,000 in race counties and 38 deaths per 100,000 in border counties.
Similar to the exposure results, the mortality estimates appear to demonstrate a distance gradient.
Although this analysis includes county-level data, the difference-in-difference approach controls for
spatially varying confounders by estimating the difference in mortality rates in adjoining years in the
same county and controls for temporally varying confounders by assessing the difference of those
differences between locations. The authors additionally adjust for potential confounders that may vary
spatially and temporally (e.g., unemployment rate and quantity of Toxic Release Inventory [TRI] lead
emissions). Hollingsworth and Rudik (2021) did not adjust for potential copollutant exposures, but
provide evidence that there is no differential effect of leaded and unleaded races on other copollutant
concentrations (i.e., CO, VOCs, PMio, PM2.5, NO2, and O3) in the weeks leading up to and following the
race. However, because the mortality rates are an annual measure, there is remaining uncertainty
regarding potential differential trends in the long-term average of other pollutants that could be correlated
with the phaseout of leaded gasoline in NASCAR and ARCA.

External Review Draft

9-66

DRAFT: Do not cite or quote

-------
1

Since Pb has been identified as being associated with renal insufficiency, previous studies have
further assessed if Pb accumulates in patients with end-stage renal disease (ERSD). In a recent
prospective cohort study in Taiwan, Lin etal. (2011) followed study subjects on maintenance
hemodialysis for a period of 18 months. Overall, subjects included in the study had higher BLLs (mean:
11.5 (ig/dL) than the general Taiwanese population (mean: 7.7 (ig/dL). It is suspected that hemodialysis
patients may experience higher BLLs since their kidneys may no longer be able to excrete Pb from the
body due to a total loss of renal function (Appendix 5). Among this group, there was a strong but
imprecise association between BLLs and all-cause mortality when comparing those in the second tertile
of BLLs (8.51-12.64 (ig/dL) to those in the first tertile of BLLs (<8.51 (ig/dL) (HR: 2.69 [95% CI: 0.47,
3.44]). This effect was higher in magnitude, but even more imprecise among those in the third tertile of
BLLs (>12.64 (ig/dL) (HR: 4.70 [95% CI: 1.92, 11.49]), compared with the first tertile of blood Pb. The
imprecise effect estimates in this analysis are likely due to a combination of the relatively small sample
size and short follow-up period, leading to a small number of deaths included in the analysis. The small
number of cases reduces statistical power, as well as the likelihood that an observed result reflects a true
effect.

In contrast to the generally consistent evidence of an association between BLLs and all-cause
mortality, a small Canadian study evaluating several trace metals observed a null association between all-
cause mortality and BLLs among hemodialysis patients (>18 years of age) (Tonelli et al.. 2018). Patients
in this cohort had relatively low BLLs (1st decile: 0.06 (ig/dL, 10th decile 1.74 |ig/dL). and there was no
observed relationship between BLLs and all-cause mortality when comparing the highest to the lowest
decile. The authors only presented quantitative results for statistically significant associations, so it is
unclear whether there was any evidence of a non-statistically significant association. Additionally, Tonelli
et al. (2018) was likely underpowered to detect a HR in the range reported in other studies of BLLs and
all-cause mortality (Figure 9-4).

9.8.4 Cause-Specific Mortality

The mortality studies available for review in the 2013 Pb ISA focused primarily on
cardiovascular mortality, and consistently reported positive associations with overall cardiovascular
mortality and cause-specific cardiovascular mortality. Recent studies also evaluate cardiovascular
mortality in addition to other cause-specific mortality outcomes.

Recent analyses further indicate a positive association between Pb exposure and cardiovascular
mortality and are further described in Section 4.10. In summary, there were several studies using
nationally representative data with low BLLs (mean <2 (ig/dL) that consistently reported increased
associations between biomarkers of Pb exposure and cardiovascular mortality. However, these
populations were largely similar (mostly from NHANES III or other more recent NHANES cycles) and

External Review Draft

9-67

DRAFT: Do not cite or quote

-------
1

still include individuals with sizeable historic exposures to Pb. For specific causes of CVD mortality (e.g.,
myocardial infarction (MI), ischemic heart disease (IHD), stroke), the measures of association were
higher in magnitude but were less precise (i.e., wider 95% CIs), likely due to the smaller number of
cause-specific cardiovascular-related deaths. Additionally, in the quasi-experimental study discussed in
Section 9.8.3, deleading of racing gas led to declines in county-level cardiovascular mortality rates
(Hollingsworth and Rudik. 2021). This evidence helps to strengthen the evidence base indicating an
association between biomarkers of Pb exposure and increased risk of cardiovascular mortality.

Several recent studies also evaluated the relationship between Pb exposure biomarkers and cancer
mortality, as described in Section 10.4. In summary, there were a limited number of studies evaluating Pb
biomarkers of exposure and overall cancer mortality. Most studies relied on nationally representative data
and yielded inconsistent but mostly null associations between Pb exposure and cancer mortality.

However, the follow-up period in many of these analyses was short (<11 years), with a small number of
cancer deaths and a lack of control of some potential influential confounders, such as comorbidities and
BMI.

Additionally, some studies evaluated alternative cause-specific mortality outcomes. A cohort
study analyzed data from five NHANES cycles (1999-2008) and reported a positive, but imprecise
association between blood Pb and Alzheimer's disease (AD) mortality rSection 3.5.4; (Horton et al..
2019)1. The imprecise effect estimate is likely due to the small number of AD mortality cases (n = 81)
that resulted from AD mortality being determined by the listing of the immediate cause of death rather
than the underlying cause of death. Additionally, Linetal. (2011) prospectively evaluated subjects on
maintenance hemodialysis for a period of 18 months and evaluated infection-caused mortality. Among
this group there was an imprecise increase in mortality (HR: 5.35 [95% CI: 1.38, 20.83]) in the highest
tertile (>12.64 (ig/dL) compared to the lowest tertile (<8.51 (.ig/dL). This association persisted (HR: 4.72
[95% CI: 1.27, 17.54]) even after correction for hemoglobin (dividing BLL by hemoglobin
concentration). Finally, a quasi-experimental reported a decrease in county-level respiratory mortality
rates in association with the phase out of leaded gasoline in automotive racing (Hollingsworth and Rudik.
2021).

9.8.5 Biological Plausibility

In evaluating the biological plausibility of reported associations between Pb exposure and total
non-accidental mortality, this section considers the biological evidence supporting health outcomes likely
to contribute to total mortality. As summarized above, studies consistently report positive associations
between Pb exposures and cardiovascular-related mortality, with much more limited evidence for
associations with other causes of mortality. Overall, cardiovascular mortality is the most common
contributor to total non-accidental mortality (i.e., accounting for about 33% of total mortality) (NHLBI.
2017). As it pertains to Pb exposure, the available evidence provides strong support for Pb-associated

External Review Draft

9-68

DRAFT: Do not cite or quote

-------
1

cardiovascular effects and supports a continuum of effects leading to cardiovascular mortality, as
described further in Appendix 4. Direct evidence for cardiovascular effects following Pb exposures comes
from numerous animal toxicological studies, and there is coherence between these animal studies and
epidemiologic studies that report associations with some of the same cardiovascular outcomes (e.g.,
increased blood pressure, changes in cardiac electrophysiology). Animal studies additionally support the
biological plausibility of the consistent epidemiologic associations reported between body Pb
concentrations and cardiovascular outcomes such as hypertension and cardiovascular mortality. Section
4.10 characterizes the strong evidence indicating the mechanisms by which exposure to Pb could
plausibly progress from initial events to endpoints relevant to the cardiovascular system, such as
hypertension, exacerbation of IHD, and potential MI or stroke. In particular, exposures to Pb can result in
oxidative stress and systemic inflammation, which could potentially lead to impaired vascular function, a
pro-atherosclerotic environment, and increases in blood pressure. There is animal toxicological evidence
demonstrating all of these effects following exposure to Pb (Section 4.8). Atherosclerosis and increased
blood pressure can then set the stage for an MI or stroke that could result in mortality. Thus, the
progression demonstrated in the available evidence for cardiovascular morbidity supports potential
biological pathways by which Pb exposure could result in cardiovascular mortality.

The current evidence strongly supports a plausible relationship between Pb exposure and
cardiovascular mortality. Additionally, Pb may act on other biological pathways leading to death. There is
some limited evidence that BLLs are associated with other causes of mortality, including AD and
infection. The strongest evidence for biologically supported pathways leading to neurodegenerative
disease include the effect of Pb on cellular protein function and subsequent initiation of oxidative stress-
and inflammation-mediated pathways (Section 3.3). AD, specifically, has been linked with increased
markers of neuroinflammation. Studies with exposure of postweaning animals to Pb have shown
increased inflammation associated with AD markers, as well as inhibition of AD markers following
postexposure treatment with anti-inflammatory and antioxidative molecules. Regarding infection-related
mortality, biological plausibility for the observed association is provided by toxicological and
epidemiologic studies demonstrating (1) skewing of T cell populations, promoting Th2 cell formation and
cytokine production, (2) decreased IFN-y production, (3) decrements in macrophage function, (4)
production of inflammatory mediators, and (5) disruption of the microbiome, all of which could lead to
immunosuppression (Section 6.6.1).

9.8.6 Summary and Causality Determination

The 2013 Pb ISA did not make a causality determination regarding the relationship between Pb
exposure and total (nonaccidental) mortality, but these studies did support the causality determinations
made within the cardiovascular disease chapter. The evidence available at the time of the last review was
limited but reported consistently positive associations between Pb biomarkers and all-cause mortality
(Menke et al.. 2006; Schober et al.. 2006). These results were additionally supported by consistent

External Review Draft

9-69

DRAFT: Do not cite or quote

-------
1

positive associations between BLLs and overall cardiovascular mortality (Section 4.10) as well as cause-
specific cardiovascular mortality (e.g., MI, IHD, stroke)). Menke et al. (2006) examined the shape of the
C-R relationship between BLLs and all-cause mortality using quadratic spline models, which generally
appeared to support a linear, no-threshold relationship, although the HRs were somewhat attenuated at
BLLs <2.5 (ig/dL. Notably, the majority of mortality studies analyzed participants from NHANES
cohorts, either NHANES II or NHANES III, so while the results are consistent, they do not represent a
range of independent study populations. Additionally, while some of the studies evaluated in the 2013 Pb
ISA examined populations with low mean BLLs (<3 (.ig/dL). study participants were born prior to the
phase-out of leaded gasoline and therefore likely had much higher past Pb exposures, making it difficult
to characterize the specific timing, duration, frequency, and level of Pb exposure that contributed to the
observed associations.

Prospective cohort studies evaluated since the completion of the 2013 Pb ISA continue to provide
consistent evidence of positive associations between Pb exposure and total (nonaccidental) mortality.
Many recent analyses further evaluated the association between BLLs and the risk of mortality using
NHANES cohorts linked to mortality databases, including an extended analysis of the NHANES III
cohort with additional years of follow-up (Lanphear et al.. 2018) and analyses of more recent NHANES
cycles (Bvun et al.. 2020; Duan et al.. 2020; van Bemmel et al.. 2011). In addition to NHANES analyses,
another analysis of participants from a nationally representative survey [KNHANES; (Bvun et al.. 2020)1
and a smaller prospective cohort study of hemodialysis patients (Linetal.. 2011) provide evidence of an
association between BLLs and total (non-accidental) mortality. These findings are supported by a quasi-
experimental study that reported a decline in county-level all-cause mortality rates following the phase
out of leaded gasoline in automotive racing (Hollingsw orth and Rudik. 2021). Recent studies continue to
include populations with low mean blood Pb concentrations, but do not address potentially large
differences in past versus current exposures. Thus, there is remaining uncertainty as to the specific timing,
duration, frequency, and level of Pb exposure that contributed to the observed associations. The observed
associations between BLLs and total mortality are large in magnitude (Figure 9-3), though uncertainty in
the levels of Pb exposure that contributed to the observed associations may also introduce uncertainty in
the magnitude of the effect. One recent study examined the C-R relationship between blood Pb and total
mortality (Lanphear et al.. 2018). Similar to Menke et al. (2006). Lanphear et al. (2018) observed
generally sigmoidal spline curves with some evidence of attenuation of the C-R relationship below
2.5 (ig/dL (Figure 9-4).

The body of evidence for total mortality is supported by strong evidence of consistent positive
associations with cardiovascular mortality (Section 4.10. which comprises a large portion of total
mortality). In addition to a greater number of studies reporting consistent associations between BLLs and
cardiovascular mortality, the evidence base includes a wider range of study populations and expanded
evidence on the C-R relationship that generally supports a linear relationship with no evidence of a
threshold. There is coherence of effects across the scientific disciplines (i.e., animal toxicological,

External Review Draft

9-70

DRAFT: Do not cite or quote

-------
1

controlled human exposure, and epidemiologic studies) and biological plausibility for Pb-related
cardiovascular disease (Appendix 4), which supports the Pb-mortality relationship.

Overall, recent epidemiologic studies build upon evidence from the 2013 Pb ISA and provide
largely consistent evidence of an association between biomarkers of Pb exposure and total mortality. A
few uncertainties remain in the evidence base, including a limited number of studies and analyses of
similar or overlapping study populations. However, these studies are supported by more robust evidence
of Pb-related cardiovascular mortality, which comprises nearly 33% of total mortality. In addition,
evidence for cardiovascular morbidity provides biologically plausible pathways through which Pb
exposure could result in mortality. There is also very limited evidence that Pb exposure is positively
associated with other causes of mortality, including AD and infection. Biological plausibility for these
outcomes is demonstrated by pathways leading from Pb exposure to neurodegenerative disease and
immunosuppression, respectively. However, although there is toxicological evidence that developmental
exposure to Pb increases the expression of proteins related to AD, the epidemiologic evidence relating Pb
exposure to incident AD remains limited. The evidence for Pb-associated all-cause and cardiovascular
mortality and strong supporting evidence for Pb-associated cardiovascular effects indicates there is
sufficient evidence to conclude that there is a causal relationship between Pb exposure and total
(nonaccidental) mortality. The key evidence, as it relates to the causal framework, is summarized in
Table 9-3.

External Review Draft

9-71

DRAFT: Do not cite or quote

-------
Table 9-3 Summary of evidence for a causal relationship between Pb exposure and total mortality.

Rationale for Causality
Determination3

Key Evidence"

Key References"

Pb Biomarker Levels Associated
with Effects0

Consistent epidemiologic
evidence from multiple studies
at relevant BLLs

Increases in total mortality in multiple nationally
represented studies. Total mortality
associations are further supported by increases
in cardiovascular mortality conducted within
nationally represented studies.

(Hollinqsworth and Rudik, 2021: Bvun et
al.. 2020: Duan et al.. 2020: Lanphear et
al.. 2018: van Bemmel et al., 2011: Menke
et al., 2006)

Median, Mean, and Geometric
Mean BLLs: 1.49-2.71 pg/dL

Epidemiologic evidence
supports no evidence of a
threshold between Pb
biomarkers of exposure and
total mortality at the
concentration ranges
examined

Recent studies provide direct evidence of a
linear or sigmoidal, no-threshold C-R
relationship at lower concentrations of BLLs.

(Menke etal.. 2006)
(Lanphear et al., 2018)

Mean BLL: 2.58 pg/dL
Geometric Mean BLL: 2.71 pg/dL

Biological plausibility from
cardiovascular morbidity
evidence

Stronq evidence for coherence of effects across Appendix 4

scientific disciplines and evidence for a range of

cardiovascular effects in response to increases

in biomarkers of Pb exposure, especially for

increases in blood pressure and hypertension.

The collective body of cardiovascular morbidity

evidence provides biological plausibility for a

relationship between biomarkers of Pb

exposure and cardiovascular mortality, which

comprises -33% of total mortality.

BLLs = blood lead levels; C-R = concentration-response; Pb = lead.

"Based on aspects considered in judgments of causality and weight-of-evidence in causal framework
'Describes the key evidence and references, supporting or contradicting, contributing most heavily to
inconsistencies. References to earlier sections indicate where the full body of evidence is described.
°Describes the Pb biomarker levels at which the evidence is substantiated.

External Review Draft 9-72 DRAFT: Do not cite or quote

in Table I and Table II of the Preamble to the ISAs (U.S. EPA. 2015).
causality determination and, where applicable, to uncertainties or

-------
9.9

Evidence Inventories - Data Tables to Summarize Study Details

Table 9-4 Epidemiologic studies of exposure to Pb and hepatic effects.

Reference and Study
Design

Study Population

Exposure Assessment Outcome

Confounders

Effect Estimates and 95%
Clsa

Direct Evaluation of Liver Injury

tZhai et al. (2017)

Yangtze River Delta

Region

China

1 yr (2014)

Cross-sectional

SPECT-China
n = 2011

General population,
>18 yr old with no history
of excessive alcohol
consumption or viral
hepatitis

Blood

Pb measured in venous
whole blood using
atomic absorption
spectrometry
Age at measurement:
>18 yr old

Median:

Males: 5.29 [jg/dL
Females: 4.49 [jg/dL

25th:

Males: 3.61 [jg/dL
Females: 2.98 [jg/dL

75th:

Males: 7.28 [jg/dL
Females: 6.59 [jg/dL

Nonalcoholic fatty liver
disease

Two doctors
performed abdominal
ultrasounds and
categorized liver
status as normal or
fatty using predefined
criteria

Age at outcome:
>18 yr old

Age, region,
education, current
smoking, current
drinking, ALT,
diabetes, waist
circumference, BMI,
LDL cholesterol, HDL
cholesterol,
triglycerides, total
cholesterol, and
blood cadmium levels

ORs for NAFLD prevalence
across blood Pb quartiles

Males



Q1:

Ref.



Q2:

1.70

(0.84, 3.42)

Q3:

1.84

(0.88, 3.86)

Q4:

2.17

(0.99, 4.75)

Females
Q1: Ref.

Q2
Q3
Q4

1.38 (0.96, 2.00)
1.50 (1.02, 2.18)
1.61 (1.08, 2.41)

External Review Draft

9-73

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tWerder et al. (2020) Gulf Long-Term Follow-up Blood

Gulf Region
United States
2012-2013
Cross-sectional

Study
n = 214

Non-smoking >30 year old
male oil spill response
workers and oil spill safety
trainees with no history of
liver disease or heavy
alcohol use

Pb measured in venous
whole blood using solid-
phase micro-extraction
with gas

chromatography/mass

spectrometry

Age at measurement:

>30

Liver injury

Cytokeratin 18 (CK18
M65 and CK18 M30)

Age at outcome:

>30

Age, race, alcohol
consumption, serum
cotinine, BMI,
diabetes dx, and
education

Change in CK18 M65 (U/L)

2.4 (-12.69, 17.49)

Change in CK18 M30 (U/L)

21.7 (9.94, 33.46)

Mean: 1.82 (1.76)

tChunq et al. (2020)

South Korea
2 yr (2016-2017)
Cross-sectional

KNHANES
n = 4420

Adults, >20 yr old

Blood

Pb measured in venous
whole blood using
GFAAS

Age at measurement:
>20 yr old

Mean: 1.81 pg/dL
Max: 20.16 pg/dL

Hepatic steatosis and
fibrosis

Hepatic steatosis (HS)
as indicated by an HS
Index = 36 (8 x
(ALT/AST ratio) + BMI
(+2 if female; +2 if had
diabetes mellitus)).
Hepatic Fibrosis (HF)
as indicated by a
fibrosis-4 (FIB-4)
score >2.67
((age * AST
level)/(platelet level *
v(ALT level)).

Age at outcome:
>20 yr old

Age, smoking status,
alcohol consumption,
hypertension status,
obesity status,
diabetes status,
hypertriglyceridemia
status, blood Hg,
blood Cd.

ORs

Hepatic Steatosis

Men

0.83 (0.66, 1.03)
Women

0.98 (0.80, 1.19)

Fibrosis

Men

0.70 (0.44, 1.09)
Women

0.72 (0.42, 1.26)

External Review Draft

9-74

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tReia et al. (2020)

United States
5 yr (2011-2016)
Cross-sectional

NHANES
n = 2499

General population >20 yr
old with nonalcoholic fatty
liver disease (NAFLD)

Blood

>20 yr old
Mean: 1.01 pg/dL
75th: 1.62 pg/dL

Liver fibrosis

NAFLD Fibrosis Score

Age at outcome:
>20 yr old (concurrent
with exposure)

Age, gender, waist
circumference,
hypertension, liver
function test,
hemoglobin A1c,
triglycerides,
smoking, and PIR

ORs (NAFLD Fibrosis Score
>0.676)

Q1
Q2
Q3
Q4

Reference
2.79 (1.39, 5.63)
3.74 (2.01, 6.96)
5.93 (2.88, 12.24)

Serum Biomarkers of Liver Function

tPollack et al. (2015)

BioCycle

Blood

ALT, ALP, AST,

Linear mixed models

AST (% change):



n = 259



Bilirubin

adjusted for age,

5.02 (-1.36, 11.41)

Buffalo, NY



Pb measured in venous



BMI, race, average

ALT (% change):

United States

Premenopausal women

whole blood using ICP-

ALT (U/L), ALP (U/L),

calories, alcohol

2 menstrual cycles (8

followed for 2 menstrual

MS

AST (U/L), Bilirubin

intake, smoking, and

6.39 (3.07, 9.72)

visits per cycle) (2005-

cycles



(mg/dL)

cycle day

ALP (% change):

2007)

Age at measurement:

Cohort



27.4 (SD: 8.2)

Age at outcome:



2.14 (-5.05, 9.33)





27.4 (SD: 8.2)









1.03 pg/dL







tChen et al. (2019)

Guangdong
China
1 yr(2015)
Cross-sectional

n = 267

Hospitalized patients from
two regions in Guangdong
(one e-waste polluted
area and a matched
control area). Patients
with heart or kidney
disease, those taking
drugs with hepatic toxicity,
and those with a history of
alcohol consumption or
smoking were excluded.

Blood

Pb was measured in
venous whole blood
using GFAAS

Age at measurement:
4 to 85 yr old

Median:

Exposed: 8.7 pg/dL;
Control: 5.1 pg/dL
75th:

Exposed: 12.2 pg/dL;
Control: 8.4 pg/dL

Abnormal liver
function

Abnormal liver
function defined as
two transminases
(AST, ALT, or GGT)
above normal range
or one at least two
times higherthan
normal range (40 U/L)

Age at outcome:
4 to 85 yr old
(concurrent with
exposure)

Age, gender, hepatic
disease, RBC, Hb,
and platelets

OR for Prevalence of
Abnormal Liver Function

1.94 (1.00, 3.73)

External Review Draft

9-75

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tChristensen et al.
(2013)

United States
2 yr (2003-2004)
Cross-sectional

NHANES
n = 1345

General population,
>12 yr old. No chronic
hepatitis or liver disease,
and no high alcohol
intake.

Blood

Pb measured in venous
whole blood using ICP-
MS

Age at measurement:
>12 yr old

Liver function

Serum ALT

Age at outcome:
>12 yr old

Sex, Race/Ethnicity,
Age, PIR, BMI

Change in ALT (U/L)

Q1
Q2
Q3
Q4

Reference

-0.068 (-0.14, 0.004)
-0.039 (-0.113, 0.035)
-0.103 (-0.185, -0.021)

Mean NR

tObena-Gvasi (2019) NHANES

United States
NHANES 2009-2016
Cross-sectional

n = 7,730 young adults
(18-44); 5,744 middle-
aged adults (45-65)

General population; ages
18-65

Blood

BLL measured in venous
whole blood using ICP-
MS

Age at measurement:
>18 yr old

Mean:

Young adults:

1.03 [jg/dL

Middle-aged adults:
1.62 [jg/dL

GGT (U/L)

Serum GGT (U/L)

Age at outcome:
>18 yr old

Gender, BMI,
income, ethnicity,
and alcohol
consumption

ORs (GGT >18 U/L)

Young Adults
1.94 (1.65, 2.28)

Middle-Aged Adults
1.34 (1.14, 1.58)

External Review Draft

9-76

DRAFT: Do not cite or quote

-------
Reference andstudy study Population Exposure Assessment	Outcome	Confounders Effect Estimates and 95%

Serum Lipids

tPeters et al. (2012)

United States
Blood Pb measured
between 1999-2008;
Serum lipids measured
3 to 4 yr after blood Pb
Cohort

Normative Aging Study
n = 426

Older male Veterans

Blood, Bone

Blood Pb measured in
venous whole blood
using GFAAS

Serum lipids

Triglycerides, total
cholesterol, HDL-C,
LDL-C

Mean: 4.01 ± 2.30 [jg/dL Age at outcome:

3 to 4 yr after blood
Pb

Age at baseline, yr
between baseline
and outcome,
education, BMI,
alcohol intake,
smoking status,
pack-yr of smoking,
hypertension status,
and statin use

ORs

Total C (>200 mg/dL)
1.08 (0.99, 1.19)

LDL-C (>130 mg/dL):
1.02 (0.91, 1.15)

HDL-C (<40 mg/dL):
0.90 (0.80, 1.00)

Triglycerides (>200 mg/dL):
0.99 (0.87, 1.13)

tXu et al. (2021)

United States
NHANES 2005-2016
Cross-sectional

NHANES
n = 7457

General population; Ages
20 to 79 yr old

Blood

Pb measured in venous
whole blood samples
using ICP-MS
Age at measurement:
Mean (SD):

43.68 (15.02) yr
GM: 1.23 [jg/dL

Dyslipidemia

Total cholesterol,
LDL-C, non-HDL-C,
triglycerides

Age at outcome:
Mean (SD): 43.68
(15.02)

Age, sex, race, BMI,
education status,
smoking status,
alcohol consumption,
physical activity, PIR,
systolic blood
pressure, serum
cotinine, and Cd

RRs

Total C (>200 mg/dL)

1.01	(1.00, 1.01)

non-HDL-C (>160 mg/dL)
1.00 (0.99, 1.01)

LDL-C (>130 mg/dL)

1.02	(1.00, 1.04)

Triglycerides (>200 mg/dL)
0.99 (0.98, 1.00)

External Review Draft

9-77

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tLee and Kim (2016)

Korea

2005-2010

Cross-Sectional

Korean National Health
and Nutrition Examination
Survey (KNHANES)
n = 7559

Korean adults aged 20+

Blood

Pb measured in venous
whole blood using
GFAAS

Age at measurement
Mean (SD):

No MetS: 42.32 (0.294)
yr; MetS: 48.36 (0.574)
yr

Geometric Mean (SD)
No MetS: 2.73 (0.024)
[jg/dL; MetS: 2.96
(0.049) [jg/dL

Serum Lipids

Low HDL cholesterol
(<40 mg/dL in women
or <50 mg/dL in men);
Elevated serum
triglycerides
(=150 mmHg)

Age at outcome
same as age at
exposure assessment

Age, BMI, residence
area, education level,
smoking and drinking
status, exercise,
serum aspartate
aminotransferase,
serum alanine
aminotransferase

ORs

HDL-C <40 mg/dL
0.84 (0.66, 1.08)

TG >150 mg/dL
1.12 (0.90, 1.39)

tEttinqer et al. (2014)

Kumasi (Ghana), Cape
Town (South Africa),
Victoria (Seychelles),
Kingston (Jamaica),
Maywood, IL (United
States)

Ghana, South Africa,
Seychelles, Jamaica,
United States

2010-2014

Cross-sectional

Modeling the
Epidemiologic Transition
Study (METS)
n = 150

Adults of African descent
from 5 countries of
varying social and
economic development

Blood

Pb measured in venous
whole blood using ICP-
MS

Age at measurement
Mean (SD):

Males: 34.7 (6.0) yr;
Females: 35.2 (6.2) yr

Geometric Mean:
1.55 [jg/dL
Median: 1.66 [jg/dL
75th: 2.6 pg/dL
Max: 31.82 pg/dL

HDL and LDL
cholesterol, blood
pressure,
triglycerides.

Height and weight
were measured by
physical examination.
Fasting glucose was
measured in blood.
Further outcome
assessment details
not provided.

Age at outcome is the
same as age at
exposure assessment

Age, sex, site
location, marital
status, education,
paid employment,
alcohol use, fish
intake

ORs (>1.66 [jg/dL vs.
<1.66 [jg/dL blood Pb)

LDL-C (>2.59 mmol/L)
0.680 (0.289, 1.597)

Triglycerides (>1.7 mmol/L)
0.09 (0.030, 0.250)

HDC-C (<1.03 [males]; <1.29
[females] mmol/L)

1.93 (0.740, 5.020)

External Review Draft

9-78

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tLiu et al. (2020)

Mexico City
Mexico

Pregnant women
recruited between
1997-1999 and 2001-
2003, follow-up among
offspring began in 2015
Cohort

Early Life Exposures in
Mexico to Environmental
Toxicants (ELEMENT)
n = 369

Mother/child pairs from a
birth cohort study of
pregnant women from 2
public hospitals serving
low to moderate-income
populations

Blood

Maternal Blood Pb
measured in venous
whole blood using
GFAAS

Age at measurement:
Maternal age (SD):

26.7 (5.6) yr

Mean of prenatal blood:
4.3 [jg/dL

Serum lipids

Total cholesterol,
triglycerides, HDL-C,
LDL-C

Age at outcome
Child's age (SD):

13.7 (1.9) yr

Child age, sex, BMI,	Change in Z-score (>5 pg/dL

number of siblings at	vs. <5 pg/dL blood Pb)

birth, maternal age,

marital status	__ . .

education, smoking	Triglycerides

history	0.58 (-0.05, 1.20)

Total cholesterol

-0.76 (-1.38, -0.13)

HDL-C

-0.64 (-1.28, 0.01)

LDL-C

-0.96 (-1.59, -0.33)

tKupsco et al. (2019)

Mexico City
Mexico

Maternal blood tested
for metals in 2nd
trimester, children
assessed at age 4-6
Cohort

Research in Obesity,
Growth Environment and
Social Stress
(PROGRESS) birth study
n = 548

Mother/child pairs from a
birth cohort study

Blood

Maternal blood Pb
measured second
trimester in venous
whole blood samples
using ICP-MS

Age at measurement
Mean (SD):

28 (5.6) yr

Mean (SD):

3.7 (2.7) [jg/dL
Max: 18 pg/dL

Serum lipids

Triglycerides and non-
HDL cholesterol

Age at outcome:

Mean: 4.8 yr; Range:
4-6 yr

Birth weight,
gestational age,
prepregnancy BMI,
education, SES,
parity, environmental
tobacco smoke

Change in Z-score

Triglycerides
0.018 (-0.028, 0.064)

non-HDL-C
-0.015 (-0.058, 0.028)

External Review Draft

9-79

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tXu etal. (2017)

United States

1999-2012

Cross-sectional

NHANES
n = 11662

General population; 12-
19 yr old

Blood

Serum lipids

Pb measured in venous Total cholesterol,
whole blood using ICP- triglycerides, HDL-C,
MS

Age at measurement:
12-19 yr

Mean (SD):
1.17 (1.20) [jg/dL

LDL-C

Age at outcome:
12-19 yr

% Increase

Age, gender,
ethnicity, PIR, waist

circumference, serum Tota/ cholesterol
cotinine, and physical
activity

0.6% (-0.1%, 1.3%)

HDL-C

0.3% (-0.5%, 1.1%)
LDL-C

2.3% (0.3%, 4.2%)

Triglycerides
-1.1% (-2.4%, 0.2%)

ALP = alkaline phosphatase; ALT = alanine aminotransferase; AST = aspartate aminotransferase; BMI = body mass index; Cd = cadmium; CI = confidence interval;
CK18 = cytokeratin 18; ELEMENT = Early Life Exposures in Mexico to Environmental Toxicants; FIB-4 = fibrosis-4; GFAAS = graphite furnace atomic absorption spectrometry;
GGT = gamma-glutamyl transferase; Hb = hemoglobin; HDL = high-density lipoprotein; HDL-C = high-density lipoprotein cholesterol; HF = hepatic fibrosis; HS = hepatic
steatosis; ICP-MS = inductively coupled plasma mass spectrometry; KNHANES = Korean National Health and Nutrition Examination Survey; LDL = low-density lipoprotein; LDL-
C = low-density lipoprotein cholesterol; MetS = metabolic syndrome; METS = Modeling the Epidemiologic Transition Study; NAFLD = nonalcoholic fatty liver disease;

NHANES = National Health and Nutrition Examination Survey; NR = not reported; OR = odds ratio; Pb = lead; PIR = poverty-income-ratio; PROGRESS = Programming Research
in Obesity, Growth Environment and Social Stress; RBC = red blood cell; RR = risk ratio; SD = standard deviation; SES = socioeconomic status; SPECT = single photon emission
computed tomography; Q = quartile; yr = year(s).

aEffect estimates are standardized to a 1 |jg/dL increase in BLL or a 10 |jg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.

fStudies published since the 2013 Integrated Science Assessment for Lead.

1

External Review Draft

9-80

DRAFT: Do not cite or quote

-------
Table 9-5 Animal toxicological studies of exposure to Pb and hepatic effects.

Study

Species (Stock/Strain),
n, Sex

Timing of
Exposure

Exposure Details BLL As Reported (pg/dL)

Endpoints Examined

Berrahal etal. (20111

Rat (Wistar)

0 mg/L Pb Acetate, M,
n = 12-16

50 mg/L Pb Acetate, M,
n = 12-16

PND40, 65

Oral, drinking water

1.76 ± 0.33 [jg/100 mL for
0 mg/L Pb Acetate,

12.67 ± 1.68 [jg/100 mL for
50 mg/mL Pb Acetate - PND 40

2.06 ± 0.35 [jg/100 mL for
0 mg/L Pb Acetate,

7.49 ± 0.78 pg/100 mL for
50 mg/mL Pb Acetate - PND 65

Plasma Alanine
Aminotransferase (ALT),
Plasma Aspartate
Aminotransferase (AST),
Plasma Alkaline Phosphatase
(ALP)

Li etal. (2017)

Mouse (BALBc)

0 mg/kg Pb Acetate, F,
n = 8

100 mg/kg Pb Acetate, F,
n = 8

Day 29 from	Oral, gavage	0.43 ± 0.05 |jg/L for 0 mg/kg Pb

exposure start	Acetate

302.20 ± 25.32 pg/L for
100 mg/kg Pb Acetate

Malondialdehyde (MDA)
Levels, Glutathione (GSH),
Glutathione Peroxidase
(GSH-PX), Total Superoxide
Dismutase (T-SOD)

Liu etal. (2013)

Rat (Wistar)

0 ppm Pb, M, n = 10
500 ppm Pb, M, n = 10

Exposure d 75 Oral, drinking water 0.0448 pg/dL for 0 ppm

0.450 pg/dL for 500 ppm

Plasma Alanine
Aminotransferase (ALT),
Plasma Aspartate
Aminotransferase (AST),
GRP78 Protein Levels,
Reactive Oxygen Species
Activity, TBARS Levels, Total
Antioxidant Capacity, ATF6
Protein Levels, ATF4 Protein
Levels, P-IRE1 Protein
Levels, T-IRE1 Protein
Levels, XBP-1 Protein Levels,
P-JNK Protein Levels, JNK
Protein Levels, PI3K Protein
Levels, P-Akt Protein Levels,
T-Akt Protein Levels

External Review Draft

9-81

DRAFT: Do not cite or quote

-------
Study

Species (Stock/Strain), Timing of
n, Sex	Exposure

Exposure Details BLL As Reported (pg/dL)

Endpoints Examined

Long etal. (2016)

Mouse (Kunming)

0% Pb Acetate, M, n = 7
0.2% Pb Acetate, M,
n =21

Six weeks
exposure

Oral, drinking water

36.42 ± 17.48 [jg/L for 0% Pb
Acetate, 214.64 ± 36.24 pg/L for
0.2% Pb Acetate

Plasma Alkaline Phosphatase
(ALP), Plasma Alanine
Aminotransferase (ALT),
Plasma Aspartate
Aminotransferase (AST),
Malondialdehyde (MDA)
Levels, Glutathione (GSH),
Glutathione Peroxidase
(GSH-PX), Total Superoxide
Dismutase (T-SOD),
Apoptosis, Bcl-2 Gene
Expression, Bax Gene
Expression, Bcl-2 Protein
Levels, Bax Protein Levels,
Nrf2 Protein Levels, HO-1
Protein Levels, Gamma-GCS
Protein Levels, Nrf-2 Gene
Expression, HO-1 Gene
Expression, Gamma-GCS
Gene Expression, GRP78
Protein Levels, Grp78 Gene
Expression, Chop Gene
Expression

Andielkovic et al. (2019)

Rat (Wistar)

0 mg Pb Acetate per kg
bw, M, n = 8
150 mg Pb Acetate per
kg bw, M, n = 6

24 h posttreatment Oral, gavage

25 pg/L for 0 mg Pb Acetate per
kg bw, 290 pg/L for 150 mg Pb
Acetate per kg bw

Plasma Aspartate
Aminotransferase (AST),
Plasma Alanine
Aminotransferase (ALT),
Plasma Alkaline Phosphatase
(ALP), Lactate
Dehydrogenase (LDH),
Malondialdehyde (MDA)
Levels, Advanced Oxidation
Protein Products Level
(AOPP), Total Thiol (-SH)
Groups Level, Prooxidative-
Antioxidative Balance (PAB),
Total Superoxide Dismutase
(T-SOD)

External Review Draft

9-82

DRAFT: Do not cite or quote

-------
Study

Species (Stock/Strain), TimmgoJ Exposure Details BLL As Reported (pg/dL)	Endpoints Examined

Dumkova et al. (2017)

Mouse (ICR)	Week 6 of

0 particles/cm3, F, n = 10 exPosure
1.23 x 10s particles/cm3,

F, n = 10

Inhalation

1.1 [jg/dL for 0 particles/cm3,
13.2 [jg/dL for 1.23 * 10®
particles/cm3, F, n = 10

Histopathology, Proliferating
Cell Nuclear Antigen (PCNA)
Immunohistochemistry,
Apoptotic Cells (TUNEL-
Positive), Na-KATPase
Expression

Barkaoui et al. (2020)

Rat (Wistar)

0	g/L Pb Acetate, M,
n =6

1	g/L Pb Acetate, M,
n =6

Exposure day 30

Oral, drinking water 11.1 ±0.12 pg/dL for 0 g/L Pb
Acetate

23.8 ± 0.912 [jg/dL fori g/L Pb
Acetate

GSH, CAT, T-SOD, GSH-PX,
MDA Levels, Histopathology,
CAT qRT-PCR, GPx qRT-
PCR, SOD qRT-PCR, NF-kB
qRT-PCR, IL-6 qRT-PCR,
TNF-alpha qRT-PCR

Gao et al. (2020)

Rat (Sprague Dawley)

0 mg/kg bw, Pb2+, M/F,
n = 10

5 mg/kg bw, Pb2+, M/F,
n = 10

Four weeks	Oral, gavage	0.02 mg/kg for 0 mg/kg bw,

postexposure	Pb2+,

0.1 ± 0.03 mg/kg for 5 mg/kg
bw, Pb2+

T-SOD, CAT, MDA Levels,
GSH, Histopathology, Plasma
AST, Plasma ALT, Cr, BUN

Dumkova et al. (2020b)

Mouse (Not Specified)
0 |jg/m3 PbO NPs, F,
n = NR, 2, 6, 11 wk
78.0 |jg/m3 PbO NPs, F,
n = NR, 6 wk followed by
0 |jg/m3 PbO NPs, 5 wk
78.0 |jg/m3 PbO NPs, F,
n = NR, 2, 6, 11 wk

Exposure week 2,
6, 11

Inhalation

0 [jg/dL for 0 |jg/m3 PbO NPs,
F, n = NR, 2, 6, 11 wk

2.7 [jg/dL for 78.0 |jg/m3 PbO
NPs, F, n = NR, 6 wk followed
by 0 |jg/m3 PbO NPs, 5 wk

10.4 ug/dL for 78.0 ug/m3 PbO
NPs-2 wk

14.8 ug/dL for 78.0 ug/m3 PbO
NPs-6 wk

17.4 [jg/dL for 78.0 |jg/m3 PbO
NPs-11 wk

Plasma Alkaline Phosphatase
(ALP), Plasma Alanine
Aminotransferase (ALT),
Plasma Aspartate
Aminotransferase (AST), Cr

External Review Draft

9-83

DRAFT: Do not cite or quote

-------
Study

Species (Stock/Strain), Timing of
n, Sex	Exposure

Exposure Details BLL As Reported (pg/dL)

Endpoints Examined

Dumkova et al. (2020a)

Mouse (CD1), (ICR)

0 |jg/m3 Pb(N03)2 NPs,
F, n = 10-3d, 2, 6,
11 wk

68.6 |jg/m3 Pb(N03)2
NPs, F, n = 10 - 3 d, 2, 6
11 wk

68.6 |jg/m3 Pb(N03)2
NPs, F, n = 10 - 6 wk,
followed by 0 |jg/m3
Pb(N03)2 NPs - 5 wk

Exposed 3 d,
11 wk

2, 6,

Inhalation

0 [jg/dL for 0 |jg/m3 - all groups
3.1 [jg/dL for 68.6 |jg/m3 -3d
4.0 pg/dL for 68.6 |jg/m3 - 2 wk
4.7 [jg/dL for 68.6 |jg/m3 - 6 wk
8.5 pg/dL for 68.6 |jg/m3 -11 wk
1.0 [jg/dL for 68.6 |jg/m3 - 6 wk
followed by 0 |jg/m3 - 5 wk

Histopathology, NF-kB qRT-
PCR, TNF-alpha qRT-PCR,
IL-1 alpha, IL-1 beta, IL-6
qRT-PCR, TGFbetal, Plasma
Alkaline Phosphatase (ALP)

Laamech et al. (2017)

Mouse (IOPS)

0 mg/kg body weight/day
Pb Acetate, M, n = 12
5 mg/kg body weight/day
Pb Acetate, M, n = 12

Exposure day 40 Oral, gavage

0.010 pg/mLforO mg/kg body
weight/day Pb Acetate,
0.18 |jg/ml_ for 5 mg/kg body
weight/day Pb Acetate

Histopathology, Plasma
Alanine Aminotransferase
(ALT), Plasma Aspartate
Aminotransferase (AST),
Total Cholesterol (TC), Total
Bilirubin (TB),
Malondialdehyde (MDA)
Levels, Protein Carbonyl
(PCO), Glutathione (GSH),
Catalase, Total Superoxide
Dismutase (T-SOD),
Glutathione Peroxidase
(GSH-PX)

ALP = alkaline phosphatase; ALT = alanine aminotransferase; AOPP = advanced oxidation protein products; AST = aspartate aminotransferase; BUN = blood urea nitrogen;
BLL = blood lead levels; CAT = catalase; Cr = chromium; D = day(s); GSH = glutathione; GSH-PX = glutathione peroxidase; LDH = lactate dehydrogenase; h = hour;
MDA = malondialdehyde; NF-kB = nuclear factor kappa B; NP = nanoparticle; PAB = prooxidative-antioxidative balance; Pb = lead; PCNA = proliferating cell nuclear antigen;
PCO = protein carbonyl; PND = postnatal day; qRT-PCR = real-time quantitative reverse transcription-polymerase chain reaction; TB = total bilirubin; TBARS = thiobarbituric acid
reactive substance; TC = total cholesterol; T-SOD = total superoxide dismutase; wk = week(s).

External Review Draft

9-84

DRAFT: Do not cite or quote

-------
Table 9-6 Epidemiologic studies of exposure to Pb and metabolic effects.

sSEE.S?	Population Exposure Assessment	Outcome	Contenders Ett^ Estimates ,nd 95%

Diabetes and Insulin Resistance - Adults

tMoon (2013)

Korea

2007-2012

Cross-Sectional

KNHANES
n = 3,184

Adults aged
>30 yr

Blood

Pb was measured in venous
whole blood using GFAAS

Age at measurement
Mean (SD):

No diabetes: 49.4 (12.4) yr
Diabetes: 58.8 (10.9) yr

Geometric Mean (SD):

No diabetes:

2.41 (1.52) [jg/dL

Diabetes:

2.47 (1.59) [jg/dL

Diabetes, HOMA-IR, HOMA-I3 Age, sex, region,

(%), fasting insulin (mlU/L)

Age at outcome is the same
as age at exposure
assessment

smoking, alcohol
consumption, regular
exercise, BMI (sex-
stratified analyses
only)

OR (95% CI) for prevalent
diabetes across blood Pb
quartiles:

Q1 (GM 1.43 pg/dL):
Reference

Q2 (GM 2.13 pg/dL):
0.91 (0.64, 1.29)

Q3 (GM 2.74 pg/dL):
0.76 (0.53, 1.09)

Q4 (GM 4.08 pg/dL):
0.75 (0.52, 1.08);

Change in HOMA-IR,
HOMA-IJ, and Fasting
Insulin per unit increase
in log-blood Pb

log(HOMA-IR)

Men: -0.04 (-0.10, -0.02),
Women: -0.04 (-0.09,
-0.01)

log(HOMA-IJ)

Men: -0.05 (-0.11, 0.01),
Women: -0.05 (-0.10,
0.01)

Fasting insulin (mlU/L)
Men: -0.53 (-1.23, 0.16)
Women: -0.27 (-1.00,
0.46)

External Review Draft

9-85

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tHansen et al. (2017)

Nord-Trondelag



Health Study

Nord-Trondelag

(HUNT3)

County

n = 883

Norway



2006-2008

Adults aged

Nested Case-Control

>20 yr. Cases



(n = 128) were



HUNT3



participants



diagnosed with



diabetes.



Controls



(n = 755) were



age- and sex-



matched HUNT3



participants



without diabetes.

Blood

Pb was measured in venous
whole blood using ICP-MS

Age at measurement
Mean (SD):

Cases: 61.4 (14.1) yr
Controls: 65.2 (10.3) yr

Median (10th—90th
percentile):

Cases: 19.9 (10.8-38.0) pg/L
Controls: 19.4 (11.0-37.2)
pg/L

Type 2 diabetes

Individuals were screened for
diabetes at a physical
examination using an oral
glucose tolerance test.
Diagnosis with type 2
diabetes was defined as
having fasting serum glucose
>7.0 mmol/L and/or 2 h
glucose >11.1 mmol/L as well
as glutamic acid
decarboxylase antibodies
(GADA) <0.08 ai.

Age at outcome is the same
as age at exposure
assessment

Age, sex, BMI, waist-
to-hip ratio,
education, income,
smoking, family
history of diabetes

OR (95% CI) for prevalent
type 2 diabetes Q4 vs Q1:

1.12 (0.58, 2.16)

tSimic etal. (2017)

Norway

2006-2008

Nested Case-Control

Nord-Trondelag
Health Study
(HUNT3)
n = 945

Adults aged
>20 yr. Cases
(n = 270) were
HUNT3
participants
diagnosed with
type 2 diabetes.
Controls
(n = 615) were
age- and sex-
matched
participants
without diabetes.

Blood

Pb was measured in venous
whole blood using ICP-MS

Age at measurement
Mean (SD):

Cases: 59.2 (12.2) yr
Controls: 65.4 (10.6) yr

Median (10th—90th
percentile):

Cases: 16.4 (9.7-35.2) pg/L
Controls: 20.2 (11.2-37.9)
pg/L

Type 2 diabetes

Type 2 diabetes was defined
as self-reported diabetes
excluding type I diabetes as
indicated by GADA index,
measured in blood at a
physical examination.

Age at outcome is the same
as age at exposure
assessment

BMI, waist-to-hip
ratio, first-degree
family history of
diabetes, smoking
habits, area,
education, economic
status, alcohol
consumption, blood
calcium

OR (95% CI) for prevalent
type 2 diabetes Q4 vs Q1:

0.24 (0.13, 0.47)

External Review Draft

9-86

DRAFT: Do not cite or quote

-------
Study DCesfgnnd	Population	Exposure Assessment	Outcome	Confounders Effect Estimates and 95%

Diabetes and Insulin Resistance - Adolescents

tLiu et al. (2020)

Mexico City
Maternal enrollment:
1997-1999 and
2001-2003

Child follow-up: 2015

Prospective Birth
Cohort

Early Life
Exposures in
Mexico to
Environmental
Toxicants
(ELEMENT)
Study
n = 369

Adolescents
aged 10-18 yr

Blood

Maternal Pb (1st trimester)
was measured in venous
whole blood using GFAAS

Age at measurement
Mean maternal age in 1st
trimester of pregnancy (SD):
26.7 (5.6) yr

Geometric Mean (95% CI):
4.3 (4.0, 4.6) [jg/dL

Fasting serum glucose Z-
score (mg/dL), HOMA-IR Z-
score

Serum fasting glucose
(mg/dL) was measured using
an enzymatic method. Serum
insulin (|jU/mL) was
measured using
immunoturbidimetric assay.
HOMA-IR was calculated as
insulin (|jU/mL)*glucose
(mg/dL)/405.

Age at outcome
Mean child age (SD):
13.7 (1.9) yr

Child age, sex, BMI z-
score, number of
siblings at birth,
maternal age, marital
status, education,
smoking history

Change in mean fasting
glucose and HOMA-IR Z-
scores for maternal blood
Pb >5 |jg/dL vs. maternal
blood Pb <5 [jg/dL

Fasting glucose z-score
All: -0.05 (-0.69, 0.60)
Boys: -0.05 (-0.34, 0.25)
Girls: -0.06 (-0.35, 0.23)

HOMA-IR z-score

All: -0.11 (-0.63, 0.42)
Boys: -0.04 (-0.28, 0.20)
Girls: 0.04 (-0.19, 0.27)

External Review Draft

9-87

DRAFT: Do not cite or quote

-------
Study DCesfgnnd	Population	Exposure Assessment	Outcome	Confounders Effect Estimates and 95%

Metabolic Syndrome (MetS) and Its Components

tMoon (2014)

Korea

2007-2012

Cross-Sectional

KNHANES
n = 3,950

Adults aged
>20 yr

Blood

Pb measured in venous
whole blood using GFAAS.

Age at measurement
Mean (SD):

No MetS: 42.7 (14.6) yr;
MetS: 54.4 (12.8) yr

Mean (SD)

No MetS: 2.08 (1.00) pg/dL;
MetS: 2.50 (1.01) pg/dL

Metabolic syndrome

MetS was defined as meeting
at least 3 of the following: (1)
elevated blood pressure (SBP
>130 mmHg or DBP
>85 mmHg or current use of
blood pressure medication),
(2) low HDL cholesterol
(<40 mg/dL in women or
<50 mg/dL in men), (3)
elevated serum triglycerides
(>150 mmHg) or current use
of antidyslipidemia
medication, (4) elevated
fasting plasma glucose levels,
(5) abdominal obesity (waist
circumference >90 cm in men
or >85 cm in women).

Age, sex, region,
smoking, alcohol
consumption, regular
exercise, BMI

OR (95% CI) for prevalent
MetS across blood Pb
quartiles

Q1 (GM 1.23 pg/dL):
Reference

Q2 (GM 1.90 pg/dL):
0.84 (0.62, 1.13)

Q3 (GM 2.51 pg/dL):
1.21 (0.90, 1.62)

Q4 (GM 3.79 pg/dL):
1.07 (0.79, 1.45)

Age at outcome is the same
as age at exposure
assessment

External Review Draft

9-88

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tRhee et al. (2013)

Korea
2008

Cross-Sectional

KNHANES
n = 1,405

Nationally
representative
survey of Korean
adults

Blood

Pb was measured in venous
whole blood using GFAAS

Age at measurement
Mean (SD):

No MetS: 40.3 (13.7) yr
MetS: 47.1 (13.3) yr

Median (25th—75th):
2.35 (1.74-3.06) pg/dL
75th: 3.06 pg/dL
Max: 19.43 pg/dL

MetS, abdominal
circumference, triglycerides,
HDL cholesterol, fasting
glucose

MetS was defined using the
Modified National Cholesterol
Education Program Adult
Treatment Panel III Criteria,
with the exception of waist
circumference measurement
cut-offs of >90 cm for males
and >85 cm for females
based on criteria from the
Korean Society for the Study
of Obesity. TC, triglycerides,
HDL cholesterol, and fasting
plasma glucose were
assessed using an automated
analyzer with enzymatic
assays. Abdominal
circumference was measured
by a professional.

Age at outcome is the same
as age at exposure
assessment

Age, sex, smoking,
education, TC,
creatinine, AST, AMT,
fasting serum insulin

OR for MetS prevalence
across log-transformed
Pb quartiles

Q1 (<1.73 pg/dL):
Reference

Q2 (1.74-2.35 pg/dL):

1.56	(0.90, 2.71)

Q3 (2.35-3.06 pg/dL):
1.63 (0.94, 2.83)

Q4 (>3.07 pg/dL):

2.57	(1.46, 4.51)

Change in outcomes per
unit increase in log-
transformed Pb

Abdominal circumference
0.051 (-0.001, 0.107) cm

Triglycerides

0.080 (0.023, 0.137) mg/dL

HDL Cholesterol
0.033 (-0.020, 0.086)
mg/dL

Fasting Glucose

0.019 (-0.029, 0.067)
mg/dL

External Review Draft

9-89

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tBulka etal. (2019)

United States

2011-2014

Cross-Sectional

NHANES
n = 1,088

Nationally
representative
survey of U.S.
adults

Blood

Pb was measured in venous
whole blood using ICP-MS

Age at measurement:
20-60 yr

Mean (SD)

NHANES 2011-2012:
1.17 (0.04) [jg/dL;

NHANES 2013-2014:

1.00 (0.03) [jg/dL

MetS, triglycerides, HDL
cholesterol, blood glucose,
abdominal obesity

MetS was defined as meeting
at least 3 of the following: (1)
elevated blood pressure (SBP
>130 mmHg or DBP
>85 mmHg or current use of
blood pressure medication),
(2) low HDL cholesterol
(<40 mg/dL in women or
<50 mg/dL in men), (3)
elevated serum triglycerides
(>150 mmHg) or current use
of antidyslipidemia
medication, (4) elevated
fasting plasma glucose levels,
(5) abdominal obesity (waist
circumference >90 cm in men
or >85 cm in women). Waist
circumference (cm) was
measured at the physical
examination by a trained
professional. Serum HDL
(|jg/dL), triglycerides (mg/dL),
and blood glucose (mg/dL)
were measured in blood
samples obtained in the
morning following an
overnight fast.

Age at outcome:

20-60 yr

Age, race/ethnicity,
family income-poverty
ratio, total caloric
intake, educational
attainment, smoking
status, average
number of drinks per
day past year,
physical activity
status, survey cycle,
BMI (excluding
abdominal obesity
analysis), serum
cotinine

PRs for outcomes across
Pb quartiles

MetS

Q1 (0.18-0.70 pg/dL):
Reference

Q2 (0.71-1.05 pg/dL):
0.90 (0.73, 1.11)

Q3 (1.06-1.63 pg/dL):
0.84 (0.69, 1.05)

Q4 (1.64-15.98 pg/dL):
0.81 (0.64, 1.03)

High Triglycerides

Q1
Q2
Q3
Q4

Reference
0.85 (0.72, 0.99)
0.76 (0.64, 0.92)
0.82 (0.67, 1.01)

Low HDL
Q1: Reference

Q2
Q3
Q4

0.90 (0.76, 1.07)
0.79 (0.65, 0.97)
0.73 (0.59, 0.89)

High Glucose
Q1: Reference

Q2
Q3
Q4

1.03 (0.86, 1.23)
0.86 (0.68, 1.08)
0.95 (0.77, 1.17)

Abdominal Obesity
Q1: Reference

Q2
Q3
Q4

0.93 (0.82, 1.07)
0.91 (0.80, 1.04)
0.66 (0.56, 0.78)

External Review Draft

9-90

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tShimetal. (2019)

Korea

2012-2014

Cross-Sectional

Korean National
Environmental
Health Survey II
(KNEHS)
n = 5,251

Nationally
representative
survey of adults
in Korea

Blood

Pb was measured in venous
whole blood using GFAAS

Age at measurement
Mean (SE):

No MetS: 49.87 (0.22) yr
MetS: 61.59 (0.50) yr

Geometric Mean (SE)
No MetS: 0.71 (0.48) pg/dL
MetS: 0.76 (0.49) pg/dL

MetS

MetS was defined as meeting
at least 3 of the following: (1)
elevated blood pressure (SBP
>130 mmHg or DBP
>85 mmHg or current use of
blood pressure medication),
(2) low HDL cholesterol
(<40 mg/dL in women or
<50 mg/dL in men), (3)
elevated serum triglycerides
(>150 mmHg) or current use
of antidyslipidemia
medication, (4) elevated
fasting plasma glucose levels,
(5) abdominal obesity (waist
circumference >90 cm in men
or >85 cm in women).

Age, sex, education,
income, marital
status, aspartate
aminotransferase,
alanine

aminotransferase

ORs for MetS prevalence
across blood Pb quartiles

Q1
Q2
Q3
Q4

Reference
0.94 (0.72, 1.24)
1.00 (0.76, 1.31)
0.86 (0.65, 1.14)

Quartile levels NR

Age at outcome is the same
as age at exposure
assessment

tWen et al. (2020) N = 2444

Taiwan
June 2016-
September 2018
Cross-Sectional

General
population

Blood

Pb was measured in venous
whole blood using ICP-MS

Age at measurement:

Mean (SD): 55.1 (13.2) yr

MetS

Age, sex, TC, LDL
cholesterol,
hemoglobin, eGFR,
uric acid

OR MetS prevalence per
log unit increase in blood
Pb:

0.86 (0.61, 1.20)

Mean: 1.6 pg/dL

External Review Draft

9-91

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

MetS was defined as meeting
at least 3 of the following: (1)
elevated blood pressure (SBP
>130 mmHg or DBP
>85 mmHg or current use of
blood pressure medication),
(2) low HDL cholesterol
(<40 mg/dL in women or
<50 mg/dL in men), (3)
elevated serum triglycerides
(>150 mmHg) or current use
of antidyslipidemia
medication, (4) elevated
fasting plasma glucose levels,
(5) abdominal obesity (waist
circumference >90 cm in men
or >85 cm in women).

Age at outcome:

Mean (SD): 55.1 (13.2) yr

External Review Draft

9-92

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tLee and Kim (2016) KNHANES
n = 9,880

Korea

2007-2012	Korean adults

Cross-Sectional	aged >20 yr

Blood

Pb measured in venous
whole blood using GFAAS

Age at measurement
Mean (SD):

Males

No MetS: 43.5 (0.23) yr
MetS: 48.7 (0.48) yr

Females

No MetS: 43.5 (0.25) yr
MetS: 51.4 (0.60) yr

Geometric Mean (SD):
Males

No MetS: 2.57 (0.02) pg/dL
MetS: 2.64 (0.04) pg/dL

Females

No MetS: 1.86 (0.01)

MetS: 1.92 (0.04) pg/dL

MetS, waist circumference
(cm), serum HDL (mg/dL),
serum triglycerides (mg/dL),
blood glucose (mg/dL)

MetS was defined as meeting
at least 3 of the following: (1)
elevated blood pressure (SBP
>130 mmHg or DBP
>85 mmHg or current use of
blood pressure medication),
(2) low HDL cholesterol
(<40 mg/dL in women or
<50 mg/dL in men), (3)
elevated serum triglycerides
(>150 mmHg) or current use
of antidyslipidemia
medication, (4) elevated
fasting plasma glucose levels,
(5) abdominal obesity (waist
circumference >90 cm in men
or >85 cm in women). Waist
circumference (cm) was
measured at the physical
examination by a trained
professional. Serum HDL
(pg/dL), triglycerides (mg/dL),
and blood glucose (mg/dL)
were measured in blood
samples obtained in the
morning following an
overnight fast.

Age at outcome is the same
as age at exposure
assessment

Age, BMI, residence
area, education level,
smoking and drinking
status, exercise, AST,
ALT

OR (95% CI) for outcomes
across blood Pb tertiles

MetS prevalence
T1 (<2.20 pg/dL):

Reference

T2 (2.20-3.01 pg/dL):
1.032 (0.788, 1.352)

T3 (>3.01 pg/dL):
0.817 (0.626, 1.065)

Waist circumference
(>85 cm)

T1
T2
T3

Reference
1.11 (0.83, 1.50)
1.11 (0.81, 1.51)

Serum HDL (<40 mg/dL)

T1
T2
T3

Reference
1.00 (0.80, 1.24)
0.76 (0.59, 0.97)

Serum triglycerides
(>150 mg/dL)
T1: Reference
T2: 1.13 (0.93, 1.39)
T3: 1.08 (0.87, 1.33)

Blood glucose

(>100 mg/dL)
T1: Reference
T2: 0.83 (0.68, 1.02)
T3: 1.04 (0.85, 1.28)

External Review Draft

9-93

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tLee and Kim (2013) KNHANES
n = 7,559

Korea

2005-2010	Korean adults

Cross-Sectional	aged >20 yr

Blood

Pb measured in venous
whole blood using GFAAS

Age at measurement
Mean (SD):

No MetS: 42.3 (0.29) yr
MetS: 48.4 (0.57) yr

Geometric Mean (SD):

No MetS: 2.734 (0.024) pg/dL
MetS: 2.957 (0.049) pg/dL

MetS, waist circumference,
serum HDL, serum
trigylcerides, blood glucose

MetS was defined as meeting
at least 3 of the following: (1)
elevated blood pressure (SBP
>130 mmHg or DBP
>85 mmHg or current use of
blood pressure medication),
(2) low HDL cholesterol
(<40 mg/dL in women or
<50 mg/dL in men), (3)
elevated serum triglycerides
(>150 mmHg) or current use
of antidyslipidemia
medication, (4) elevated
fasting plasma glucose levels,
(5) abdominal obesity (waist
circumference >90 cm in men
or >85 cm in women). Waist
circumference (cm) was
measured at the physical
examination by a trained
professional. Serum HDL
(pg/dL), triglycerides (mg/dL),
and blood glucose (mg/dL)
were measured in blood
samples obtained in the
morning following an
overnight fast.

Age at outcome is the same
as age at exposure
assessment

Age, BMI, residence
area, education level,
smoking and drinking
status, exercise,
serum aspartate
aminotransferase,
serm alanine
aminotransferase

OR (95% CI) for outcomes
across blood Pb tertiles

MetS Prevalence
T1 (<2.362 pg/dL):

Reference

T2 (>2.362-3.282 pg/dL):
1.267 (0.950, 1.690)

T3 (>3.282 pg/dL:
0.984 (0.735, 1.317)

Waist circumference
(>85 cm)

T1
T2
T3

Reference
1.04 (0.75, 1.45)
0.89 (0.64, 1.24)

Serum HDL (<40 mg/dL)
T1: Reference

T2
T3

0.98 (0.79, 1.23)
0.96 (0.77, 1.20)

Serum triglycerides
(>150 mg/dL)
T1: Reference

T2
T3

1.01 (0.82, 1.24)
1.07 (0.87, 1.32)

Blood glucose
(>100 mg/dL)
T1: Reference
T2: 1.00 (0.81, 1.24)
T3: 1.14 (0.91, 1.44)

External Review Draft

9-94

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tWana etal. (2018c) NHANES
n = 9537

United States

2003-2014

Cross-Sectional

NHANES
participants aged
20+,2003-2014

Blood

Pb was measured in venous
whole blood using ICP-DRC-
MS

Age at measurement
Mean (SD): 49.2 (18.0) yr

Geometric mean (SD):
1.32 (2.00) [jg/dL

Waist circumference (cm)

Waist circumference (cm)
was measured during minimal
respiration to the nearest 0.1
cm at the level of the iliac
crest at the time of NHANES
physical examination.

Age at outcome:

Mean (SD): 49.2 (18.0) yr

Age, sex,
race/ethnicity,
education, smoking
status, physical
activity, NHANES
cycle, and urinary
creatinine

Change in waist
circumference (cm) per 1-
SD increase in log(10)-
transformed Pb (SD NR):

0.008 (-0.010, -0.006)

tPeters et al. (2012) Normative Aging Blood, Bone

Serum lipids

United States
Blood Pb measured
between 1999-2008;
Serum lipids
measured 3 to 4 yr
after blood Pb
Cohort

Study
n = 426

Older male
Veterans

Blood Pb measured in venous Triglycerides, HDL-C
whole blood using GFAAS

Age at outcome:

Mean: 4.01 ± 2.30 [jg/dL 3 to 4 yr after blood Pb

Age at baseline, yr
between baseline and
outcome, education,
BMI, alcohol intake,
smoking status, pack-
yr of smoking,
hypertension status,
and statin use

ORs

Low HDL-C (<40 mg/dL):

0.899 (0.804, 1.004)

High Triglycerides
(>200 mg/dL):

0.993 (0.874, 1.129)

tEttinqer et al. (2014)

Kumasi, Ghana; Cape
Town, South Africa;
Victoria, Seychelles;
Kingston, Jamaica;
Maywood, Illinois
(United States)
2010-2014
Prospective Cohort

Modeling the
Epidemiologic
Transition Study
(METS)
n = 150

Adults of African
descent from 5
countries of
varying social
and economic
development

Blood

Pb was measured in venous
whole blood using DRC-ICP-
MS

Age at measurement
Mean (SD):

Males: 34.7 (6.0) yr
Females: 35.2 (6.2) yr

Geometric Mean (95% CI):
1.55 (1.30, 1.85) [jg/dL

Waist Circumference >94 cm
(males) or >80 cm (females),
Fasting Glucose >100 mg/dL

Fasting glucose was
measured in blood. Further
outcome assessment details
not provided.

Age at outcome is the same
as age at exposure
assessment

Age, sex, site
location, marital
status, education,
paid employment,
alcohol use, fish
intake, percent body
fat

ORs for blood Pb above
the median (1.66 (jg/dL)
vs below the median

Waist Circumference

[>94 cm (m) or >80 cm (f)]
4.53 (1.06, 19.48)

Fasting Glucose
(>100 mg/dL)

4.99 (1.97, 12.69)

Median (95% CI):
1.66 (1.34, 1.93) [jg/dL

External Review Draft

9-95

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

75th: 2.6 pg/dL
Max: 31.82 pg/dL

Body Weight

tWana etal. (2018a)

China
2014

Cross-Sectional

SPECT-China
n = 3922

Chinese citizens
aged >18 yr who
had lived in their
current area for
6+ mo

Blood

Pb was measured in venous
whole blood using GFAAS

Age at measurement:

Mean (SD):

Normal weight subjects:
50.9 (13.9) yr

Overweight subjects:
54.0 (12.3) yr

Obese subjects:

56.2 (11.2) yr

Median

(25th-75th percentiles)

Normal weight:

3.9 (2.6, 5.6) pg/dL

Overweight subjects:

4.3	(2.9, 6.1) pg/dL

Obese subjects:

4.4	(2.7, 6.2) pg/dL

BMI (kg/m2)

BMI was calculated as weight
(kg) divided by squared
height (m2). Overweight
(including obese) was defined
as BMI >25 kg/m2.

Age at outcome is the same
as age at exposure
assessment

Age, sex, economic
status, rural/urban
residence, current
smoking, diabetes,
hypertension,
dyslipidemia

OR (95% CI) for
overweight or obese (BMI
>25 kg/m2) across blood
Pb quartiles

Q1 (<2.69 pg/dL):
Reference

Q2 (2.69-4.01 pg/dL):
1.09 (0.89, 1.33)

Q3 (4.01-5.60 pg/dL):
1.15 (0.94, 1.40)

Q4 (>5.60 pg/dL):
1.40 (1.14, 1.71)

External Review Draft

9-96

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tEttinaer et al. (2014)

Kumasi, Ghana; Cape
Town, South Africa;
Victoria, Seychelles;
Kingston, Jamaica;
Maywood, Illinois
(United States)
2010-2014
Prospective Cohort

Modeling the
Epidemiologic
Transition Study
(METS)
n = 150

Adults of African
descent from 5
countries of
varying social
and economic
development

Blood

Pb was measured in venous
whole blood using DRC-ICP-
MS

Age at measurement
Mean (SD):

Males: 34.7 (6.0) yr
Females: 35.2 (6.2) yr

Geometric Mean (95% CI):
1.55 (1.30, 1.85) [jg/dL

Overweight (BMI >25), Obese Age, sex, site

(BMI >30)

Height and weight were
measured by physical
examination.

Age at outcome is the same
as age at exposure
assessment

location, marital
status, education,
paid employment,
alcohol use, fish
intake, percent body
fat

ORs for blood Pb above
the median (1.66 (jg/dL)
vs below the median

Overweight (BMI >25)
0.88 (0.31, 2.51)

Obese (BMI >30)
2.70 (0.75, 9.75)

Median (95% CI):
1.66 (1.34, 1.93) [jg/dL

75th: 2.6 pg/dL
Max: 31.82 pg/dL

External Review Draft

9-97

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study
Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
CIs

tGuoetal. (2019) N = 145

China
2015

Cross-Sectional

Blood

BMI (kg/m2)

Adult men	Pb was measured using ICP- Age outcome

recruited through	MS	Mean (SD): 39 (12) yr

a physical

examination	Age measurement

center	Mean (SD): 39(12) yr

Age

Change in BMI (kg/m2)
per log increase in blood
Pb:

0.05 (-3.64, 3.74)

Mean (SD): 8.5 (3.8) pg/dL;
Median: 7.9 pg/dL
75th: 10.8 pg/dL
Max: 28.2 pg/dL

ALT = alanine aminotransferase; AST = aspartate aminotransferase; BMI = body mass index; CI = confidence interval; DBP = diastolic blood pressure; DRC-ICP-MS = dynamic
reaction cell for inductively coupled plasma mass spectrometry; eGFR = estimated glomerular filtration rate; ELEMENT = Early Life Exposures in Mexico to Environmental
Toxicants; GADA = glutamic acid decarboxylase antibodies; GFAAS = graphite furnace atomic absorption spectrometry; GM = geometric mean; HDL = high-density lipoprotein;
HDL-C = high-density lipoprotein cholesterol; HOMA-IR = Homeostatic Model Assessment for Insulin Resistance; HOMA- (B = HOMA of (B-cell function; ICP-MS = inductively
coupled plasma mass spectrometry; KNHANES = Korean National Health and Nutrition Examination Survey; MetS = metabolic syndrome; METS = Modeling the Epidemiologic
Transition Study; NR = not reported; OR = odds ratio; Pb = lead; SBP = systolic blood pressure; SD = standard deviation; SPECT = single photon emission computed
tomography; TC = total cholesterol; Q = quartile.

aEffect estimates are standardized to a 1 |jg/dL increase in BLL or a 10 |jg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.
fStudies published since the 2013 Integrated Science Assessment for Lead.

1

External Review Draft

9-98

DRAFT: Do not cite or quote

-------
Table 9-7 Animal toxicological studies of exposure to Pb and metabolic effects.

Study Species ^Stock/Strain), T|m|ngof Exposure BLL As Reported (pg/dL)	Endpoints Examined

'	Exposure

Faulk et al.

Mouse (Agouti), 0.0 ppm Mo 3, 6, 9

Oral, drinking

Mean maternal BLL,

Oxygen Consumption, CO2 Production, Food Intake, Body

(2014)

Pb, M/F, n = 30
2.1 ppm Pb, M/F, n = 28
16 ppm Pb, M/F, n = 33
32 ppm Pb, M/F, n = 29

(Longitudinal phenotypic
measures were taken
from a total of 120 a/a
mice, on average 2.7
mice per litter)

water

tested at weaning, were
below the LOD for the
control group, and 4.1
(61.3) [jg/dL, 25.1 (67.3)
[jg/dL, and 32.1 (611.4)
[jg/dL in the three
exposure groups,
2.1 ppm, 16 ppm, and
32 ppm, respectively

Weight, Body Fat

Rahman et al.

Rat (Wistar) PND21.30

Oral, drinking

2.2 ± 0.07 [jg/dL for 0%

Serum 25(OH)D, Serum 1,25(OH)2D, Hepatic 25-

(2018)

0% Pb Acetate, M/F,

water

Pb Acetate,

Hydroxylase Protein Levels, Hepatic 25-Hydroxylase



n = 37



12.4 ± 3.3 [jg/dL for 0.2%

Immunohistochemistry



0.2% Pb Acetate, M/F,



Pb Acetate - PND 21





n = 38



3.3 ± 1.7 [jg/dL for 0% Pb







Acetate, 22.7 ± 6.0 [jg/dL
for 0.2% Pb Acetate -
PND 30



External Review Draft	9-99	DRAFT: Do not cite or quote

-------
Study

Species (Stock/Strain),
n, Sex

Timing of
Exposure

Exposure
Details

BLL As Reported (pg/dL)

Endpoints Examined

Zhou et al. Rat (Sprague Dawley),
(2018)	0%Pb Acetate, M,

n = 20

0.5% Pb Acetate, M,
n = 20

1% Pb Acetate, M,
n = 20

2% Pb Acetate, M,
n = 20

PND 52	Oral, drinking 11.4 |jg/L for 0%

water	147 |jg/L for 0.5%

226 |jg/L for 1 %
289 |jg/L for 2%

Cholesterol Content, mRNA level of SREBP2 in the
cortex, mRNA level of SREBP2 in the hippocampus,
mRNA level of LDL-R in the cortex, mRNA level of HMG-
CR in the hippocampus, mRNA level of HMG-CR in the
cortex, protein level of SREBP2 in the cortex, mRNA level
of LDL-R in the hippocampus, protein level of HMG-CR in
the cortex, protein level of LDL-R in the cortex, protein
level of SREBP2 in the hippocampus, protein level of
HMG-CR in the hippocampus, protein level of LDL-R in
the hippocampus, immunohistochemistry ofSREBP2 in
the cortex, immunohistochemistry of HMG-CR in the
cortex, immunohistochemistry of LDL-R in the cortex,
immunohistochemistry of SREBP2 in the hippocampus,
immunohistochemistry of HMG-CR LDL-R in the
hippocampus, immunohistochemistry of LDL-R in the
hippocampus, mRNA level of LXR-a in the cortex, mRNA
level of ABCA1 in the cortex, mRNA level of LXR-a in the
hippocampus, mRNA level of ABCA1 in the hippocampus,
protein level of LXR-a in the cortex, protein level of
ABCA1 in the cortex, protein level of LXR-a in the
hippocampus, protein level of ABCA1 in the hippocampus

ABCA1 = ATP-binding cassette transporter ABCA1 (member 1 of human transporter sub-family ABCA); BLL = blood lead level; C02 = carbon dioxide; F = female; HMG-CR = 3-
Hydroxy-3-Methylglutaryl-Coenzyme A Reductase; LDL-R = low-density lipoprotein receptor; LOD = limit of detection; LXR-a = liver X receptor alpha; M = male; mRNA = messenger
ribonucleic acid; Pb = lead; PND = postnatal day; SREBP2 = Sterol Regulatory Element Binding Transcription Factor 2.

External Review Draft

9-100

DRAFT: Do not cite or quote

-------
Table 9-8 Animal toxicological studies of exposure to Pb and gastrointestinal effects.

Study	Species (Stoc^k/Strain), Timing of Exposure Exposure Details	BLL ^gftfL)Ort0d Endpoints Examined

Kosik-Boaacka et al. Rat (Wistar), Control Day 270	Oral, drinking water 0.34±0.23 [jg/dL for transepithelial electrical

(2011)	(distilled water), M, n = 9	0.0%, 7.21 ± 1.27 pg/dL potential difference (PD),

0.1% Pb, M, n = 9	for 0.1%	changes in the

transepithelial electrical
potential difference
during mechanical
stimulation (dPD),
transepithelial electrical
resistance (R)

Reddv et al. (2018)

Rat (Sprague Dawley),

Microbiome Counts at

Oral, gavage

2.3 ± 1.16 pg/dL-CD, M

Fecal Lactobacilli



Control Diet (CD), M,

Week 0, 4, 8, 10, 12



19.3±6.23 pg/dL -

(Counts), Fecal E. Coli



n = 10

BLL at End of Week 8



CD + Pb, M

(Counts), Fecal Yeast



Control Diet, F, n = 10





2.5 ± 0.89 pg/dL- ID, M

(Counts)



Iron Deficient (ID), M,





47.5 ± 3.78 pg/dL-





n = 10





ID + Pb, M





Iron Deficient, F, n = 10





1.9 ±0.81 pg/dL-CD, F





Control Diet + Pb, M,





13.5 ± 3.52 pg/dL-





n = 10





CD + Pb, F





Control Diet + Pb, F,





1.5 ± 0.31 pg/dL - ID, F





n = 10





29.80 ± 8.30 pg/dL-





Iron Deficient + Pb, M,





ID + Pb, F





n = 10











Iron Deficient, F, n = 10









BLL = blood lead level; dPD = transepithelial electrical potential difference during mechanical stimulation; F = female; M = male; PD = transepithelial electrical potential difference;
R = resistance

External Review Draft

9-101

DRAFT: Do not cite or quote

-------
Table 9-9 Epidemiologic studies of exposure to Pb and endocrine effects.

Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tChen etal. (2013) NHANES
n = 5,418

United States

2007-2008
Cross-sectional

Adolescents and adults in
the general U.S.
population who had no
reported thyroid diseases,
thyroid medications,
pregnancy, and sex
steroid medications.

Blood Pb

Blood Pb was
measured in venous
whole blood using
GFAAS

Age at measurement:
>12 yr old

Mean: 0.93 [jg/dL
Max: 9.20 [jg/dL

TSH, thyroglobulin (Tg), Age, sex, race/ethnicity,

and thyroid hormones (T3,
FT3, T4, FT4)

TSH and thyroid
hormones measured in
serum using the Beckman
Immunoassay System.

Age at outcome:

>12 yr old

creatinine-adjusted
urinary iodine, BMI Z-
score, and serum
cotinine level

Change in T4 ([jg/dL)b

Adolescents (12-19 yr old)
(-0.02, 0.04)

Adults (>19 yr old)
-0.01 (-0.02, 0.01)
Change in FT4 (ng/dL)b
Adolescents (12-19 yr old)
(-0.01, 0.04)

Adults (>19 yr old)
0.01 (-0.01, 0.02)

Change in T3 (ng/dL)b
Adolescents (12-19 yr old)
(-0.01, 0.04)

Adults (>19 yr old)
-0.0004 (-0.02, 0.02)
Change in FT3 (pg/ml_)b
Adolescents (12-19 yr old)
(-0.002, 0.04)

Adults (>19 yr old)
0.01 (-0.001, 0.02)
Change in TSH ([jlU/mL)b
Adolescents (12-19 yr old)
-0.05 (-0.18, 0.07)

Adults (>19 yr old)
-0.01 (-0.06, 0.04)

External Review Draft

9-102

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

Change in Tg (ng/ml_)b

Adolescents (12-19 yr old)
0.05 (-0.13, 0.24)

Adults (>19yrotd)
0.01 (-0.03, 0.06)

tKrieq (2019)

United States

1988-1994

Cross-sectional

NHANES III
n = 16,573

General population,
>20 yr old

Blood Pb

Blood Pb was
measured in venous
whole blood using
AAS

Age at measurement:
>20 yr old

Mean: 3.55 [jg/dL
(SE = 0.10)

TSH and T4

TSH and thyroid
hormones measured in
serum using the Beckman
Immunoassay System.

Age at outcome:

>20 yr old

Linear regression model
adjusted for race-
ethnicity, sex, age,
session, BMI, pregnant,
menopause, hormone
pill use, vaginal cream
use, hormone patch
use, urinary creatinine

Change in TSH (%)

-1.2 (-5.6, 3.3)

Change in T4 (%)

-38.9 (-51.3, -23.4)

Change in Logio-TSH
(HU/mL)b

Male

0.01 (-0.04, 0.05)
Female (Not pregnant)
-0.04 (-0.08, 0.01)
Female (Pregnant)
-0.03 (-0.26, 0.20)

Change in Logio-T4 ([jg/dL)b

Male

-0.15 (-0.48, 0.18)

Female (Not pregnant)
-0.52 (-0.83, -0.21)

Female (Pregnant)
-2.01 (-3.09, -0.93)

External Review Draft

9-103

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tMendv et al. (2013) NHANES
n = 4,652

United States

2007-2008

Cross-sectional

General population >20 yr
old, excluding pregnant
women, individuals with a
history of thyroid disease,
or under treatment for
thyroid dysfunction

Blood Pb

Blood Pb was
measured in venous
whole blood using
GFAAS

Age at measurement:
>20 yr old

Mean (SD):
1.52 ± 1.20 [jg/dL
Max: 33.12 pg/dL

TSH and thyroid
hormones (T3, FT3, T4,
FT4)

TSH and thyroid
hormones measured in
serum using the Beckman
Immunoassay System

Age at outcome:

>20 yr old

Age, gender,
race/ethnicity, smoking,
alcohol consumption,
BMI, physical activity,
iodine intake,
medications, and bone
mineral density

Change in T3 (ng/dL)

-0.774 (-2.269, 0.722)
Change in FT3 (pg/mL)
0.015 (-0.007, 0.037)
Change in T4 ((jg/dL)
-0.162 (-0.321, -0.004)
Change in FT4 (ng/mL)
(-0.011, 0.011)

Change in TSH (mlll/mL)

0.015 (-0.088, 0.118)

tChristensen (2012)

United States

2007-2008

Cross-sectional

NHANES
n = 1,587

General population,
>20 yr old, excluding
individuals with thyroid
disease or cancer, or
were taking thyroid
medications

Blood Pb

Blood Pb was
measured in venous
whole blood using
GFAAS

Age at measurement:
>20 yr old

Median: 1.3 pg/dL
75th: 2.1 pg/dL

TSH and thyroid
hormones (T3, T4)

TSH and thyroid
hormones measured in
serum using the Beckman
Immunoassay System.

Age at outcome:

>20 yr old

Age, sex, race, BMI,
serum lipids, serum
cotinine, pregnancy and
menopausal status, and
use of selected
medications

Change in ln(T3) (ng/dL)

0.004 (-0.016, 0.023)

Change in ln(FT3) (pg/mL)

0.008 (-0.002, 0.017)

Change in ln(T4) ((jg/dL)

-0.018 (-0.036, 0)

Change in ln(FT4) (pg/mL)

-0.001 (-0.018, 0.015)

Change in In(TSH) (mlU/L)

0.027 (-0.031, 0.085)

External Review Draft

9-104

DRAFT: Do not cite or quote

-------
tLuo and Hendrvx

(2014)

NHANES
n = 6,231

Blood Pb

United States

2007-2010

Cross-sectional

General population >20 yr
old, excluding pregnant
women, individuals with
history of thyroid disease,
or missing data.

Blood Pb was
measured in venous
whole blood using
GFAAS

Age at measurement:
>20 yr old

Mean: 1.82 [jg/dL
Max: 33.10 [jg/dL

External Review Draft

TSH, thyroglobulin (Tg),
and thyroid hormones (T3,
FT3, T4, FT4)

TSH and thyroid
hormones measured in
serum using the Beckman
Immunoassay System.

Age at outcome:

>20 yr old

Adjusted for age, sex,
race and ethnicity,
serum cotinine, BMI,
and creatinine-adjusted
urinary iodine

Change in T3 across tertiles
(ng/dL)b

T1: Reference

T2: 1.02 (-0.90, 2.94)

T3: 0.69 (-2.37, 3.76)

Women Only

T1: Reference

T2: -0.36 (-3.72, 3.00)

T3: 0.61 (-5.02, 6.23)

Men Only

T1: Reference

T2: 1.96 (-0.98, 4.91)

T3: 0.69 (-2.59, 3.97)

Change in FT3 across
tertiles (pg/ml_)b:

T1: Reference

T2: 0.03 (0.001, 0.07)

T3: 0.04 (0.01, 0.08)

Women Only

T1: Reference

T2: 0.02 (-0.04, 0.08)

T3: 0.03 (-0.04, 0.11)

Men Only

T1: Reference

T2: 0.03 (-0.01, 0.07)

T3: 0.05 (0.01, 0.09)

Change in T4 across tertiles
(jjg/dL)b:

T1: Reference

T2: 0.01 (-0.16, 0.14)

T3: -0.09 (-0.28, 0.11

Women Only

9-105

DRAFT: Do not cite or quote

-------
External Review Draft

T1: Reference

T2: 0.12 (-0.10, 0.35)

T3: 0.02 (-0.29, 0.33)

Men Only

T1: Reference

T2: -0.14 (-0.35, 0.08)

T3: -0.20 (-0.40, 0.01)

Change in FT4 across
tertiles (ng/dL)b:

T1: Reference

T2: 0.007 (-0.01, 0.02)

T3: 0.002 (-0.01, 0.01)

Women Only

T1: Reference

T2: 0.02 (0.01, 0.04)

T3: 0.02 (-0.003, 0.04)

Men Only

T1: Reference

T2: -0.02 (-0.03, 0.005)

T3: -0.01 (-0.04, 0.008)

Change in Log-Tg across
tertiles (ng/ml_)b:

T1: Reference

T2: 0.04 (-0.04, 0.13)

T3: 0.02 (-0.07, 0.12)

Women Only

T1: Reference

T2: 0.08 (-0.03, 0.19)

T3: -0.06 (-0.19, 0.08)

Men Only

T1: Reference

T2: -0.001 (-0.13, 0.17)

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

T3: 0.05 (-0.08, 0.17)

Change in Log-TSH across
tertiles (ulll/mL)b:

T1: Reference

T2: 0.01 (-0.05, 0.07)

T3: 0.02 (-0.06, 0.09)

Women Only

T1: Reference

T2: 0.05 (-0.06, 0.16)

T3: 0.02 (-0.09, 0.14)

Men Only

T1: Reference

T2: -0.04 (-0.13, 0.06)

T3: -0.02 (-0.11, 0.07)

External Review Draft

9-107

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tNie etal. (2017)

Shanghai and 7

provinces

China

2014

Cross-sectional

SPECT-China study
n = 5,628

Residents of these
regions are 99.5% Han
Chinese.

Exclusion criteria
included age under 18 yr
old, less than 6 mo spent
at current residence, and
severe communication
problems or acute illness
(thyroid resection or
iodine-131 therapy,
malignant tumor,
subacute thyroiditis, liver
cirrhosis)

Blood Pb

Whole blood
measured using AAS
Age at measurement:
18-93 yr old

Median:

Men: 44.00 |jg/L
Women: 37.87 |jg/L

Mean:

Men:

29.00±62.18 |jg/L
Women:

25.03±54.61 |jg/L

TSH, thyroid hormones
(T3, T4), thyroid
peroxidade antibody
(TPOAb) and thyroglobulin
antibodies (TGAb)

Thyroid dysfunction and
subclinical thyroid
dysfunction were
measured by

immunochemiluminometric
assays

Age at outcome:

18-93 yr old

Linear and logistic
regression model
adjusted for age, BMI
smoking status (men
only) and drinking status Women

1.41 (0.00, 2.84)

Change in TPOAb (%)

Men

0.50 (-0.80, 1.82)

Change in TGAb (%)

Men

-0.60 (-1.88, 0.70)
Women

0.20 (-1.09, 1.51)

Change in TSH (%)

Men

-0.40 (-1.29, 0.40)
Women

1.11 (0.30, 1.82)

tKahn etal. (2014)

Pristina and Mitrovica
Yugoslavia
1985-1986
Cross-sectional

Yugoslavia Prospective
Study of Environmental
Lead Exposure

n = 291

Pregnant women in
second trimester, major
central nervous system
defects, multiple births,
and residence >10 km
from clinic

Blood Pb

Whole blood samples
taken in Yugoslavia
and transported on
wet ice to Columbia
University. Blood Pb
measured using
GFAAS.

Age at measurement:
16-41 yr old

Mean [jg/dL (SD):
Pristina: 5.57 (2.01)
Mitrovica: 20.00 (6.99)

TSH, thyroid hormones
(FT4), and thyroid
peroxidase antibodies
(TPOAb)

FT4 and TPOAb

were measured by a
radioimmunoassay
procedure. TSH was
measured using an
immunoradiometric assay

Age at outcome:

16-41 yr old

Logistic regression
model adjusted for:

FT4: height, ethnicity,
BMI, fetal gestational
age, maternal
education, adults per
room; TSH: hemoglobin,
ethnicity, BMI, fetal
gestational age,
maternal age; TPOAb:
ethnicity, fetal
gestational age,
maternal age, adults per
room.

Change in FT4 (ng/dL)b

-0.074 (-0.10, -0.046)

Change in Log-TSH
(HlU/mL)b

0.026 (-0.065, 0.12)

Change in Log-TPOAb
(IU/mL)b

0.31 (0.17, 0.46)

ORb

TPOAb >vs. <10 lU/mL
2.41 (1.53, 3.82)

External Review Draft

9-108

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tSouza-Talarico et al.
(2017)

Sao Paulo
Brazil

Cross-sectional

N = 126

105 women and 21 men
ages 50-82 yr old with a
mean of 9.8 (±4.5) yr of
education

Blood Pb

Fasting blood Pb was
measured using ICP-
MS

Age at measurement:
50-82 yr old

Median: 2.1 [jg/dL
(SD: ±0.9)
Max: 6.1 [jg/dL

Cortisol concentration and
allostatic load

Six neuroendocrine,
metabolic, and
anthropometric biomarkers
were analyzed, and values
were transformed into an
AL index using clinical
reference cut-offs. Salivary
samples were collected at
home over 2 d at
awakening, 30-min after
waking, afternoon, and
evening periods to
determine Cortisol levels.

Age, gender, time of
awakening,
socioeconomic status
(SES), GDS, and PSS
scores

Change in CAR (pg/dL min)b

0.791 (0.672, 1.073)

Change in total AUC (jjg/dL
hr)b

0.889 (0.829, 0.953)

Age at outcome:
50-82 yr old

External Review Draft

9-109

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure
Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tNaueta et al. (2018) Study of Genetics, Stress Blood Pb

Montreal
Canada
2004-2006
Cross-sectional

and Cognitive
Development
n = 65

75% of participants were
women, 95% were
Caucasian, 90% were
current smokers

Blood Pb levels were
determined using
inductively coupled
plasma mass
spectroscopy
Age at measurement:
50-67 yr old

Median: 2.48 [jg/dL
Mean: 2.41 [jg/dL
(SD = 0.15)

Diurnal basal Cortisol
levels and acute Cortisol
responsivity

Basal Cortisol:
Participants were
instructed to collect saliva
five times per day during
three consecutive
weekdays: upon
awakening, 30 min after
awakening, at 2:00 p.m.,
at 4:00 p.m., and at
bedtime

Linear model adjusted
for age, gender, waist-
hip ratio, smoking status
and income levels.

Change in basal Cortisol
levels (|jg/dL)

-0.01 (-0.05, 0.02)

Change in reactive Cortisol
levels (|jg/dL)

-0.01 (-0.03 0.01)

Stress reactivity: A total of
nine saliva samples were
collected for measurement
of salivary Cortisol: two
baseline samples, one
postanticipatory, and six
post-TSST tasks: one after
15 min and then five
sampled every 10 min

Age at outcome:

50-67 yr old

AAS = atomic absorption spectrometry; BMI = body mass index; CAR = Cortisol awakening response; CI = confidence interval; d = day(s); GFAAS = graphite furnace atomic
absorption spectrometry; FT3 = free triiodothyronine; FT4 = free thyroxine; ICP-MS = inductively coupled plasma mass spectrometry; NHANES = National Health and Nutrition
Examination Survey; Pb = lead; SD = standard deviation; SE = standard error; SES = socioeconomic status; SPECT = single photon emission computed tomography;
Tg = thyroglobulin; T = fertile; TGAb = thyroglobulin antibodies; TPOAb = thyroid peroxidade antibody; TSH = thyroid stimulating hormone; yr = year(s)

aEffect estimates are standardized to a 1 |jg/dL increase in BLL or a 10 |jg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.
bEffect estimate unable to be standardized due to insufficient distribution information.
fStudies published since the 2013 Integrated Science Assessment for Lead.

External Review Draft

9-110

DRAFT: Do not cite or quote

-------
Table 9-10 Animal toxicological studies of exposure to Pb and endocrine effects.

Study

Species (Stock/Strain), n, Sex

Timing of
Exposure

Exposure Details
(Concentration,
Duration)

BLL As Reported (jjg/dL)b Corticosterone Levels

Rossi-George et al.
(2011)

Rat (Long-Evans)	GD-61 to

Control (untreated), M/F, n = 10 P^D 304
dams

50 ppm, M/F, n = 9 dams
150 ppm, M/F, n = 11 dams

Dams were dosed
starting 2 mo prior
to mating through
lactation. Pups
were weaned on
PND 21 and
continued on the
regimen of their
dam until
euthanasia post-
testing at
approximately
10 mo of age.

0.979 |jg/dL for 0 ppm,
19.091 |jg/dL for 50 ppm,
35.245 [jg/dL for 150 ppm ¦
PND 21 Females

I.469	|jg/dL for 0 ppm,

II.259	|jg/dLfor50 ppm,
25.699 [jg/dL for 150 ppm ¦
PND 61 Females

I.713	|jg/dL for 0 ppm,

II.993	|jg/dLfor50 ppm,
29.615 [jg/dL for 150 ppm ¦
PND 304 Females

Adrenal Weight,
Corticosterone Levels

1.935 [jg/dL for 0 ppm,
19.597 [jg/dL for 50 ppm,
31.935 [jg/dL for 150 ppm-
PND 21 Males

2.177 |jg/dL for 0 ppm,
12.581 |jg/dL for 50 ppm,
26.855 [jg/dL for 150 ppm -
PND 61 Males

1.694 |jg/dL for 0 ppm,
15.968 [jg/dL for 50 ppm,
29.274 |jg/dL for 150 ppm -
PND 304 Males

Graham et al.

Rat (Sprague Dawley)

PND 4 to

Rats were gavaged

0.267 [jg/dL for 0 mg/kg,

Adrenal Weight,

(2011)

Control (vehicle), M/F,

PND 28

every other day

3.27 |jg/dL for 1 mg/kg

Corticosterone Levels





from P4 until P28.





n = 12-18 (6-8/6-8)



12.5 |jg/dL for 10 mg/kg -
PND 29



External Review Draft

9-111

DRAFT: Do not cite or quote

-------
Study

Species (Stock/Strain), n, Sex

Timing of
Exposure

Exposure Details

(Concentration,
Duration)

BLL As Reported (jjg/dL)b Corticosterone Levels



1 mg/kg Pb, M/F, n = 12-18









(6-8/6-8)









10 mg/kg Pb, M/F, n = 12-18









(6-8/6-8)







Corv-Slechta et al.
(2013)

Mouse (C57BL.6)	GD -61 to

Control (untreated), M, n = 8-17 PND 365

Control (untreated), F, n = 8-13

100 ppm Pb, M, n = 8-17

100 ppm Pb, F, n = 8-13

Dams were
exposed starting
2 mo prior to
mating. Offspring
were continued on
the same exposure
as their dams until
the end of the
experiment at
12 mo of age.

0.34 [jg/dL for 0 ppm Fl males

0.11 [jg/dL for 0 ppm FS males

0.34 [jg/dL for 0 ppm Fl females

0.16 [jg/dL for 0 ppm FS
females

6.94 |jg/dL for 100 ppm Fl
males

6.16 [jg/dL for 100 ppm FS
males

9.38 |jg/dL for 100 ppm Fl
females

7.07 |jg/dL for 100 ppm FS
females

Adrenal Weight,
Corticosterone Levels

Amos-Kroohs et al.
(2016)

Rat (Sprague Dawley)

Control (vehicle, see notes),
M/F, n = 16 (8/8)

1 mg/kg Pb, M/F, n = 16 (8/8)
10 mg/kg Pb, M/F, n = 16 (8/8)

P4 until P10,
18, or 28.

Rats were gavaged
every other day
from PND4 until
PND10, 18, or 28.

1.24 |jg/dL for 0 mg/kg Pb
2.79 |jg/dL for 1 mg/kg Pb
9.07 [jg/dL for 10 mg/kg Pb

Corticosterone Levels

Sobolewski et al.
(2020)

Mouse (C57BL.6)

F0 Control (assume untreated),

F, n = 10

100 ppm Pb, F, n = 10

20 females were in control and
20 received Pb but these groups
were further divided, and some
received prenatal stress and
others did not.

GD -61 to
PND 21

Exposure started 2 F1 0.0 [jg/dL for Control

mo prior to mating
and continued
through PND 21
(weaning) of the
F1.

F3 was technically
not directly
exposed.

12.5 |jg/dL for 100 ppm -
PND 6-7

F3 0.0 [jg/dL for Control

Corticosterone Levels

F# = filial generation; F = female; GD = gestational day; M = male; mo = month(s); Pb = lead; PND = postnatal day.

External Review Draft

9-112

DRAFT: Do not cite or quote

-------
Table 9-11 Epidemiologic studies of exposure to Pb and musculoskeletal effects.

Reference and
Study Design

Study Population

Exposure Assessment Outcome

Confounders

Effect Estimates and
95% Clsa

Osteoporosis and Bone Mineral Density

tChoetal. (2012)

South Korea
2008

Cross-Sectional

KNHANES
n = 481

Postmenopausal women

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:

Mean (SD):

Q1: 64.03 (8.52) yr
Q2 and Q3: NR
Q4: 61.78 (8.62) yr

Median: 2.32 [jg/dL
25th: 1.83 pg/dL
75th: 2.88 pg/dL

Osteoporosis

BMD measured in hip,
neck, and spine using X-
ray absorptiometry.
Osteoporosis defined as
T-score <2.5 at any of
the measurement sites

Age at outcome is the
same as the age at
exposure assessment

Age, BMI, alcohol
intake, cigarette
smoking, exercise,
use of oral
contraceptive pill,
hormone therapy,
caloric intake,
calcium intake, fish
consumption, and
vitamin D level

OR Osteoporosis
Prevalence

Q1
Q2
Q3
Q4

Ref.

1.41 (0.75,
1.34 (0.70,
1.50 (0.79,

2.67)
2.56)
2.86)

tWana et al. (2019)

United States

2013-2014

Cross-sectional

NHANES	Blood, Urine

n = 1859

Blood Pb measured in whole
General population; >40 yr blood using ICP-MS
old

Age at measurement:

>40 yr

Mean: 1.24 pg/dL
75th: 1.81 pg/dL

BMD and fracture risk

BMD measured via DXA
scan; Fracture risk
measured via Fracture
Risk Assessment score -
a composite index of
fracture risk factors

Age at outcome:

>40 yr

Age, race/ethnicity,
BMI, serum 25(OH)D
level, smoking,
drinking, treatment
for osteoporosis, and ^a/es
use of prednisone

Change in BMD (g/cm2)

Femur

-0.01 (-0.03, 0.01)

Premenopausal Women
-0.06 (-0.08, -0.03)

Menopausal Women
0.01 (-0.01, 0.03)

Spine

External Review Draft

9-113

DRAFT: Do not cite or quote

-------
StuedynDesfgnn	Study Population	Exposure Assessment	Outcome	Confounders Eff8Ct95% cis^ ^

Males

0.01 (-0.01, 0.03)

Premenopausal Women
-0.05 (-0.08, -0.02)

Menopausal Women
0.02 (-0.01, 0.04)

Change in 10-yr
Fracture Risk Score

Hip

0.45 (0.28, 0.62)
Major

1.22 (0.68, 1.77)

tLee and Kim
(2012)

South Korea

2008-2009

Cross-Sectional

KNHANES
n = 832

Women ages >40 yr

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:

Mean (SD): 56.1 (10.4) yr

GM: 2.182 pg/dL

BMD

BMD in the femoral neck,
trochanter,
intertrochanter, Ward
triangle, total femur, and
lumbar 1-4. Measured
using DXA

Age at outcome:

Mean (SD): 56.1 (10.4)
yr

Residence area,
obesity, educational
level, smoking status,
drinking status,
number of
pregnancies,
hormone treatment,
contraceptive oral pill
and daily calcium
intake for pre- and
postmenopausal, and
time since
menopause for
postmenopausal

Change in BMD (g/cm2)

Premenopausal Women

Total Femur
-0.15 (-0.33, 0.03)

Trochanter
-0.18 (-0.41, 0.05)

Intertrochanter
-0.11 (-0.25, 0.03)

Femoral Neck
-0.11 (-0.28, 0.07)

External Review Draft

9-114

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

Ward's Triangle
-0.11 (-0.26, 0.03)

Lumbar 1-4
-0.09 (-0.24, 0.06)

Menopausal Women

Total Femur
-0.28 (-0.45, -0.11)

Trochanter
-0.30 (-0.55, -0.06)

Intertrochanter
-0.22 (-0.35, -0.08)

Femoral Neck
-0.21 (-0.39, -0.02)

Ward's Triangle
-0.13 (-0.29, 0.03)

Lumbar 1-4
-0.17 (-0.31, -0.04)

External Review Draft

9-115

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tPollack et al.
(2013)

Western New York
United States
2005-2007
Cross-Sectional

BioCycle Study
n = 248

Premenopausal women
ages 18-44 yr

Blood

Blood Pb measured in venous
whole blood using ICP-MS

Age at measurement:

Mean (SD): 27.4 (8.2) yr

Mean: 1.03 [jg/dL

Bone mineral density

BMD in the hip, spine,
wrist, and whole body
(g/cm2) measured via
DXA

Age at outcome:

Mean (SD): 27.4 (8.2) yr

Age, BMI, race,	Change in BMD (g/cm2)

parity, caloric intake,

and age at menarche	„ ,

a	Whole Body

-0.004 (-0.03, 0.021)

Total Hip

-0.002 (-0.035, 0.031)

Lumbar Spine
-0.016 (-0.048, 0.016)

Wrist

0.001 (-0.012, 0.014)

tLi et al. (2020b)

Sichuan Province
China

Cross-sectional

n = 799

Blood, Urine

BMD

Study area included two Blood Pb measured in venous Osteoporosis (BMD T-

rural towns, one with a
history of heavy metal
contamination. Generally
healthy adults ages 40-
75 yr old who lived in
study area for >15 yr and
subsisted on rice and
vegetables grown in study
area.

whole blood using ICP-MS

Age at measurement:
40-75 yr

Median 3.4 [jg/dL
75th: 4.7 pg/dL

score <2.0); BMD
measured via X-ray
absorptiometry

Age at outcome:
40-75 yr

Age, BMI, and
smoking status

OR Osteoporosis
Prevalence (>3.4 pg/dL
vs. <3.4 [jg/dL blood Pb)

Males

0.6 (0.24, 1.49)

Females
1.33 (0.61, 2.88)

Non-Smoking Females
0.94 (0.4, 2.21)

External Review Draft

9-116

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tLimetal. (2016) KNHANES

South Korea

2008-2011

Cross-Sectional

n = 2429

General population; >18 yr
old

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:

>18 yr

Median: 2.22 [jg/dL

25th: 1.66 pg/dL
75th: 2.93 pg/dL

BMD (osteoporosis and Age, sex, smoking

osteopenia)

Ostopenia (BMD T-score
<-1.0) and Osteoporosis
(BMD T-score <-2.5)

Age at outcome:

>18 yr

status, alcohol
consumption,
geographic region,
education level,
occupation, and
family income

ORs for Osteoporosis or
Osteopenia prevalence
across blood Pb
quartiles

Q1
Q2
Q3
Q4

Ref.

1.08 (0.85, 1.37)
1.18 (0.91, 1.53)
1.49 (1.12, 1.98)

tLee and Park
(2018)

Ansung and Ansan
South Korea
2001-2002
Cross-Sectional

Korean Association
Resource (KARE) Cohort
n = 443

Adults aged 40-65 yr from
two South Korean
communities, on rural
(Ansung) and one urban
(Ansan)

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:
40-65 yr

GM: 4.44 pg/dL

BMD

Age, sex, geographic Change in BMD T-score

region, income, and

BMD (T-score)	physical activity

measured via ultrasound

Age at outcome:

40-65 yr

All

-0-0.26 (-0.45, -0.07)

Ever Smokers
-0.47 (-0.85, -0.09)

Current Smokers
-0.60 (-1.02, -0.17)

Never Smokers
-0.15 (-0.37, 0.07)

External Review Draft

9-117

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

Osteoarthritis

tPark and Choi
(2019)

South Korea
4 Years (2010—
2012)

Cross-sectional

KNHANES
n = 884

Women, >55 yr old

Blood

BLL measured in venous whole
blood using GFAAS

Age at measurement:

Mean: 62.9 yr

Median: 2.22 [jg/dL
Max: 7.84 pg/dL

Osteoarthritis

Radiographic and
symptomatic
osteoarthritis.
Radiographic OA (rOA)
assessed in the hip,
knee, and spine using
the Kellgren-Lawrence
grading system.
Symptomatic OA (sxOA)
assessed using a
combination of
radiographic evidence
and self-reported
symptoms

Age at outcome:

Mean: 62.9 yr

Age, smoking status,
alcohol use, physical
activity, education,
occupation, income,
diabetes,
hypertension, and
BMI

ORs for Osteoarthritis
prevalence per In-unit
increase in blood Pb
(Mg/dL)

rOA Knee
1.77 (1.17, 2.67)

sxOA Knee
1.50 (0.90, 2.53)

rOA Back
1.05 (0.70, 1.59)

sxOA Back
0.68 (0.39, 1.18)

tNelson et al.
(2011a)

Johnston County,
N.C.

United States
2003-2004 and
2006-2008
Cross-Sectional

Johnston County
Osteoarthritis Project
n = 668

African American and
White adults ages >45 yr
old

Blood

Blood Pb measured in venous
whole blood using ICP-MS

Age at measurement
Mean (SD):

Females: 62.4 (9.4) yr

Males: 64.5 (10.8) yr

Median:

Females: 1.9 [jg/dL
Males: 2.2 [jg/dL

Max:

Females: 25.4 [jg/dL

Osteoarthritis

Urine and serum
biomarkers of joint tissue
metabolism

Age at outcome:

Mean (SD):

Females: 62.4 (9.4) yr
Males: 64.5 (10.8) yr

Age, race, BMI,
smoking status

and % Change in urine and
serum biomarkers of
joint tissue metabolism

Males

uNTX-l
1.2% (-1.0,

UCTX-II
1.4% (-0.6,

COMP
1.6% (-0.1,

3.4%)

3.4%)

3.2%)

External Review Draft

9-118

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

Males: 25.1 [jg/dL

C2C

0.0% (-1.0, 1.0%)

CPII

-0.2% (-1.4, 1.0%)

C2C.CPII
0.0% (-1.4, 1.4%)

HA

0.2% (-2.5, 3.0%)

Females

uNTX-l

7.7% (3.9, 11.7%)
UCTX-II

5.1% (0.8, 9.5%)
COMP

-0.8% (-2.8, 1.2%)
C2C

0.0% (-1.6, 1.6%)
CPII

1.7% (-0.6, 4.1%)
C2C.CPII

-1.2% (-3.5, 1.1%)

External Review Draft

9-119

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

HA

-0.8% (-6.6, 5.3%)

tNelson et al.
(2011b)

Johnston County,
N.C.

United States
2003-2004 and
2006-2008
Cross-Sectional

Johnston County
Osteoarthritis Project
n = 1635

African American and
White adults ages >45 yr
old

Blood

Blood Pb measured in venous
whole blood using ICP-MS

Age at measurement:

Mean (SD): 65.3 (11.0) yr

Mean: 2.4 [jg/dL

Osteoarthritis

Radiographic and
symptomatic
osteoarthritis.
Radiographic OA (rOA)
assessed in the knee
using the Kellgren-
Lawrence grading
system. Symptomatic OA
(sxOA) assessed using a
combination of
radiographic evidence
and self-reported
symptoms

Age at outcome:

Mean (SD): 65.3 (11.0)
yr

Age, sex, race,
ethnicity, BMI,
current smoking,
current drinking

and

ORs for Prevalent
Osteoarthritis of the
Knee

rOA

1.10 (1.00,

sxOA
1.08 (0.96,

1.20)

1.20)

Oral Health - Adults

tWon et al. (2013)

South Korea
2009

Cross-Sectional

KNHANES
n = 1966

General population; >19 yr
old

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:

>19 yr

Mean NR

T1
T2
T3

<1.73 [jg/dL
1.73-3.04 [jg/dL
>3.04 [jg/dL

Periodontal disease

Community Periodontal
Index (code >3,
corresponding to pockets
>3.5 mm)

Age at outcome:

>19 yr

Age, sex, family
income, education
level, use of floss,
use of interproximal
toothbrush, alcohol
consumption,
smoking status, ETS
in workplace,
diabetes,
hypertension, and
oral health status

ORs for Prevalent
Periodontal Disease
across blood Pb tertiles

T1
T2
T3

Ref.

1.37 (0.97,
1.31 (0.88,

1.93)
1.96)

External Review Draft

9-120

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

1-Han etal. (2013)

South Korea

2008-2010

Cross-Sectional

KNHANES
n = 4716

General population; >19 yr
old

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:

>19 yr

GM:

Periodontitis: 2.60 [jg/dL
No periodontitis: 2.12 [jg/dL

Periodontal disease

Community Periodontal
Index (code >3,
corresponding to pockets
>3.5 mm)

Age at outcome:

>19 yr

Age, gender, income,
education, frequency
of daily
toothbrushing,
regular dental check-
up, smoking, alcohol
consumption,
physical activity,
fasting plasma
glucose, BMI, white
blood cell count and
urine cotinine
concentration.

ORs for Prevalent
Periodontal Disease
across blood Pb
quintiles

Q1 (<1.59 [jg/dL)

Ref.

Q2 (1.59-2.05 [jg/dL)
1.36 (1.00, 1.85)

Q3 (2.05-2.52 pg/dL)
1.3 (0.96, 1.76)

Q4 (2.52-3.57 pg/dL)
1.55 (1.13, 2.13)

Q5 (>3.17 pg/dL)
1.6 (1.15, 2.22)

tKim and Lee
(2013)

South Korea

2008-2009

Cross-Sectional

KNHANES	Blood

n = 3996

Blood Pb measured in venous
General population; >20 yr whole blood using GFAAS
old

Age at measurement:

>20 yr

GM: 2.31 pg/dL

Periodontal Disease

Community Periodontal
Index (code >3,
corresponding to pockets
>3.5 mm)

Age at outcome:

>20 yr

Age, body mass
index (BMI),
residence area,
education level,
household income,
smoking and drinking
status, hemoglobin,
glucose, use of floss
or interproximal
toothbrush, decayed,
missing, or filled
permanent teeth
(DMFT), and active
caries

ORs for Prevalent
Periodontal Disease
across blood Pb
quintiles

Males

1.854 (1.265, 2.717)

Males (adjusted for Hg,
Cd)

1.699 (1.154, 2.502)
Females

1.301 (0.883, 1.917)

Females (w/ Hg and Cd)

External Review Draft

9-121

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

1.242 (0.833, 1.851)

Oral Health - Children and Adolescents

tWu etal. (2019)

Mexico City
Mexico

Initial Recruitment:
1997-2005; Follow-
up: 2008-2013
Cohort

Early Life Exposures in
Mexico to Environmental
Toxicants (ELEMENT)
n = 173 to 386 (depending
on exposure metric)

Mother/child pairs
recruited from 2 public
hospitals serving low-to
moderate-income
populations

Blood

Maternal and child blood Pb
measured in venous whole
blood using GFAAS. Maternal
bone Pb measured using K-XRF

Age at measurement:

Maternal BLL:

1st, 2nd, and 3rd trimester

Dental caries

Teeth evaluated by
trained examiners who
assigned decayed,
missing, filled tooth
(DMFT) scores

Age at outcome:
Adolescence (10 to
18 yr)

Child BLL:

1, 2, 3, and 4 yr, and in
adolescence (10 to 18 yr)

Maternal bone:

Postnatally

Mean (males, females):
1st trimester: 6.06, 6.36 [jg/dL
2nd trimester: 5.24, 5.25 [jg/dL
3rd trimester: 5.67, 5.73 [jg/dL
Childhood: 15.48, 15.18 pg/dL
Adolescence: 3.60, 3.34 pg/dL
Maternal tibia: 8.64, 9.68 pg/g
Maternal patella: 7.18, 8.64 pg/g

Sex, cohort, mother's Rate Ratio of Decayed,
education, sugar Missing, and Filled
sweetened	Teeth per In-unit

beverages intake increase in blood or
bone Pb

1st Trimester BLL
1.07 (0.90, 1.27)

2nd Trimester BLL
1.12 (0.94, 1.32)

3rd Trimester BLL
1.17 (0.99, 1.37)

Childhood BLL
1.14 (0.94, 1.38)

Adolescent BLL
0.97 (0.81, 1.16)

Maternal Patella Pb
0.95 (0.88, 1.03)

Maternal Tibia Pb
0.98 (0.88, 1.08)

External Review Draft

9-122

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tKimetal. (2017)

Seoul, Daegu,
Cheonan, and
Busan

South Korea

2005-2010

Cross-sectional

The Children's Health and Blood
Environment Research
(CHEER) group
n = 1,565 (children w/
permanent teeth) and
1,241 (children w/
deciduous teeth)

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:
"School-aged"

School-aged children from
urban, rural, and
industrialized areas with
BLLs <5 [jg/dL

GM: 1.53 [jg/dL

Dental caries

DMFS sum by trained
dental hygienists

Age at outcome:
"School-aged"

Sex, age
(categorical),
household income
(categorical), and
urinary cotinine level
(categorical)

PR for Decayed and
Filled Surfaces

Deciduous Teeth

Decayed Surfaces
1.16 (0.91, 1.49)

Filled Surfaces
1.11 (0.98, 1.25)

DMFS

1.14 (1.02, 1.27)

Permanent Teeth

Decayed Surfaces
0.69 (0.45, 1.07)

Filled Surfaces
0.87 (0.73, 1.04)

DMFS

0.83 (0.69, 0.99)

tWiener et al.
(2015)

United States

1988-1994

Cross-Sectional

NHANES III
n = 3127

General population; 2 to
6 yr old

Blood

Blood Pb measured in venous
whole blood using GFAAS

Age at measurement:

2 to 6 yr

Mean NR
28.2% <2 [jg/dL;

48.3% 2 to <5 [jg/dL;
18.4% 5 to <10 [jg/dL;

Dental caries

Number of teeth with at
least one decayed or
filled surface as detected
by trained examiners

Age at outcome:

2 to 6 yr

Sex, race/ethnicity,
age, urban status,
census region,
poverty index, family
education, ETS
exposure, birth
weight, breastfed,
dental visit, and
parental perception
of oral health

PR for Decayed and
Filled Surfaces

<2 [jg/dL:

Ref.

2-5 [jg/dL:
1.84 (1.36, 2.50)

5-10 [jg/dL:

External Review Draft

9-123

DRAFT: Do not cite or quote

-------
StuedynDesfgnn	Study Population	Exposure Assessment	Outcome	Confounders Eff8Ct95% cis^ ^

5.1% >10 [jg/dL	2.14(1.36,3.36)

>10 [jg/dL:
1.91 (1.17, 3.11)

BLL = blood lead level; BMD = bone mineral density; BMI = body mass index; CHEER = Children's Health and Environment Research; CI = confidence interval; C2C = serum
cleavage neoepitope of type II collagen; COMP = cartilage oligomeric matrix protein; CPU = carboxypropeptide of type II collagen; DMFS =; DMFT = decayed, missing, and filled
teeth; DXA = Dual-energy X-ray absorptiometry; ELEMENT = Early Life Exposures in Mexico to Environmental Toxicants; ETS = environmental tobaccos smoke; GFAAS = Graphite
furnace atomic absorption spectrometry; ICP-MS = inductively coupled plasma mass spectrometry; KARE = Korean Association Resource; KNHANES = Korean National Health and
Nutrition Examination Survey; NHANES = National Health and Nutrition Examination Survey; NR = not reported; OA = osteoarthritis; Pb = lead; PR = prevalence
ratio; rOA = radiographic osteoarthritis; sxOA = symptomatic osteoarthritis; SD = standard deviation; Q = quartile; yr = year(s).

aEffect estimates are standardized to a 1 |jg/dL increase in BLL or a 10 |jg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.
fStudies published since the 2013 Integrated Science Assessment for Lead.

1

Table 9-12 Animal toxicological studies of exposure to Pb and musculoskeletal effects.

Study

Species	Timing of Exposure

(Stock/Strain), n, Sex Exposure Details

BLL As Reported
(Hg/dL)

Endpoints Examined

Beieretal. (2017)

Mouse (C57BL.6),
0 ppm Pb, M/F, n = NR
100 ppm Pb, M/F,
n = NR

PND 240 Oral, 0.17 ± 0.19 ng/dLfor Serum Protein Levels of Dickkopf-1, Serum Protein Levels of
drinking 0 ppm,	Sclerostin (scl), Serum Protein Levels of C-terminal telopeptide

water 58.67 ± 4.61 ng/dL (CTx-1), Serum Protein Levels of type 1 procollagen (P1NP),
for 100 ppm -	Energy to Femur Failure (Males, 8 moo), Femur Yield Load /

PND 240	Maximum Load (Males, 8 moo), Maximum Femur Stiffness

(Males, 8 moo), Osteoclast Surface/Bone Surface (Oc.S/BS) by
Micro-Computed Tomography (microCT), Osteoclast
Number/Trabecular Area (N.Oc/Tb.Ar) by Micro-Computed
Tomography (microCT), Osteoblast Number/Trabecular Area
(N.Ob/Tb.Ar) by Micro-Computed Tomography (microCT),
Adipocyte size (Ad Size) by Micro-Computed Tomography
(microCT), Adipocyte Volume/Total Volume (AV/TV) by Micro-
Computed Tomography (microCT), Bone Volume to Total Volume
(BV/TV) by Micro-Computed Tomography (microCT)

External Review Draft

9-124

DRAFT: Do not cite or quote

-------
Study

Species	Timing of Exposure

(Stock/Strain), n, Sex Exposure Details

BLL As Reported
(Hg/dL)

Endpoints Examined

Beieretal. (2016)

Mouse (C57BL.6), PND 30, Oral,
0 ppm Pb, F, 200 ppm 90,180, drinking
Pb, F, 500 ppm Pb,/F 360	water

0 ng/mL for 0 ppm, Femur Length, Areal Bone Mineral Density (aBMD), Bone Mass,
50 ng/mL for	Bone Weight, Body Fat, Femur Diameter, P1NP (ng/mL),

100 ppm, 100 ng/mL TRAP5b (U/L), CTx (ng/mL), Calcitonin (pg/mL), 17 beta-estradiol
for 300 ppm -	(ng/mL), Dkk-1 (ng/mL), Femoral BV/TV, Tb.N, Tb.Sp, Conn.D,

PND 28	SMI, Cort Th, Cort BA, Tb Extension, Bone Strength, Beta-

Catenin Protein Levels, TNF-Alpha Protein Levels, NF-kB Protein
Levels, b-catenin RT-PCR, Peroxisome Proliferator-Activated
Receptor-c RT-PCR, CD47 RT-PCR, Nuclear Factor Of Activated
T Cells RT-PCR, CTSK RT-PCR

aBMD = areal bone mineral density; AV/TV = adipocyte volume/total volume; BV/TV = bone volume to total volume; CTx-1 = C-terminal telopeptide; mo = month(s);
microCT = Micro-Computed Tomography; NF-kB = nuclear factor kappa B; N.Oc/Tb.Ar = Osteoclast Number/Trabecular Area; Oc.S/BS = Osteoclast Surface/Bone Surface;
P1 NP = type 1 procollagen; PND = postnatal day; RT-PCR = reverse transcription-polymerase chain reaction; scl = sclerostin; TNF = tumor necrosis factor.

Table 9-13 Epidemiologic studies of exposure to Pb and ocular effects.

Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

Glaucoma

tWana et al. (2018b)

United States
1991-1999 (Follow-up
through 2014)

Cohort

Veterans Affairs NAS
n = 702

Healthy male Veterans at
time of enrollment in the
NAS (1963) and without
glaucoma at baseline
(time of bone lead
measurement)

Bone

Tibia and patella lead
measured using K-XRF
Age at measurement:
Mean age: 66.8

Mean -

Tibia: 21.7 |jg/g
Patella: 31.0 |jg/g

Glaucoma

Incident cases of primary
open-angle glaucoma
identified using validated
criteria to assess medical
records

Age, BMI,
education, job
type, pack-yr,
diabetes mellitus,
systemic

hypertension, and
ocular

hypertension.

HRs for Glaucoma
Incidence

Tibia Pb

1.28 (0.99, 1.65)

Patella Pb
1.42 (1.11, 1.82)

External Review Draft

9-125

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tPark and Choi (2016)

KNHANES

Blood

Intraocular pressure

Age, sex, smoking

Change in intraocular



n = 8371





status, alcohol

pressure (mmHg):

South Korea



Blood Pb was measured in

Intraocular pressure measured

consumption, job

0.088 (0.06, 0.117)

2008-2012

General population,

venous whole blood using

using a Goldmann applanation

status, education,

Cross-sectional

>20 yr old with no history

GFAAS

tonometer

residence,





of glaucoma

Age at measurement:



hypertension







>20 yr old

Age at outcome:

medication use,









>20 yr old

and family history







GM: 2.19 [jg/dL



of glaucoma



+Lin etal. (2015)

KNHANES

Blood

Glaucoma

Age, sex, exercise,

OR for Glaucoma



n = 2680





and ferretin and

Prevalence13:

South Korea



Blood Pb was measured in

Presence of glaucoma was

aspartate

1.04 (0.84, 1.29)

2008-2009

General population,

venous whole blood using

assessed by testing of visual

aminotransferase

Cross-sectional

>19 yr old with no history

GFAAS

function using frequency-

levels





of retinal disease or

Age at measurement:

doubling technology.







stroke

>19 yr old













Age at outcome:









Mean -

19 yr old









w/ glaucoma: 2.70 [jg/dL











w/o glaucoma: 2.52 [jg/dL







+Lee etal. (2016)

KNHANES

Blood

Glaucoma

Age group, region

ORs for Glaucoma



n = 5198





of residence,

Prevalence13

South Korea



Blood Pb was measured in

Presence of glaucoma was

occupation,



2008-2012

General population,

venous whole blood using

assessed by testing of visual

education level,

A/rt cm a /

Cross-sectional

>19 yr old without a

GFAAS

function using frequency-

smoking status,

IvUifllal !\Jr



history of glaucoma or

Age at measurement:

doubling technology.

hypertension,

0.93 (0.65, 1.34)



age-related macular

>19 yr old



family history of





degeneration



Age at outcome:

glaucoma, and

Low-Teen IOP





GM -

>19 yr old

IOP

1.16 (0.74, 1.83)





No Glaucoma: 2.32 [jg/dL;







Glaucoma: 2.28 [jg/dL

High-Teen IOP
0.65 (0.36, 1.18)

External Review Draft

9-126

DRAFT: Do not cite or quote

-------
Reference^nd Study study Population	Exposure Assessment	Outcome	Confounders EffeCt95yJ c?iaS a"d

Age-Related Macular Degeneration

tPark etal. (2015)

South Korea

2008-2011

Cross-sectional

KNHANES
n = 3865

General population,
>40 yr old

Blood

Blood Pb was measured in
venous whole blood using
GFAAS

Age at measurement:
>40 yr old

Mean: 2.69 [jg/dL

Age-related macular
degeneration

Macular degeneration was
assessed using retinal
photographs. Photographs
were graded at least twice
using a standardized protocol.

Age at outcome:

>40 yr old

Age, sex, smoking Early-Stage AMD (OR):

status, occupation,
residence,
household income,
anemia, BMI

1.12 (1.02, 1.23)

Late-Stage AMD (OR):

1.25 (1.05, 1.50)

External Review Draft

9-127

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tHwana et al. (2015)

South Korea

2008-2012

Cross-sectional

KNHANES
n = 4933

General population,
>40 yr old

Blood

Blood Pb was measured in
venous whole blood using
GFAAS

Age at measurement:
>40 yr old

Mean: 3.15 [jg/dL

Quintile 1

<1.75 [jg/dL

Quintile 2

1.75-2.25 [jg/dL

Quintile 3

2.25-2.73 [jg/dL

Quintile 4

2.73-3.38

Quintile 5

>3.38 [jg/dL

Age-related macular
degeneration

Macular degeneration was
assessed using retinal
photographs. Photographs
were graded twice using a
standardized protocol.

Age at outcome:

>40 yr old

NA

ORs (Early-Stage
AMD; Quintiles)

Q1
Q2
Q3
Q4
Q5

Reference
1.04 (0.62, 1.73)
1.14 (0.70, 1.84)
1.26 (0.78, 2.06)
1.55 (0.94, 2.53)

Men Only:

Q1
Q2
Q3
Q4
Q5

Reference
0.66 (0.31, 1.40)
1.32 (0.68, 2.56)
0.80 (0.40, 1.60)
1.32 (0.68, 2.54)

Women Only:
Q1: Reference
Q2: 1.72 (0.86, 3.46)
Q3: 1.83 (0.90, 3.73)
Q4: 1.41 (0.72, 2.77)
Q5: 1.92 (1.06, 3.48)

External Review Draft

9-128

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tWu et al. (2014)

United States

2005-2008

Cross-sectional

NHANES
n = 5390

General population,
>40 yr old

Blood

Blood Pb was measured in
venous whole blood using ICP-
MS

Age at measurement:

>40 yr old

GM: 1.61 [jg/dL; Median:
1.77 [jg/dL
75th: 2.61 pg/dL
Max: 26.8 pg/dL

Age-related macular
degeneration

Macular degeneration was
assessed using retinal
photographs. Photographs
were graded twice using a
standardized protocol.

Age at outcome:

>40 yr old

Age, aged-
squared, gender,
race, education,
BMI, pack-yr

ORs for AMD
Prevalence (Quartiles)

Q1
Q2
Q3
Q4

Reference
0.86 (0.60, 1.22)
1.00 (0.68, 1.48)
0.86 (0.59, 1.26)

Quartile 1
Quartile 2
Quartile 3
Quartile 4

0.18-1.2 pg/dL
1.21-1.77 pg/dL
1.78-2.61 pg/dL
2.62-26.8 pg/dL

Other Ocular Effects

tWanqetal. (2016)

United States

1999-2008

Cross-sectional

NHANES
n = 9763

General population,
>50 yr old

Blood

Blood Pb was measured in

venous whole blood using AAS

(1999-2002) and GFAAS

(2003-2008)

Age at measurement:

50+ yr old

Cataract surgery

Self-reported cataract surgery

Age at outcome:

>50 yr old

Age, race, gender, OR for Cataract

education,
diabetes mellitus,
BMI, serum
cotinine, and pack-
yr

Surgery per doubling
of BLL:

0.97 (0.88, 1.06)

GM: 1.97 pg/dL

External Review Draft

9-129

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tJuna and Lee (2019)

South Korea

2010-2012

Cross-sectional

KNHANES
n = 23376

General population,
>40 yr old

Blood

Blood Pb was measured in
venous whole blood using
GFAAS

Age at measurement:
>40 yr old

GM -

Male: 2.82 pg/dL;

Female: 2.05 pg/dL

Dry eye disease

Self-reported symptoms of dry
eye disease

Age at outcome:

>40 yr old

Age, sex, smoking ORs for Dry Eye

status, alcohol
consumption,
region, education,
occupation, family
income, family
history of
ophthalmologic
disease, and
history of
ophthalmologic
surgery

Disease Prevalence
(Tertiles)

T1: Reference
T2: 1.12 (0.85,
T3: 0.79 (0.56,

1.48)
1.1)

Tertile 1
Tertile 2
Tertile 3

<2.03 pg/dL
2.03-2.82 pg/dL
>2.82 pg/dL

AAS = atomic absorption spectrometry; AMD = age-related macular degeneration; BMI = body mass index; GFAAS = Graphite furnace atomic absorption spectrometry;
GM = geometric mean; HR = hazard ratio; ICP-MS = inductively coupled plasma mass spectrometry; IOP = intraocular pressure; KNHANES = Korean National Health and Nutrition
Examination Survey; K-XRF = K-Shell X-Ray Fluorescence; NA = not available; NAS = Normative Aging Study; NHANES = National Health and Nutrition Examination Survey;
OR = odds ratio; Pb = lead; T = tertile; yr = year(s).

aEffect estimates are standardized to a 1 pg/dL increase in BLL or a 10 pg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.
bPer natural log unit increase in pg/dL of blood Pb.

fStudies published since the 2013 Integrated Science Assessment for Lead.

External Review Draft

9-130

DRAFT: Do not cite or quote

-------
Table 9-14 Animal toxicological studies of Pb exposure and ocular effects.

Study	Species (Stock/Strain), n, Timing of Exposure Exposure Details BLL As Reported (pg/dL)	Endpoints Examined

Shen et al. (2016) Rat (Sprague Dawley), 0 ppm
Pb, M, n = 12 (BLL), n = 6
(other endpoints)

55 ppm Pb (0.01%), M, n = 12
(BLL), n = 6 (other endpoints)

109 ppm Pb (0.02%), M,
n = 12 (BLL), n = 6 (other
endpoints)

Blot Protein Levels of
Occludin, Western Blot Protein
Levels of Claudin-5, Western
Blot Protein Levels of pAkt
(Ser473), Western Blot Protein
Levels of pAkt (Thr308)

BLL weeks 1, 2, 3, 4, Oral, drinking water
5, 6; Other
Endpoints week 6

1.11 ±0.08 |jg/dL for 0.00%
12.58 ±2.42 [jg/dL for 0.01 %
19.00 ±2.59 [jg/dL for 0.02%

Retinal Thickness, Blood-
Retina-Barrier Permeability,
Occludin Protein Levels,
Claudin 5 Protein Levels,
Immunofluorescence Protein
Levels of Occludin,
Immunofluorescence Protein
Levels of Claudin 5, Western

Perkins et al. Mouse (C57BL.6), Bcl-xL
(2012)	Transgenic ((C57BL.6),

Background)Wild Type 0.0%
Pb Acetate, M/F, n = 3 to 7,
varying between groups and
between assays
Wild Type 0.015% Pb Acetate,
M/F, n = 3 to 7, varying
between groups and between
assays

Transgenic 0.0% Pb Acetate,
M/F, n = 3 to 7, varying
between groups and between
assays

Transgenic 0.015% Pb
Acetate, M/F, n = 3 to 7,
varying between groups and
between assays

BLL PND 21,
PND60

Other Endpoints
PND 60 to 70

Oral, drinking water

1.9 ± 1.0 pg/dl for 0.0%,
20.6 ±4.7 |jg/l for 0.015% Pb-
-PND 21

3.6 ± 1.8 pg/dl for 0.0%,
5.6 ± 2.7 |jg/l for 0.015% Pb -
PND 60

Conventional Transmission
Electron Microscopy (TEM) of
Cell and Organelle Structure,
Three-Dimensional Electron
Microscope Tomography of
Cell and Organelle Structure,
Mitochondrial Cristae
Measurements in Rod
Spherules, Mitochondrial
Cristae Measurements in Cone
Pedicles, Mitochondrial Crista
Junction Diameter and Density
in Rod Spherules,

Mitochondrial Crista Junction
Diameter and Density in Cone
Pedicles, Photoreceptor and
Synaptic Terminal Oxygen
Consumption (Light-Adapted)

BLL = blood lead level; CI = confidence interval; F = female; M = male; pAkt = phosphorylated Akt; Pb = lead.

External Review Draft

9-131

DRAFT: Do not cite or quote

-------
Table 9-15 Epidemiologic studies of Pb exposure and respiratory effects.

SyDesfg^	Study Population	Exposure Assessment	Outcome	Confounders	and

Children and Adolescents

tMadriqal et al. (2018)

United States
2011-2012
Cross-sectional

NHANES
n:1234

Children and adolescents
aged 6-17 yr

Blood Pb measured in
venous whole blood
using ICP-MS.

Age at measurement:
6-17 yr old

Median:

0.56

pg/dL

25th

percentile:

0.44

pg/dL

75th

percentile:

0.85

pg/dL

Pulmonary function:
FEVi, FVC, FEVi:
FVC, and FEF25-75%

Spirometry was
performed in the
standing position
using a standardized
protocol according to
the recommendations
of the American
Thoracic Society for
FEV1 and FVC.

Age at outcome:
6-17 yr old

Age, sex, race, height,
family income to poverty
ratio, serum cotinine, use
of anti-asthmatic,
bronchodilator, or inhaler
medications

Change in lung function
parameters across
blood Pb quartiles

FEVi

Q1
Q2
Q3
Q4

Ref.

4.8 (-98.3, 107.8)
22.3 (-49.3, 93.9)
41.9 (-46.9, 130.6)

FVC
Q1: Ref.

Q2: 1.6 (-88.5, 91.7)
Q3: 23.8 (-46.4, 94.0)
Q4: 45.5 (-49.2, 140.2)

FEVr.FVC
Q1: Ref.

Q2: 0.0003 (-0.01, 0.01)
Q3: -0.001 (-0.01, 0.01)
Q4: 0.002 (-0.01, 0.02)

FEF25-75%

Q1: Ref.

Q2: -8.1 (-229.8, 213.7)
Q3: -28.9 (-160.5,
102.7)

Q4: 0.71 (-193.1, 192.5)

External Review Draft

9-132

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tZena et al. (2017)

Guiyu, Xiashan, and
Haojiang

Guangdong Province,
China

November -
December 2013

Cross-sectional

Preschool children aged 5-
7 yr

n = 206 (n = 100 from Guiyu,
n = 54 from Xiashan, n = 52
from Haojiang)

Blood Pb measured in
venous whole blood
using GFAAS

Age at measurement:
5-7 yr old

Median

Exposed (Guiyu):
5.53 [jg/dL

Unexposed (Xiashan
and Haojiang):
3.57 [jg/dL

75th Percentile:
Exposed: 7.04 [jg/dL
Unexposed: 4.86 [jg/dL

Lung function
parameters: FVC and
FEV1

Spirometry was
conducted with a
portable spirometer;
results of three
readings were
recorded and the
highest FVC and
FEV1 was used in the
analysis

Age at outcome:
5-7 yr old

Age, gender, height,
family member daily
smoking, family income
level, parental education
level, daily outdoor play
time, and living area

Change in lung function
parameters per In-unit
increase in blood Pb
(Hg/dL)

FEVi (mL)

-15 (-93, 63)

FVC (mL)

-29 (-100, 43)

tLittle et al. (2017)

Legnica-Glogow

District

Poland

1995 and 2007
Cross-sectional

Polish schoolchildren aged
10-15 yr

n = 184 male
n = 189 female

Blood Pb measured in
venous whole blood
using GFAAS

Age at measurement:
10-15 yr

FVC

A Spiro ProVR unit
was used to measure
pulmonary function.
FVC was computed by
the instrument as a
percentage of gender-
, age- and height-
specific normative
data.

Adjusted for height

Change in FVC (mL) per
logio-unit increase in
blood Pb (|jg/dL)

Boys

-5.1 (-13.9, 3.7)

Girls

-12.9 (-23.2, -2.6)

Age at outcome:
10-15 yr

External Review Draft

9-133

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tZena et al. (2016) Children age 3-8

Guiyu and Haojiang
China

December 2012 to
January 2013
Cross-sectional

n = 470 children
n = 170 from Haojiang and
n = 300 from Guiyu)

Blood Pb measured in
venous whole blood
using GFAAS.

Age at measurement:
3-8 yr old

Medians

Guiyu: 6.24 |jg/dL
Haojiang: 4.75 [jg/dL

75th: BLL:

Guiyu: 8 [jg/dL
Haojiang: 5.76 [jg/dL

Respiratory
symptoms: wheeze,
cough, dyspnea, and
phlegm

The respiratory
symptoms such as
wheeze, cough,
phlegm, and dyspnea
were defined by the
standard

questionnaire from the
European Community
Respiratory Health
Survey (ECRHS)

Age at outcome:
3-8 yr old

Age, gender, passive
smoking, living in Guiyu,
whether use home as
workshop, whether home
close to e-waste recycling
site, and whether child
contact e-waste

OR (>5 [jg/dL vs.
<5 [jg/dL blood Pb)

Wheeze

0.64 (0.32, 1.27)

Dyspnea
0.64 (0.23, 1.79)

Cough

0.95 (0.6, 1.52)

Phlegm

1.2 (0.72, 2.01)

Adults

tPaketal. (2012)

Shiwha and Banwol
Korea

2005 and 2007
(Shiwha) and 2006
and 2008 (Banwol)
Cohort

Shiwha and Banwol
Environmental Health Cohort
(SBEHC)

Men and women over the age
of 30 residing in Shiwha or
Banwoi and completed both
pulmonary function tests
during cycle 1 (2005-2006)
and cycle 2 (2007-2008)

Blood Pb measured in
venous whole blood
using GFAAS

GM (GSD):

Cycle 1: 1.55 (1.76)

pg/dL

Cycle 2: 1.96 (1.66)
pg/dL

FEVi and FVC

Pulmonary function
was measure via
spirometry

Age at outcome: 30+

Age, sex, baseline height,
baseline FVC,
methacholine, cotinine
level

Accelerated FVC
Decline

177.0 (24.1, 329.9)

Accelerated FEVi
Decline

107.0 (-0.8, 214.8)

n = 263 (n = 112 males)

External Review Draft

9-134

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population	Exposure Assessment	Outcome

Confounders

Effect Estimates and
95% Clsa

tLeem et al. (2015)

Korea

2008-2012

Cross-sectional

KNHANES
n = 5972

Adults >20 yr who completed
spirometry and had blood
measurements

Blood Pb measured in
venous whole blood
using GFAAS

Age at measurement:
20+

Mean BLL

non-OLF: 2.36 [jg/dL
OLF: 2.77 pg/dL

Obstructive lung
function (OLF)

Spirometry was used
for lung function. OLF
was defined as
FEV1/FVC <0.7

Age at outcome:

20+

Age, sex, BMI, and
smoking status

Change in lung function
parameters per In-unit
increase in blood Pb
(Hg/dL)

FEVi (mL)

0 (-116, 116)

FVC (mL)

9 (-3, 21)

FEVi/FVC (%)
-0.002 (-0.004, 0)

External Review Draft

9-135

DRAFT: Do not cite or quote

-------
Reference and
Study Design

Study Population

Exposure Assessment

Outcome

Confounders

Effect Estimates and
95% Clsa

tRokadia and
Aqarwal (2013)

United States

2007-2010

Cross-sectional

NHANES

n = 9575 (1164 OLD and 8411
non-OLD)

Serum Pb measured
from venous whole blood
samples using ICP-MS

Age at measurement:
General population; >18 yr old <| 8-79 yr

Geom. mean (SE)
non-OLD: 1.18 (1.0)

[jg/dL OLD:
pg/dL

1.73 (1.02)

Obstructive lung
disease (OLD)

Spirometric data were
collected from
NHANES participants;
Participants with OLD
were defined as FEV
1 /FVC <0.7; Mild
OLD: FEV1 = 80%
predicted; Moderate-
severe OLD: FEV1
<80% predicted

Age at outcome:
18-79 yr

Age, sex, race, BMI,
chronic kidney disease,
diabetes, hyperlipidemia,
hypertension, stroke,
coronary artery disease,
smoking, serum C-
reactive protein
concentration, and serum
cotinine concentration

ORs for OLD
Prevalence

All OLD
1.94 (1.10,

Mild OLD
1.21 (0.55,

3.42)

2.66)

Moderate-Severe OLD
3.49 (1.70, 7.16)

BLL = blood lead level; BMI = body mass index; CI = confidence interval; ECRHS = European Community Respiratory Health Survey; FEF = forced expiratory flow; FEV1 = forced
expiratory volume; FVC = forced vital capacity; GFAAS = graphite furnace atomic absorption spectrometry; GM = geometric mean; GSD = gestational sac diameter; ICP-
MS = inductively coupled plasma mass spectrometry; KNHANES = Korean National Health and Nutrition Examination Survey; NHANES = National Health and Nutrition Examination
Survey; OLD = obstructive lung disease; OLF = obstructive lung function; OR = odds ratio; Pb = lead; SBEHC = Shiwha and Banwol Environmental Health Cohort; Q = quartile;
yr = year(s).

aEffect estimates are standardized to a 1 |jg/dL increase in BLL or a 10 |jg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.

fStudies published since the 2013 Integrated Science Assessment for Lead.

1

External Review Draft

9-136

DRAFT: Do not cite or quote

-------
Table 9-16 Animal toxicological studies of exposure to Pb and respiratory effects.

Study Species (Stock/Strain), n, Sex Exposure

Exposure Details
(Concentration,
Duration)

BLL As Reported (pg/dL) b

Endpoints
Examined

Dumkova et al. (2017) Mouse (ICR) NR
experiment 1

Control (clean air), F, n = 5

1.23 x 10s PbO particles/cm3, F, n = 5

experiment 2

Control (clean air), F, n = 5

0.956 x 10s PbO particles/cm3, F, n = 5

Mice were exposed to
PbO NPs 24 hr/d for
6 wk.

<11 ng/g for control
(<1.166 pg/dL)

132 ng/g for Pb-exposed
(13.992 pg/dL)

IHC, Histology

Dumkova et al.	Mouse (CD1) (ICR)	NR

(2020b)

Control (clean air), F, n = 10 (2 wk, 6 wk,

11 wk)

PbO, F, n = 10 (2 wk, 6 wk, 11 wk)

PbO recovery, F, n = 10 (6 wk PbO, 5 wk
clean air)

174 ng/g PbO 11 wk
(17.4 ijg/dL)

27 ng/g PbO recovery
(6 wk/clean air 5 wk)
(2.7 pg/dL)

Mice were exposed to <3 ng/g in control (2 wk, 6 wk, Western blot,
clean air or PbO np	11 wk) (<0.3 pg/dL)	Histology, IHC,

24 hr/d 7 d/wkfor2 wk,	PCR

6 wk, or 11 wk. a
recovery group was
exposed to PbO for 6 wk
and then clean air for

5 wk (11 wk total)	148 ng/g PbO 6 wk

104 ng/g PbO 2 wk
(10.4 pg/dL)

(14.8 pg/dL)

External Review Draft

9-137

DRAFT: Do not cite or quote

-------
Study

Species (Stock/Strain), n, Sex

Timing of
Exposure

Exposure Details

(Concentration,
Duration)

BLL As Reported (pg/dL) b

Endpoints
Examined

Dumkova et al.
(2020a)

Mouse (ICR)

Control (clean air), F, n = 10 (d 3, 2 wk,
6 wk, 11 wk)

Pb(N03)2 (68.6 pg/m3), F, n = 10 (d 3,
2 wk, 6 wk, 11 wk)

Recovery (Pb(N03)2 68.6 |jg/m3), F,
n = 10 (6 wk Pb/5 wk recovery)

6 wk - 8 wk Mice were exposed to
at start	Pb(N03)2 np or clean air

24 hr/d, 7 d/wk for 3 d,
2 wk, 6 wk, or 11 wk. To
assess recovery, a
separate group of mice
were exposed to
Pb(N03)2 for 6 wk and
then clean air for 5 wk.

<3 ng/g for control at all
timepoints (d 3, 2 wk, 6 wk,
11 wk) (<0.3 pg/dL)

31 ng/g for Pb(N03)2 d 3
(3.1 pg/dL)

40 ng/g for Pb(N03)2 2 wk
(4.0 pg/dL)

47 ng/g for Pb(N03)2 6 wk
(4.7 pg/dL)

PCR, Histology,
IHC

85 ng/g for Pb(N03)2 11 wk
(8.5 pg/dL)

10 ng/g for Pb(N03)2
exposure 6 wk and clean air
for 5 wk (1.0 pg/dL)

BLL = blood lead level; BMI = body mass index; d = day(s); hr = hour(s); IHC = immunohistochemistry; NP = nanoparticle; Pb = lead; Pb(N03)2 = lead nitrate; PbO = lead monoxide;
PCR = polymerase chain reaction; wk = week(s).

External Review Draft

9-138

DRAFT: Do not cite or quote

-------
Table 9-17 Epidemiologic studies of Pb exposure and total mortality.

Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

Menke et al. (2006)

NHANES III 1988-1994,
mortality follow-up in 2001

-12 yr of follow-up
Cohort

NHANES III
n = 13,946, >20 yr

Average individual
born -1946

Blood (GFAAS with
Zeeman correction)
(Hg/dL)

Mean: 2.58
Tertiles
T1 <1.93
T2 1.94-3.62
T3 >3.63

Age of measurement
Mean 44.4

All-cause mortality

Cox proportional hazard
regression analysis adjusted
age, race/ethnicity, sex, urban
residence, cigarette smoking,
alcohol consumption,
education, physical activity,
household income,
menopausal status, BMI,
CRP, TC, diabetes mellitus,
hypertension, GFR category

HR

All-cause 1.09 (1.05, 1.14)

Schober et al. (2006)

NHANES III 1988-1994,
mortality follow-up in 2006

-8.55 yr of follow-up
Cohort

NHANES III
n = 9,686, >40 yr

Average individual
born in or before
-1951

Blood (GFAAS with
Zeeman correction)
(pg/dL)

T1 <5 (median 2.6)
T2 5-9 (median 6.3)
T3 >10 (median 11.8)

All-cause mortality

Cox proportional hazard
regression analysis adjusted
for sex, age, race/ethnicity,
smoking, education level.

Did not evaluate BMI or
cormorbidities

HR

All-cause 1.05 (1.03, 1.08)

Age of measurement
>40 yr

External Review Draft

9-139

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

Lustbera and Silberaeld
(2002)

NHANES II 1976-1980,
mortality follow-up in 1992
Cohort

NHANES II

n = 4,190, aged 30-
74

Average individual
born -1924

Blood (GFAAS with
Zeeman correction)15
(Hg/dL)

Mean (SD) 14.0 (5.1)
Median: 13
T1: <10
T2: 10-19
T3: 20-29

Age of measurement
Mean (SD) 54.1 (13.2)

All-cause and
circulatory mortality

Cox proportional hazard
regression analysis adjusted
for age, sex, location,
education, race, income,
smoking, BMI, exercise

HR(T1: Referent)0

All-cause

T2: 1.40 (1.16-1.69)
T3: 2.02 (1.62-2.52)

Khaliletal. (2009)

Baltimore, MD and
Monongahela Valley, PA

Blood Pb measured 1990-
1991, mortality follow-up for
-12 yr

Study of
Osteoporotic
Fractures
n = 533

women, ages 65-
87 yr

Blood (GFAAS with
Zeeman correction)
(pg/dL)

Mean (SD) 5.3 (2.3)
Range 1-21

Age of measurement
Mean 70

All-cause mortality

Cox proportional hazards
regression analysis adjusted
forage, clinic, BMI, education,
smoking, alcohol intake,
estrogen use, hypertension,
total hip BMD, walking for
exercise, and diabetes

HR (>8 [jg/dL vs. <8 [jg/dL
blood Pb)c

All-cause: 1.59 (1.02, 2.49)

tLanphear et al. (2018)
United States

1988-1994 mortality follow-
up in 2011

-19 yr of follow-up (IQR
17.6-21.0 yr)

Cohort

NHANES III
n = 14,289 >20 yr

Average individual
born -1947

Blood (GFAAS with
Zeeman correction)
(pg/dl_)

Geometric Mean 2.71
Geometric SE 1.31
10th percentile 1.0
90th percentile 6.7

Age of measurement
Mean 44.1

All-cause, CVD, and Cox proportional hazards
IHD mortality	regression analysis adjusting

forage, sex, household
income, ethnic origin, BMI,
smoking status, alcohol
consumption, physical activity,
concentration of cadmium in
urine, blood pressure, healthy
eating index tertiles, HbA1C,
and serum cholesterol

HR

All-cause: 1.06 (1.03, 1.09)
CVD: 1.10 (1.05, 1.15)
IHD: 1.14 (1.08, 1.20)

External Review Draft

9-140

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tvan Bemmel et al. (2011)

United States

1988-1994, follow-up
through 2007
-7.8 yr of follow-up for
those with low blood Pb
-7.5 yr of follow-up for
those with high blood Pb

Cohort

NHANES III
n = 3,349
Adult age >40 yr

Average individual
born -1932

Blood (GFAAS with
Zeeman correction)
(pg/dL)

Median
<5 pg/dL 2.6
>5 pg/dL 7.5

Age of measurement
<5 pg/dL 57
>5 pg/dL 61

All-cause and CVD Cox proportional hazards

mortality

adjusting for age, education,
sex, smoking status, and
race/ethnicity

HR

All-cause

All: 1.04 (0.98, 1.10)
ALADGG 1.03 (0.98, 1.08)
ALADCG/GG 1.09 (0.93, 1.28)

tDuan et al. (2020)

United States

1999-2014, follow-up
through end of 2015
- 7.1 yr of follow-up

NHANES
n = 18,602
Age >20 yr

Average individual
born -1960

Blood (ICP-MS) (pg/dL)d
Median (IQR)

1.49 (0.93, 2.31)

Age of measurement
Mean (SD) 45.9 (0.3)

All-cause mortality

Poisson regression analyses
adjusted for: sex, age,
ethnicity, education, poverty-
income-ratio (PIR), cotinine
category, BMI, physical
activity, hypertension, and
diabetes

RR

All-cause: 1.39 (1.28, 1.51)

Cohort

External Review Draft

9-141

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tBvun et al. (2020)

Korea

2007-2015, mortality
follow-up in 2018 (between
3-11 yr of follow-up)

Cohort

KNHANES
n = 7,308

Individuals with a
BLL less than
10 [jg/dL, who were
aged 30 yr and over
at the baseline
examination, and
who were not
diagnosed with
cancer or IHD

Average individual
born in or before
-1981

Blood (GFAAS with
Zeeman background
correction) (pg/dL)

Geometric mean: 2.26
Blood Pb tertiles:

T1
T2
T3

<1.91

1.91-2.71

>2.71

Age at measurement:
>30 yr

All-cause mortality

Cox proportional hazard
models adjusted for age and
sex, household income,
education, occupation,
smoking status, drinking
frequency, BMI, and physical
activity, high-lead-containing
food intake (grains,
vegetables, and seafood)

HRC

T1
T2
T3

Reference
2.02 (1.20,
1.91 (1.13,

3.40)
3.23)

External Review Draft

9-142

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tLin etal. (2011)

Taiwan

Years not reported
Cohort (18 mo of follow-up)

n = 927

Taiwanese adult
patients with end-
stage renal disease
(ERSD) on
hemodialysis for
>6 mo, age >18

Baseline blood Pb
(ETAAS) (pg/dL)
Mean: 11.5
Median: 10.4

All-cause, and

Infection-cause

mortality

T1
T2
T3

<8.51

8.51-12.64
<12.64

Age of measurement
Mean (SD) 55.2 (13.5)

Multivariate Cox model
adjusting forage, previous
cardiovascular diseases
(stroke, Ml, PID, congestive
heart failure (CHF)), education
level, hemodialysis vintage,
using fistula, normalized
protein catabolic rate,
hemoglobin, serum albumin,
creatinine, cardiothoracic
ratio, and logarithmic
transformation of high-
sensitivity C-reactive protein
(CRP)

HR(T1: Referent)0

All-cause

T2 2.69 (0.47, 3.44)
T3 4.70 (1.92, 11.49)

Infection-cause
T2 4.33 (0.35, 6.54)
T3 5.35 (1.38, 20.83)

Hemoglobin-corrected:

All-cause:

T2: 3.52 (0.41, 5.01)
T3: 4.98 (1.86, 13.33)

Infection-cause:
T2: 3.02 (0.23, 2.07)
T3: 4.72 (1.27, 17.54)

External Review Draft

9-143

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tTonelli etal. (2018)
Canada

Cohort (2 yr of follow-up)

n = 1,278

Patients on incident
hemodialysis
>18 yr

Plasma Pb (ICP-MS)

(pg/dL)

Deciles

1	0.06

2	0.19

3	0.28

4	0.35

5	0.44

6	0.55

7	0.68

8	0.83

9	1.08

10	1.74

All-cause mortality

Logistic regression adjusting
for age, sex, race/ethnicity,
unemployment prior to
dialysis, yr dialysis initiated,
dialysis duration, predialysis
care, arteriovenous access,
comorbidities (atrial fibrillation,
Ml, BMI, cancer,
cerebrovascular disease,
CHF, lung disease, diabetes,
dementia, hypertension, liver
disease, peripheral vascular
disease, psychiatric disease,
substance misuse), albumin,
and creatinine.

*AII variables were considered
candidate variables and were
included based on stepwise
regression results

Authors indicate a null
relationship between blood
Pb deciles and all-cause
mortality; quantitative
results not reported

External Review Draft

9-144

DRAFT: Do not cite or quote

-------
Reference and Study
Design

Study Population Exposure Assessment

Outcome

Confounders

Effect Estimates and 95%
Clsa

tHollinasworth and Rudik
(2021) United States

Quasi-experimental design

Elderly population
(>65 yr)

Assessed the
change in deaths
(National Vital
Statistics System)
occuring among this
age group before
and after the
phaseout of leaded
gasoline in
professional racing
(NASCAR, ARCA).

County-level blood Pb
measurements in
children

All-cause mortality

Difference-in-difference
approach controlling for SES
at the county level (median
income, unemployment rates,
percent minority population),
TRI Pb emissions data

Decline in age-standardized
mortality rate per 100,000
population

Race counties: 91
Border counties: 38

Compared mortality
rates in race-
counties to
bordering counties

Average individual
born in or before
-1942

ARCA = Automobile Racing Club of America; BLL = blood lead level; BMD = bone mineral density; BMI = body mass index; CHF = congestive heart failure; CI = confidence interval;
CHF = congestive heart failure; CRP = C-reactive protein; CVD = cardiovascular disease; ERSD = end-stage renal disease; ETAAS = electrothermal atomic absorption spectrometry;
GFAAS = graphite furnace atomic absorption spectrometry; GFR = glomerular filtration rate; HR = hazard ratio; ICP-MS = inductively coupled plasma mass spectrometry;
IHD = ischemic heart disease; IQR = interquartile range; KNHANES = Korean National Health and Nutrition Examination Survey; Ml = myocardial infarction; mo = month(s); NASCAR
= National Association for Stock Car Auto Racing; NHANES = National Health and Nutrition Examination Survey; Pb = lead; PIR = poverty-income-ratio; RR = risk ratio; SD = standard
deviation; SES = socioeconomic status, T = fertile; TC = total cholesterol; wk = week(s); yr = year(s).

aEffect estimates are standardized to a 1 |jg/dL increase in BLL or a 10 |jg/g increase in bone Pb level, unless otherwise noted. For studies that report results corresponding to a
change in log-transformed Pb biomarkers, effect estimates are assumed to be linear within the 10th to 90th percentile interval of the biomarker and standardized accordingly.

b Blood Pb analysis method unclear, assumed based on data source.

0 Unable to be standardized.

d Units assumed to be |jg/dL (written as |jg/L in the paper).

fStudies published since the 2013 Integrated Science Assessment for Lead.

External Review Draft

9-145

DRAFT: Do not cite or quote

-------
External Review Draft	9-146	DRAFT: Do not cite or quote

-------
9.10 References

Abbas. S: Khan. K: Khan. MP: Nagar. GK: Tewari. D: Maurva. SK: Dubev. J: Ansari. NG: Bandvopadhvav. S:

Chattopadhvav. N. (2013). Developmental exposure to As, Cd, and Pb mixture diminishes skeletal growth
and causes osteopenia at maturity via osteoblast and chondrocyte malfunctioning in female rats. Toxicol
Sci 134: 207-220. http://dx.doi.org/10.1093/toxsci/kft093.

Abu-Khudir. R: Habieb. ME: Mohamed. MA: Hawas. AM: Mohamed. TM. (2017). Anti-apoptotic role of spermine
against lead and/or gamma irradiation-induced hepatotoxicity in male rats. Environ Sci Pollut Res Int 24:
24272-24283. http://dx.doi.org/10.1007/sll356-017-0069-Q.

Adegbesan. BO: Adenuga. GA. (2007). Effect of lead exposure on liver lipid peroxidative and antioxidant defense
systems of protein-undernourished rats. Biol Trace Elem Res 116: 219-225.
http://dx.doi.org/10.1007/BF02685932.

Ademuviwa. O: Agarwal. R: Chandra. R: Behari. JR. (2009). Lead-induced phospholipidosis and cholesterogenesis
in rat tissues. Chem Biol Interact 179: 314-320. http://dx.doi.org/10.1016/i.cbi.2008.10.057.

Almasmoum. H: Refaat. B: Ghaith. MM: Almaimani. RA: Idris. S: Ahmad. J: Abdelghanv. AH: Basalamah. MA:
El-Boshv. M. (2019). Protective effect of VitaminD3 against lead induced hepatotoxicity, oxidative stress,
immunosuppressive and calcium homeostasis disorders in rat. Environ Toxicol Pharmacol 72: 103246.
http://dx.doi.Org/10.1016/i.etap.2019.103246.

Alvarez-Lloret. P: Lee. CM: Conti. MI: Terrizzi. AR: Gonzalez-Lopez. S: Martinez. MP. (2017). Effects of chronic
lead exposure on bone mineral properties in femurs of growing rats. Toxicology 377: 64-72.
http://dx.doi.Org/10.1016/i.tox.2016.ll.017.

Amos-Kroohs. RM: Graham. PL: Grace. CE: Braun. AA: Schaefer. TL: Skelton. MR: Vorhees. CV: Williams. MT.
(2016). Developmental stress and lead (Pb): Effects of maternal separation and/or Pb on corticosterone,
monoamines, and blood Pb in rats. Neurotoxicology 54: 22-33.
http://dx.doi.Org/10.1016/i.neuro.2016.02.011.

Andielkovic. M: Diordievic. AB: Antoniievic. E: Antoniievic. B: Stanic. M: Kotur-Stevulievic. J: Spasoievic-

Kalimanovska. V: Jovanovic. M: Boricic. N: Wallace. D: Bulat. Z. (2019). Toxic effect of acute cadmium
and lead exposure in rat blood, liver, and kidney. Int J Environ Res Public Health 16: 274.
http://dx.doi.org/10.3390/iierphl6020274.

Arguello. G: Balboa. E: Arrese. M: Zanlungo. S. (2015). Recent insights on the role of cholesterol in non-alcoholic
fatty liver disease [Review]. Biochim Biophys Acta 1852: 1765-1778.
http://dx.doi.Org/10.1016/i.bbadis.2015.05.015.

Arora. M: Weuve. J: Weisskopf. MG: Sparrow. D: Nie. H: Garcia. RI: Hu. H. (2009). Cumulative lead exposure and
tooth loss in men: The normative aging study. Environ Health Perspect 117: 1531-1534.
http://dx.doi.org/10.1289/ehp.0900739.

Barkaoui. T: Hamimed. S: Bellamine. H: Bankaii. I: Sleimi. N: Landoulsi. A. (2020). Alleviated actions of Plantago
albicans extract on lead acetate-produced hepatic damage in rats through antioxidant and free radical
scavenging capacities. J Med Food 23: 1201-1215. http://dx.doi.org/10.1089/imf.2019.0246.

Beier. EE: Holz. JD: Sheu. TJ: Puzas. JE. (2016). Elevated lifetime lead exposure impedes osteoclast activity and
produces an increase in bone mass in adolescent mice. Toxicol Sci 149: 277-288.
http://dx.doi.org/10.1093/toxsci/kfv234.

External Review Draft

9-147

DRAFT: Do not cite or quote

-------
Beier. EE: Sheu. TJ: Dang. D: Holz. JD: Ubavawardena. R: Babii. P: Puzas. JE. (2015). Heavy metal ion regulation
of gene expression: Mechanisms by which lead inhibits osteoblastic bone-forming activity through
modulation of the Wnt/(3-catenin signaling pathway. J Biol Chem 290: 18216-18226.
http://dx.doi.org/10.1074/ibc.Ml 14.629204.

Beier. EE: Sheu. TJ: Resseguie. EA: Takahata. M: Awad. HA: Corv-Slechta. DA: Puzas. JE. (2017). Sclerostin
activity plays a key role in the negative effect of glucocorticoid signaling on osteoblast function in mice.
Bone Res 5: 17013. http://dx.doi.org/10.1038/boneres.2017.13.

Berrahal. AA: Lasram. M: El Eli. N: Kerkeni. A: Gharbi. N: El-Fazaa. S. (2011). Effect of age-dependent exposure
to lead on hepatotoxicity and nephrotoxicity in male rats. Environ Toxicol 26: 68-78.
http://dx.doi.org/10.1002/tox.2053Q.

Biswas. NM: Ghosh. PK. (2006). Protection of adrenal and male gonadal functions by androgen in lead-treated rats.
Kathmandu Univ Med J 4: 218-221.

Bulka. CM: Perskv. VW: Daviglus. ML: Durazo-Arvizu. RA: Argos. M. (2019). Multiple metal exposures and

metabolic syndrome: A cross-sectional analysis of the National Health and Nutrition Examination Survey
2011-2014. Environ Res 168: 397-405. http://dx.doi.Org/10.1016/i.envres.2018.10.022.

Bvun. G: Kim. S: Kim. SY: Park. D: Shin. MJ: Oh. H: Lee. JT. (2020). Blood lead concentrations and mortality in
Korean adults: The Korea National Health and Nutrition Examination Survey with mortality follow-up. Int
J Environ Res Public Health 17: 6898. http://dx.doi.org/10.3390/iierphl7186898.

Campbell. JR: Auinger. P. (2007). The association between blood lead levels and osteoporosis among adults-results
from the Third National Health and Nutrition Examination Survey (NHANES III). Environ Health Perspect
115: 1018-1022. http://dx.doi.org/10.1289/ehp.9716.

Chen. A: Kim. SS: Chung. E: Dietrich. K. (2013). Thyroid hormones in relation to lead, mercury, and cadmium
exposure in the National Health and Nutrition Examination Survey, 2007-2008. Environ Health Perspect
121: 181-186. http://dx.doi.org/10.1289/ehp.1205239.

Chen. HS: Tsai. YC: Chen. KK: Tseng. YC: Hsu. KJ. (2012). Detrimental effects of maternal lead exposure during
pregnancy and lactation on molar development in the young rat. Bull Environ Contam Toxicol 89: 240-244.
http://dx.doi.org/10.1007/s00128-012-Q683-v.

Chen. YR: Xu. XJ: Zeng. ZJ: Lin. XO: Oin. OL: Huo. X. (2019). Blood lead and cadmium levels associated with

hematological and hepatic functions in patients from an e-waste-polluted area. Chemosphere 220: 531-538.
http://dx.doi.Org/10.1016/i.chemosphere.2018.12.129.

Cho. GJ: Park. HT: Shin. JH: Hur. JY: Kim. SH: Lee. KW: Kim. T. (2012). The relationship between blood mercury
level and osteoporosis in postmenopausal women. Menopause 19: 576-581.
http://dx.doi.org/10.1097/gme.0b013e3182377294.

Christensen. KLY. (2012). Metals in blood and urine, and thyroid function among adults in the United States 2007-
2008. Int J Hyg Environ Health 216: 624-632. http://dx.doi.Org/10.1016/i.iiheh.2012.08.005.

Christensen. KLY: Carrico. CK: Sanval. AJ: Gennings. C. (2013). Multiple classes of environmental chemicals are
associated with liver disease: NHANES 2003-2004. Int J Hyg Environ Health 216: 703-709.
http://dx.doi.Org/10.1016/i.iiheh.2013.01.005.

Chrostek. L: Supronowicz. L: Panasiuk. A: Cvlwik. B: Gruszewska. E: Flisiak. R. (2014). The effect of the severity
of liver cirrhosis on the level of lipids and lipoproteins. Clin Exp Med 14: 417-421.
http://dx.doi.org/10.1007/sl0238-013-Q262-5.

Chung. SM: Moon. JS: Yoon. JS: Won. KC: Lee. HW. (2020). The sex-specific effects of blood lead, mercury, and
cadmium levels on hepatic steatosis and fibrosis: Korean nationwide cross-sectional study. J Trace Elem
Med Biol 62: 126601. http://dx.doi.Org/10.1016/i.itemb.2020.126601.

Corv-Slechta. DA: Merchant-Borna. K: Allen. JL: Liu. S: Weston. D: Conrad. K. (2013). Variations in the nature of
behavioral experience can differentially alter the consequences of developmental exposures to lead,
prenatal stress, and the combination. Toxicol Sci 131: 194-205. http://dx.doi.org/10.1093/toxsci/kfs26Q.

External Review Draft

9-148

DRAFT: Do not cite or quote

-------
de Figueiredo. FAT: Gerlach. RF: da Veiga. MAM. S: Nakadi. FY: Ramos. J: Kawakita. ER: Guerra. CD: Issa.
JPM. (2014). Reduced bone and body mass in young male rats exposed to lead. BioMed Res Int 2014:
571065. http://dx.doi.org/10.1155/2014/571Q65.

Duaa W: Xu. C: Liu. O: Xu. J: Weng. Z: Zhang. X: Basnet. TB: Dahal. M: Gu. A. (2020). Levels of a mixture of
heavy metals in blood and urine and all-cause, cardiovascular disease and cancer mortality: A population-
based cohort study. Environ Pollut 263: 114630. http://dx.doi.Org/10.1016/i.envpol.2020.114630.

Dumkova. J: Smutna. T: Vrlikova. L: Docekal. B: Kristekova. D: Vecera. Z: Husakova. Z: Jakesova. V: Jedlickova.
A: Mikuska. LP: Alexa. T: Coufalik. P: Tvrdonova. A: Krumal. K: Vaculovic. DT: Kanickv. GV: Hampl.
GA: Buchtova. MM. (2020a). A clearance period after soluble lead nanoparticle inhalation did not
ameliorate the negative effects on target tissues due to decreased immune response. International Journal of
Molecular Sciences 21: 8738. http://dx.doi.org/10.3390/iims21228738.

Dumkova. J: Smutna. T: Vrlikova. L: Kotasova. H: Docekal. B: Capka. L: Tvrdonova. M: Jakesova. V: Pelkova. V:
Krumal. K: Coufalik. P: Mikuska. P: Vecera. Z: Vaculovic. T: Husakova. Z: Kanickv. V: Hampl. A:
Buchtova. M. (2020b). Variability in the clearance of lead oxide nanoparticles is associated with alteration
of specific membrane transporters. ACS Nano 14: 3096-3120. http://dx.doi.org/10.1021/acsnano.9b08143.

Dumkova. J: Smutna. T: Vrlikova. L: Le Coustumer. P: Vecera. Z: Docekal. B: Mikuska. P: Capka. L: Fictum. P:
Hampl. A: Buchtova. M. (2017). Sub-chronic inhalation of lead oxide nanoparticles revealed their broad
distribution and tissue-specific subcellular localization in target organs. Part Fibre Toxicol 14:55.
http://dx.doi.org/10.1186/sl2989-017-Q236-v.

Egan. KB: Cornwell. CR: Courtney. JG: Ettinger. AS. (2021). Blood lead levels in U.S. children ages 1-11 years,
1976-2016. Environ Health Perspect 129: 37003. http://dx.doi.org/10.1289/EHP7932.

El-Tantawv. WH. (2016). Antioxidant effects of Spirulina supplement against lead acetate-induced hepatic injury in
rats. J Tradit Complement Med 6: 327-331. http://dx.doi.Org/10.1016/i.itcme.2015.02.001.

Erie. JC: Good. JA: Butz. JA. (2009). Excess lead in the neural retina in age-related macular degeneration. Am J
Ophthalmol 148: 890-894. http://dx.doi.Org/10.1016/i.aio.2009.07.001.

Ettinger. AS: Bovet. P: Plange-Rhule. J: Forrester. TE: Lambert. EV: Lupoli. N: Shine. J: Dugas. LR: Shoham. D:
Durazo-Arvizu. RA: Cooper. RS: Luke. A. (2014). Distribution of metals exposure and associations with
cardiometabolic risk factors in the "Modeling the Epidemiologic Transition Study". Environ Health 13: 90.
http://dx.doi.org/10.1186/1476-069X-13-90.

Faulk. C: Barks. A: Sanchez. BN: Zhang. ZZ: Anderson. OS: Peterson. KE: Dolinov. DC. (2014). Perinatal lead
(Pb) exposure results in sex-specific effects on food intake, fat, weight, and insulin response across the
murine life-course. PLoS ONE 9: el04273. http://dx.doi.org/10.1371/iournal.pone.0104273.

Gao. W: Guo. Y: Wang. L: Jiang. Y: Liu. Z: Lin. H. (2020). Ameliorative and protective effects of fucoidan and
sodium alginate against lead-induced oxidative stress in Sprague Dawley rats. Int J Biol Macromol 158:
662-669. http://dx.doi.Org/10.1016/i.iibiomac.2020.04.192.

Graham. PL: Grace. CE: Braun. AA: Schaefer. TL: Skelton. MR: Tang. PH: Vorhees. CV: Williams. MT. (2011).
Effects of developmental stress and lead (Pb) on corticosterone after chronic and acute stress, brain
monoamines, and blood Pb levels in rats. Int J Dev Neurosci 29: 45-55.
http://dx.doi.Org/10.1016/i.iidevneu.2010.09.008.

Gump. BB: Stewart. P: Reihman. J: Lonkv. E: Darvill. T: Parsons. PJ: Granger. DA. (2008). Low-level prenatal and
postnatal blood lead exposure and adrenocortical responses to acute stress in children. Environ Health
Perspect 116: 249-255. http://dx.doi.org/10.1289/ehp.10391.

Guo. X: Yang. O: Zhang. W: Chen. Y: Ren. J: Gao. A. (2019). Associations of blood levels of trace elements and
heavy metals with metabolic syndrome in Chinese male adults with microRNA as mediators involved.
Environ Pollut 248: 66-73. http://dx.doi.Org/10.1016/i.envpol.2019.02.015.

Han. DH: Lee. HJ: Lim. S. (2013). Smoking induced heavy metals and periodontitis: Findings from the Korea
National Health and Nutrition Examination Surveys 2008-2010. J Clin Periodontal 40: 850-858.
http://dx.doi.org/10.llll/icpe.12133.

External Review Draft

9-149

DRAFT: Do not cite or quote

-------
Hansen. AF: Simic. A: Asvold. BO: Romundstad. PR: Midthiell. K: Svversen. T: Flaten. TP. (2017). Trace elements
in early phase type 2 diabetes mellitus-A population-based study. The HUNT study in Norway. J Trace
Elem Med Biol 40: 46-53. http://dx.doi.Org/10.1016/i.itemb.2016.12.008.

Hasanein. P: Kazemian-Mahtai. A: Khodadadi. I. (2016). Bioactive peptide carnosin protects against lead acetate-
induced hepatotoxicity by abrogation of oxidative stress in rats. Pharmaceutical Biology 54: 1458-1464.
http://dx.doi.org/10.3109/13880209.2015.110470Q.

Hollingsworth. A: Rudik. I. (2021). The effect of leaded gasoline on elderly mortality: Evidence from regulatory
exemptions. Am Econ J EconPolicy 13: 345-373. http://dx.doi.org/10.1257/pol.20190654.

Holz. JD: Beier. E: Sheu. TJ: Ubavawardena. R: Wang. MN: Sampson. ER: Rosier. RN: Zuscik. M: Puzas. JE.

(2012). Lead induces an osteoarthritis-like phenotype in articular chondrocytes through disruption of TGF-
(3 signaling. J Orthop Res 30: 1760-1766. http://dx.doi.org/10.1002/ior.22117.

Hortoa CJ: Weng. HY: Wells. EM. (2019). Association between blood lead level and subsequent Alzheimer's
disease mortality. Environmental Epidemiology 3: e045.
http://dx.doi.org/10.1097/EE9.0000000000000Q45.

Hwang. HS: Lee. SB: Jee. D. (2015). Association between blood lead levels and age-related macular degeneration.
PLoS ONE 10: e0134338. http://dx.doi.org/10.1371/iournal.pone.0134338.

Jung. SJ: Lee. SH. (2019). Association between three heavy metals and dry eye disease in Korean adults: Results of
the Korean National Health and Nutrition Examination Survey. Korean J Ophthalmol 33: 26-35.
http://dx.doi.org/10.3341/kio.2018.0Q65.

Jurczuk. M: Brzoska. MM: Moniuszko-Jakoniuk. J. (2007). Hepatic and renal concentrations of vitamins E and C in
lead- and ethanol-exposed rats. An assessment of their involvement in the mechanisms of peroxidative
damage. Food Chem Toxicol 45: 1478-1486. http://dx.doi.Org/10.1016/i.fct.2007.02.007.

Jurczuk. M: Moniuszko-Jakoniuk. J: Brzoska. MM. (2006). Involvement of some low-molecular thiols in the
peroxidative mechanisms of lead and ethanol action on rat liver and kidney. Toxicology 219: 11-21.
http://dx.doi.Org/10.1016/i.tox.2005.10.022.

Kahn. LG: Liu. XH: Raiovic. B: Popovac. D: Oberfield. S: Graziano. JH: Factor-Litvak. P. (2014). Blood lead

concentration and thyroid function during pregnancy: Results from the Yugoslavia Prospective Study of
Environmental Lead Exposure. Environ Health Perspect 122: 1134-1140.
http://dx.doi.org/10.1289/ehp.1307669.

Khalil. N: Wilson. JW: Talbott. EO: Morrow. LA: Hochberg. MC: Hillier. TA: Muldoon. SB: Cummings. SR:

Caulev. JA. (2009). Association of blood lead concentrations with mortality in older women: a prospective
cohort study. Environ Health 8: 15. http://dx.doi.org/10.1186/1476-069x-8-15.

Khotimchenko. MY: Kolenchenko. EA. (2007). Efficiency of low-esterified pectin in toxic damage to the liver

inflicted by lead treatment. Bull Exp Biol Med 144: 60-62. http://dx.doi.org/10.1007/sl0517-007-0254-Q.

Kim. Y: Lee. BK. (2013). Association between blood lead and mercury levels and periodontitis in the Korean

general population: Analysis of the 2008-2009 Korean National Health and Nutrition Examination Survey
data. Int Arch Occup Environ Health 86: 607-613. http://dx.doi.org/10.1007/s00420-012-Q796-v.

Kim. YS: Ha. M: Kwon. HJ: Kim. HY: Choi. YH. (2017). Association between Low blood lead levels and increased
risk of dental caries in children: A cross-sectional study. BMC Oral Health 17: 42.
http://dx.doi.org/10.1186/sl2903-017-Q335-z.

Koiima. M: Ashino. T: Yoshida. T: Iwakura. Y: Degawa. M. (2012). Involvement of interleukin-1 in lead nitrate-

induced hypercholesterolemia in mice. Biol Pharm Bull 35: 246-250. http://dx.doi.org/10.1248/bpb.35.246.

Kosik-Bogacka. PI: Baranowska-Bosiacka. I: Marchlewicz. M: Kolasa. A: Jakubowska. K: Olszewska. M:

Lanocha. N: Wiernicki. I: Millo. B: Wiszniewska. B: Chlubek. D. (2011). The effect of L-ascorbic acid
and/or tocopherol supplementation on electrophysiological parameters of the colon of rats chronically
exposed to lead. Med Sci Monit 17: BR16-BR26. http://dx.doi.org/10.12659/MSM.881323.

External Review Draft

9-150

DRAFT: Do not cite or quote

-------
Krieg. EF. Jr. (20191. The relationships between blood lead levels and serum thyroid stimulating hormone and total
thyroxine in the third National Health and Nutrition Examination Survey. J Trace Elem Med Biol 51: ISO-
IS?. http://dx.doi.Org/10.1016/i.itemb.2018.10.010.

Kupraszewicz. E: Brzoska. MM. (20131. Excessive ethanol consumption under exposure to lead intensifies disorders
in bone metabolism: a study in a rat model. Chem Biol Interact 203: 486-501.
http://dx.doi.Org/10.1016/i.cbi.2013.01.002.

Kupsco. A: Kioumourtzoglou. MA: Just. AC: Amarasiriwardena. C: Estrada-Gutierrez. G: Cantoral. A: Sanders.
AP: Braun. JM: Svensson. K: Brennan. KJM: Oken. E: Wright. RO: Baccarelli. AA: Tellez-Roio. MM.
(2019). Prenatal metal concentrations and childhood cardiometabolic risk using Bayesian kernel machine
regression to assess mixture and interaction effects. Epidemiology 30: 263-273.
http://dx.doi.org/10.1097/EDE.000000000000Q962.

Laamech. J: El-Hilalv. J: Fetoui. H: Chtourou. Y: Gouitaa. H: Tahraoui. A: Lvoussi. B. (2017). Berberis vulgaris L.
effects on oxidative stress and liver injury in lead-intoxicated mice. J Complement Integr Med 14.
http://dx.doi.org/10.1515/icim-2015-0079.

Lanphear. BP: Rauch. S: Auinger. P: Allen. RW: Hornung. RW. (2018). Low-level lead exposure and mortality in
US adults: A population-based cohort study. The Lancet Public Health 3: el77-el84.
http://dx.doi.org/10.1016/S2468-2667(18)30025-2.

Lee. BK: Kim. Y. (2012). Association between bone mineral density and blood lead level in menopausal women:
Analysis of 2008-2009 Korean National Health and Nutrition Examination Survey data. Environ Res 115:
59-65. http://dx.doi.Org/10.1016/i.envres.2012.03.010.

Lee. BK: Kim. Y. (2013). Blood cadmium, mercury, and lead and metabolic syndrome in South Korea: 2005-2010
Korean National Health and Nutrition Examination Survey. Am J Ind Med 56: 682-692.
http://dx.doi.org/10.1002/aiim.221Q7.

Lee. BK: Kim. Y. (2016). Association of blood cadmium level with metabolic syndrome after adjustment for
confounding by serum ferritin and other factors: 2008-2012 Korean National Health and Nutrition
Examination Survey. Biol Trace Elem Res 171: 6-16. http://dx.doi.org/10.1007/sl2011-015-Q499-9.

Lee. HS: Park. T. (2018). Nuclear receptor and VEGF pathways for gene-blood lead interactions, on bone mineral
density, in Korean smokers. PLoS ONE 13: e0193323. http://dx.doi.org/10.1371/iournal.pone.0193323.

Lee. J: Vali. Y: Boursier. J: Puffin. K: Verheii. J: Brosnan. MJ: Zwinderman. K: Anstee. OM: Bossuvt. PM:
Zafarmand. MH. (2020). Accuracy of cytokeratin 18 (M30 and M65) in detecting non-alcoholic
steatohepatitis and fibrosis: A systematic review and meta-analysis. PLoS ONE 15: e0238717.
http://dx.doi.org/10.1371/iournal.pone.0238717.

Lee. SH: Kang. EM: Kim. GA: Kwak. SW: Kim. JM: Bae. HW: Seong. GJ: Kim. CY. (2016). Three Toxic Heavy
Metals in Open-Angle Glaucoma with Low-Teen and High-Teen Intraocular Pressure: A Cross-Sectional
Study from South Korea. PLoS ONE 11: eO 164983. http://dx.doi.org/10.1371/iournal.pone.0164983.

Leem. AY: Kim. SK: Chang. J: Kang. YA: Kim. YS: Park. MS: Kim. SY: Kim. EY: Chung. KS: Jung. JY. (2015).

Relationship between blood levels of heavy metals and lung function based on the Korean National Health
and Nutrition Examination Survey IV-V. Int J Chron Obstruct Pulmon Dis 10: 1559-1570.
http://dx.doi.org/10.2147/COPD.S86182.

Li. B: Jin. D: Yu. S: Evivie. SE: Muhammad. Z: Huo. G: Liu. F. (2017). In vitro and in vivo evaluation of
Lactobacillus delbrueckii subsp. bulgaricus KLDS 1.0207 for the alleviative effect on lead toxicity.

Nutrients 9: 845. http://dx.doi.org/10.3390/nu9080845.

Li. D: Liang. H: Li. Y: Zhang. J: Qiao. L: Luo. H. (2020a). Allicin Alleviates Lead-Induced Bone Loss by

Preventing Oxidative Stress and Osteoclastogenesis Via SIRTl/FOXOl Pathway in Mice. Biol Trace Elem
Res 199: 237-243. http://dx.doi.org/10.1007/sl2011-020-Q2136-5.

Li. XM: Li. RJ: Yan. JM: Song. Y: Huo. J: Lan. Z: Chen. JY: Zhang. LS. (2020b). Co-exposure of cadmium and
lead on bone health in a southwestern Chinese population aged 40-75 years. J Appl Toxicol 40: 352-362.
http://dx.doi.org/10.1002/iat.3908.

External Review Draft

9-151

DRAFT: Do not cite or quote

-------
Lim. HS: Lee. HH: Kim. TH: Lee. BR. (20161. Relationship between heavy metal exposure and bone mineral
density in Korean adult. JBM 23: 223-231. http://dx.doi.Org/10.11005/ibm.2016.23.4.223.

Lin. JL: Lin-Tan. DT: Hsu. CW: Yen. TH: Chen. KH: Hsu. HH: Ho. TC: Hsu. KH. (2011). Association of blood
lead levels with mortality in patients on maintenance hemodialysis. Am J Med 124: 350-358.
http://dx.doi.Org/10.1016/i.amimed.2010.10.022.

Lin. SC: Singh. K: Lin. SC. (20151. Association Between Body Levels of Trace Metals and Glaucoma Prevalence.
JAMA Ophthalmol 133: 1144-1150. http://dx.doi.org/10.1001/iamaophthalmol.2Q15.2438.

Little. BB: Ignasiak. Z: Slawinska. T: Poslusznv. P: Malina. RM: Wiegman. PL. (2017). Blood lead levels,
pulmonary function and agility in Polish schoolchildren. Ann Hum Biol 44: 723-728.
http://dx.doi.org/10.1080/03014460.2Q17.1387284.

Liu. CM: Ma. JO: Sun. YZ. (2011). Protective role of puerarin on lead-induced alterations of the hepatic glutathione
antioxidant system and hyperlipidemia in rats. Food Chem Toxicol 49: 3119-3127.
http://dx.doi.Org/10.1016/i.fct.2011.09.007.

Liu. CM: Ma. JO: Sun. YZ. (2012). Puerarin protects the rat liver against oxidative stress-mediated DNA damage
and apoptosis induced by lead. Exp Toxicol Pathol 64: 575-582.
http://dx.doi.Org/10.1016/i.etp.2010.ll.016.

Liu. CM: Zheng. GH: Ming. OL: Sun. JM: Cheng. C. (2013). Protective effect of quercetin on lead-induced

oxidative stress and endoplasmic reticulum stress in rat liver via the IRE1/JNK and PI3K/Akt pathway.

Free Radic Res 47: 192-201. http://dx.doi.org/10.3109/10715762.2012.76Q198.

Liu. Y: Ettinger. AS: Tellez-Roio. M: Sanchez. BN: Zhang. ZZ: Cantoral. A: Hu. H: Peterson. KE. (2020). Prenatal
lead exposure, type 2 diabetes, and cardiometabolic risk factors in Mexican children at age 10-18 years. J
Clin Endocrinol Metab 105: 210-218. http://dx.doi.org/10.1210/clinem/dgz038.

Long. M: Liu. Y: Cao. Y: Wang. N: Dang. M: He. JB. (2016). Proanthocyanidins attenuation of chronic lead-
induced liver oxidative damage in Kunming mice via the Nrf2/ARE pathway. Nutrients 8: 656.
http://dx.doi.org/10.3390/nu8100656.

Luo. JH: Hendrvx. M. (2014). Relationship between blood cadmium, lead, and serum thyroid measures in US adults
- The National Health and Nutrition Examination Survey (NHANES) 2007-2010. Int J Environ Health Res
24: 125-136. http://dx.doi.org/10.1080/09603123.2013.80Q962.

Lustberg. M: Silbergeld. E. (2002). Blood lead levels and mortality. Arch Intern Med 162: 2443-2449.
http://dx.doi.org/10.1001/archinte.162.21.2443.

Ma. Y: Fu. D. a: Liu. Z. (2012). Effect of lead on apoptosis in cultured rat primary osteoblasts. Toxicol Ind Health
28: 136-146. http://dx.doi.org/10.1177/0748233711407956.

Mabrouk. A: Bel Hadi Salah. I: Chaieb. W: Ben Cheikh. H. (2016). Protective effect of thymoquinone against lead-
induced hepatic toxicity in rats. Environ Sci Pollut Res Int 23: 12206-12215.
http://dx.doi.org/10.1007/sll356-016-6419-5.

Machida. M: Sun. SJ: Qguma. E: Kavama. F. (2009). High bone matrix turnover predicts blood levels of lead among
perimenopausal women. Environ Res 109: 880-886. http://dx.doi.Org/10.1016/i.envres.2009.06.005.

Madrigal. JM: Perskv. V: Pappalardo. A: Argos. M. (2018). Association of heavy metals with measures of

pulmonary function in children and youth: Results from the National Health and Nutrition Examination
Survey (NHANES). Environ Int 121: 871-878. http://dx.doi.Org/10.1016/i.envint.2018.09.045.

Mazumdar. I: Goswami. K. (2014). Chronic exposure to lead: a cause of oxidative stress and altered liver function in
plastic industry workers in kolkata, India. Indian J Clin Biochem 29: 89-92.
http://dx.doi.org/10.1007/sl2291-013-Q337-9.

Mendv. A: Gasana. J: Vieira. ER. (2013). Low blood lead concentrations and thyroid function of American adults.
Int J Environ Health Res 23: 461 -473. http://dx.doi.org/10.1080/09603123.2Q12.755155.

External Review Draft

9-152

DRAFT: Do not cite or quote

-------
Menke. A: Muntner. P: Batuman. V: Silbergeld. EK: Guallar. E. (2006). Blood lead below 0.48 (imol/1 (10 jxg/dl)
and mortality among US adults. Circulation 114: 1388-1394.
http://dx.doi.org/10.1161/circulationaha.106.628321.

Moon. SS. (20131. Association of lead, mercury and cadmium with diabetes in the Korean population: The Korea
National Health and Nutrition Examination Survey (KNHANES) 2009-2010. Diabet Med 30: el43-el48.
http://dx.doi.org/10. Ill 1/dme. 12103.

Moon. SS. (20141. Additive effect of heavy metals on metabolic syndrome in the Korean population: The Korea
National Health and Nutrition Examination Survey (KNHANES) 2009-2010. Endocrine 46: 263-271.
http://dx.doi.org/10.1007/sl2020-013-0Q61-5.

Mosad. SM: Ghanem. AA: El-Fallal. HM: El-Kannishv. AM: El Baiomv. AA: Al-Diastv. AM: Arafa. LF. (2010).

Lens cadmium, lead, and serum vitamins C, E, and beta carotene in cataractous smoking patients. Curr Eye
Res 35: 23-30. http://dx.doi.org/10.3109/0271368090336288Q.

Moss. ME: Lanphear. BP: Auinger. P. (1999). Association of dental caries and blood lead levels. JAMA 281: 2294-
2298. http://dx.doi.org/10 1001 /iama.281.24.2294.

Nelson. AE: Chaudharv. S: Kraus. VB: Fang. F: Chen. JC: Schwartz. TA: Shi. XYA: Renner. JB: Stabler. TV:

Helmick. CG: Caldwell. K: Poole. AR: Jordan. JM. (201 la). Whole blood lead levels are associated with
bio markers of joint tissue metabolism in African American and white men and women: The Johnston
County Osteoarthritis Project. EnvironRes 111: 1208-1214.
http://dx.doi.Org/10.1016/i.envres.2011.08.002.

Nelson. AE: Shi. XA: Schwartz. TA: Chen. JC: Renner. JB: Caldwell. KL: Helmick. CG: Jordan. JM. (2011b).
Whole blood lead levels are associated with radiographic and symptomatic knee osteoarthritis: a cross-
sectional analysis in the Johnston County Osteoarthritis Project. Arthritis Res Ther 13: R37.
http://dx.doi.org/10.1186/ar3270.

Ngueta. G: Verner. MA: Fiocco. AJ: Lupien. S: Plusauellec. P. (2018). Blood lead levels and hypothalamic -
pituitary-adrenal function in middle-aged individuals. EnvironRes 160: 554-561.
http://dx.doi.Org/10.1016/i.envres.2017.10.032.

NHLBI (National Institutes of Health. National Heart Lung and Blood Institute). (2017). NHLBI fact book, fiscal
year 2012: Disease statistics. Available online at

https://web.archive.org/web/20170711012213/http://www.nhlbi.nih.gov/about/documents/factbook/2012/c
hapter4 (accessed August 23, 2017).

Nie. XM: Chen. Y: Chen. YC: Chen. C: Han. B: Li. O: Zhu. CF: Xia. FZ: Zhai. HL: Wang. NT: Ln. Yl. (2017).
Lead and cadmium exposure, higher thyroid antibodies and thyroid dysfunction in Chinese women.

Environ Pollut 230: 320-328. http://dx.doi.Org/10.1016/i.envpol.2017.06.052.

Obeng-Gvasi. E. (2019). Lead exposure and cardiovascular disease among young and middle-aged adults. Med Sci
7: 103. http://dx.doi.org/10.3390/medsci7110103.

Olchowik. G: Widomska. J: Tomaszewski. M: Gospodarek. M: Tomaszewska. M: Jagiello-Woitowicz. E. (2014).
The influence of lead on the biomechanical properties of bone tissue in rats. Ann Agric Environ Med 21:
278-281. http://dx.doi.org/10.56Q4/1232-1966.1108591.

Pak. YS: Oh. A: Kho. YL: Paek. D. (2012). Lung function decline and blood lead among residents nearby to

industrial complex. Int Arch Occup Environ Health 85: 951-959. http://dx.doi.org/10.1007/sQ0420-012-
0743-v.

Pal. PB: Sinha. K: Sil. PC. (2013). Mangiferin, a Natural Xanthone, Protects Murine Liver in Pb(II) Induced Hepatic
Damage and Cell Death via MAP Kinase, NF-kappa B and Mitochondria Dependent Pathways. PLoS ONE
8: e56894. http://dx.doi.org/10.1371/iournal.pone.0056894.

Pandva. CD: Pillai. PP: Gupta. SS. (2010). Lead and cadmium co-exposure mediated toxic insults on hepatic steroid
metabolism and antioxidant system of adult male rats. Biol Trace Elem Res 134: 307-317.
http://dx.doi.org/10.1007/sl2011-0Q9-8479-6.

External Review Draft

9-153

DRAFT: Do not cite or quote

-------
Park. S: Choi. NK. (20161. Associations of blood heavy metal levels with intraocular pressure. Ann Epidemiol 26:
546-550. http://dx.doi.Org/10.1016/i.annepidem.2016.07.002.

Park. S: Choi. NK. (20191. The relationships of blood lead level, body mass index, and osteoarthritis in

postmenopausal women. Maturitas 125: 85-90. http://dx.doi.Org/10.1016/i.maturitas.2019.04.215.

Park. SJ: Lee. JH: Woo. SJ: Kang. SW: Park. KH. (2015). Five heavy metallic elements and age-related macular

degeneration: Korean National Health and Nutrition Examination Survey, 2008-2011. Ophthalmology 122:
129-137. http://dx.doi.Org/10.1016/i.ophtha.2014.07.039.

Perez Aguilar. RC: Honore. SM: Genta. SB: Sanchez. SS. (2014). Hepatic fibrogenesis and transforming growth

factor/Smad signaling activation in rats chronically exposed to low doses of lead. J Appl Toxicol 34: 1320-
1331. http://dx.doi.org/10.1002/iat.2955.

Perkins. GA: Scott. R: Perez. A: Ellisman. MH: Johnson. JE: Fox. DA. (2012). Bcl-xL-mediated remodeling of rod
and cone synaptic mitochondria after postnatal lead exposure: Electron microscopy, tomography and
oxygen consumption. Mol Vis 18: 3029-3048.

Peters. JL: Kubzanskv. LP: Ikeda. A: Fang. SC: Sparrow. D: Weisskopf. MG: Wright. RO: Vokonas. P: Hu. H:

Schwartz. J. (2012). Lead concentrations in relation to multiple biomarkers of cardiovascular disease: The
Normative Aging Study. Environ Health Perspect 120: 361-366. http://dx.doi.org/10.1289/ehp. 1103467.

Pollack. AZ: Mumford. SL: Mendola. P: Perkins. NJ: Rotman. Y: Wactawski-Wende. J: Schisterman. EF. (2015).
Kidney biomarkers associated with blood lead, mercury, and cadmium in premenopausal women: A
prospective cohort study. J Toxicol Environ Health A 78: 119-131.
http://dx.doi.org/10.1080/15287394.2014.94468Q.

Pollack. AZ: Mumford. SL: Wactawski-Wende. J: Yeung. E: Mendola. P: Mattison. PR: Schisterman. EF. (2013).
Bone mineral density and blood metals in premenopausal women. Environ Res 120: 76-81.
http://dx.doi.Org/10.1016/i.envres.2012.06.001.

Oi. S: Zheng. H: Chen. C: Jiang. H. (2019). Pu-Zhong (Eucommia ulmoides Oliv.) Cortex Extract Alleviates Lead
Acetate-Induced Bone Loss in Rats. Biol Trace Elem Res 187: 172-180. http://dx.doi.org/10.10Q7/sl2011-
018-1362-6.

Rahman. A: Al-Awadi. AA: Khan. KM. (2018). Lead affects vitamin P metabolism in rats. Nutrients 10: 264.
http://dx.doi.org/10.3390/nulQ030264.

Reckziegel. P: Pias. VT: Benvegnu. PM: Boufleur. N: Barcelos. RCS: Segat. HJ: Pase. CS: Pos Santos. CMM:

Flores. EMM: Burger. ME. (2016). Antioxidant protection of gallic acid against toxicity induced by Pb in
blood, liver and kidney of rats. Toxicol Rep 3: 351-356. http://dx.doi.Org/10.1016/i.toxrep.2016.02.005.

Reddv. YS: Srivalliputturu. SB: Bharatrai. PK. (2018). The effect of lead (Pb) exposure and iron (Fe) deficiency on
intestinal lactobacilli, E. coli and yeast: A study in experimental rats. J Occup Health 60: 475-484.
http://dx.doi.org/10.1539/ioh.2017-0267-QA.

Reia. P: Makar. M: Visaria. A: Karanfilian. B: Rustgi. V. (2020). Blood lead level is associated with advanced liver
fibrosis inpatients with non-alcoholic fatty liver disease: A nationwide survey (NHANES 2011-2016). Ann
Hepatol 19: 404-410. http://dx.doi.Org/10.1016/i.aohep.2020.03.006.

Rhee. SY: Hwang. YC: Woo. JT: Sinn. PH: Chin. SO: Chon. S: Kim. YS. (2013). Blood lead is significantly

associated with metabolic syndrome in Korean adults: An analysis based on the Korea National Health and
Nutrition Examination Survey (KNHANES), 2008. Cardiovasc Piabetol 12: 9.
http://dx.doi.org/10.1186/1475-2840-12-9.

Rokadia. HK: Agarwal. S. (2013). Serum heavy metals and obstructive lung disease: Results from the National
Health and Nutrition Examination Survey. Chest 143: 388-397. http://dx.doi.org/10.1378/chest.12-0595.

Rossi-George. A: Virgolini. MB: Weston. P: Thiruchelvam. M: Corv-Slechta. PA. (2011). Interactions of lifetime
lead exposure and stress: Behavioral, neurochemical and HPA axis effects. Neurotoxicology 32: 83-99.
http://dx.doi.Org/10.1016/i.neuro.2010.09.004.

Saraiva. MCP: Taichman. RS: Braun. T: Nriagu. J: Eklund. SA: Burt. BA. (2007). Lead exposure and periodontitis
in US adults. J Periodontal Res 42: 45-52. http://dx.doi.Org/10.llll/i.1600-0765.2006.00913.x.

External Review Praft

9-154

PRAFT: Po not cite or quote

-------
Schober. SE: Mirel. LB: Graubard. BI: Brodv. DJ: Flegal. KM. (2006). Blood lead levels and death from all causes,
cardiovascular disease, and cancer: Results from the NHANES III Mortality Study. Environ Health
Perspect 114: 1538-1541. http://dx.doi.org/10.1289/ehp.9123.

Sharma. A: Sharma. V: Kansal. L. (20101. Amelioration of lead-induced hepatotoxicity by Allium sativum extracts
in Swiss albino mice. Libyan Journal of Medicine 5: 4621. http://dx.doi.org/10.4176/0911Q7.

Shen. XF: Huang. P: Fox. DA: Lin. Y: Zhao. ZH: Wang. W: Wang. JY: Liu. XO: Chen. JY: Luo. WJ. (2016). Adult
lead exposure increases blood-retinal permeability: A risk factor for retinal vascular disease.
Neurotoxicology 57: 145-152. http://dx.doi.Org/10.1016/i.neuro.2016.09.013.

Sheng. Z: Wang. S: Zhang. X: Li. X: Li. B: Zhang. Z. (2020). Long-Term Exposure to Low-Dose Lead Induced
Deterioration in Bone Microstructure of Male Mice. Biol Trace ElemRes 195: 491-498.
http://dx.doi.org/10.1007/sl2011-019-Q1864-7.

Shim. YH: Ock. JW: Kim. YJ: Kim. Y: Kim. SY: Kang. D. (2019). Association between heavy metals, bisphenol A,
volatile organic compounds and phthalates and metabolic syndrome. Int J Environ Res Public Health 16:
671. http://dx.doi.org/10.3390/iierphl6040671.

Simic. A: Hansen. AF: Asvold. BO: Romundstad. PR: Midthiell. K: Svversen. T: Flaten. TP. (2017). Trace element
status inpatients with type 2 diabetes in Norway: The HUNT3 Survey. J Trace Elem Med Biol 41: 91-98.
http://dx.doi.Org/10.1016/i.itemb.2017.03.001.

Sobolewski. M: Abston. K: Conrad. K: Marvin. E: Harvey. K: Susiario. M: Corv-Slechta. DA. (2020). Lineage- and
sex-dependent behavioral and biochemical transgenerational consequences of developmental exposure to
lead, prenatal stress, and combined lead and prenatal stress in mice. Environ Health Perspect 128: 27001.
http://dx.doi.org/10.1289/EHP4977.

Souza-Talarico. JN: Suchecki. D: Juster. RP: Plusauellec. P: Junior. FB: Bunscheit. V: Marcourakis. T: de Matos.
TM: Lupien. SJ. (2017). Lead exposure is related to hypercortisolemic profiles and allostatic load in
Brazilian older adults. Environ Res 154: 261-268. http://dx.doi.Org/10.1016/i.envres.2017.01.012.

Sun. KH: Mei. WH: Mo. S: Xin. L: Lei. XJ: Huang. MT: Chen. OF: Han. L: Zhu. XF. (2019). Lead exposure

inhibits osteoblastic differentiation and inactivates the canonical Wnt signal and recovery by icaritin in
MC3T3-E1 subclone 14 cells. Chem Biol Interact 303: 7-13. http://dx.doi.Org/10.1016/i.cbi.2019.01.039.

Swarup. D: Naresh. R: Varshnev. VP: Balagangatharathilagar. M: Kumar. P: Nandi. D: Patra. RC. (2007). Changes
in plasma hormones profile and liver function in cows naturally exposed to lead and cadmium around
different industrial areas. Res Vet Sci 82: 16-21. http://dx.doi.Org/10.1016/i.rvsc.2006.05.002.

Tonelli. M: Wiebe. N: Bello. A: Field. CJ: Gill. JS: Hemmelgarn. BR: Holmes. DT: Jindal. K: Klarenbach. SW:

Manns. BJ: Thadhani. R: Kinniburgh. D. (2018). Concentrations of trace elements and clinical outcomes in
hemodialysis patients: A prospective cohort study. Clin J Am Soc Nephrol 13: 907-915.
http://dx.doi.org/10.2215/CJN.11451017.

U.S. EPA (U.S. Environmental Protection Agency). (2006). Air quality criteria for lead [EPA Report]. (EPA/600/R-
05/144aF-bF). Research Triangle Park, NC. https://cfpub.epa.gov/ncea/risk/recordisplav.cfm?deid= 158823.

U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for lead [EPA Report].
(EPA/600/R-10/075F). Washington, DC. https://nepis.epa. gov/Exe/ZvPURL,cgi?Dockev=P 100K82L.txt.

U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the Integrated Science Assessments [EPA
Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
of Research and Development, National Center for Environmental Assessment, RTP Division.
https ://cfpub .epa. gov/ncea/isa/recordisplav. cfm?deid=310244.

Upadhvav. AK: Mathur. R: Bhadauria. M: Nirala. SK. (2009). Therapeutic influence of zinc and ascorbic acid
against lead induced biochemical alterations. Therapie 64: 383-388.
http ://dx. doi.org/10.2515/therapie/2009055.

vanBemmel. DM: Li. Y: McLean. J: Chang. MH: Dowling. NF: Graubard. B: Raiaraman. P. (2011). Blood lead
levels, ALAD gene polymorphisms, and mortality. Epidemiology 22: 273-278.
http://dx.doi.org/10.1097/EDE.0b013e3182Q93f75.

External Review Draft

9-155

DRAFT: Do not cite or quote

-------
Wang. N.T: Lu. M: Chen. C: Xia. FZ: Han. B: Li. O: Cheng. J: Chen. Y: Zhu. C'F: Jensen. MP: Lu. YL. (2018a).

Adiposity genetic risk score modifies the association between blood lead level and body mass index. J Clin
Endocrinol Metab 103: 4005-4013. http://dx.doi.org/10.1210/ic.2018-0Q472.

Wang. W: Moroi. S: Bakulski. K: Mukheriee. B: Weisskopf. MG: Schaumberg. D: Sparrow. D: Vokonas. PS: Hu.
H: Park. SK. (2018b). Bone lead levels and risk of incident primary open-angle glaucoma: The VA
normative aging study. Environ Health Perspect 126: 087002. http://dx.doi.org/10.1289/EHP3442.

Wang. WJ: Wu. CC: Jung. WT: Lin. CY. (2019). The associations among lead exposure, bone mineral density, and
FRAX score: NHANES, 2013 to 2014. Bone 128: 115045. http://dx.doi.Org/10.1016/i.bone.2019.115045.

Wang. WY: Schaumberg. DA: Park. SK. (2016). Cadmium and lead exposure and risk of cataract surgery in U.S.
adults. Int J Hyg Environ Health 219: 850-856. http://dx.doi.Org/10.1016/i.iiheh.2016.07.012.

Wang. X: Mukheriee. B: Park. SK. (2018c). Associations of cumulative exposure to heavy metal mixtures with
obesity and its comorbidities among U.S. adults in NHANES 2003-2014. Environ Int 121: 683-694.
http://dx.doi.Org/10.1016/i.envint.2018.09.035.

Wen. WL: Wang. CW: Wu. DW: Chen. SC: Hung. CH: Kuo. CH. (2020). Associations of heavy metals with
metabolic syndrome and anthropometric indices. Nutrients 12: 2666.
http://dx.doi.org/10.3390/nul2092666.

Werder. EJ: Beier. JI: Sandler. DP: Falkner. KC: Gripshover. T: Wahlang. B: Engel. LS: Cave. MC. (2020). Blood
BTEXS and heavy metal levels are associated with liver injury and systemic inflammation in Gulf states
residents. Food Chem Toxicol 139: 111242. http://dx.doi.org/10.1016/i.fct.2020.111242.

Wiener. RC: Long. PL: Jurevic. RJ. (2015). Blood levels of the heavy metal, lead, and caries in children aged 24-72
months: NHANES III. Caries Res 49: 26-33. http://dx.doi.org/10.1159/00Q365297.

Won. YS: Kim. JH: Kim. YS: Bae. KH. (2013). Association of internal exposure of cadmium and lead with

periodontal disease: A study of the Fourth Korean National Health and Nutrition Examination Survey. J
Clin Periodontal 40: 118-124. http://dx.doi.org/10. Ill 1/icpe. 12033.

Wu. EW: Schaumberg. DA: Park. SK. (2014). Environmental cadmium and lead exposures and age-related macular
degeneration in U.S. adults: The National Health and Nutrition Examination Survey 2005 to 2008. Environ
Res 133: 178-184. http://dx.doi.Org/10.1016/i.envres.2014.05.023.

Wu. Y: Jansen. EC: Peterson. KE: Foxman. B: Goodrich. JM: Hu. H: Solano-Gonzalez. M: Cantoral. A: Tellez-
Roio. MM: Martinez-Mier. EA. (2019). The associations between lead exposure at multiple sensitive life
periods and dental caries risks in permanent teeth. Sci Total Environ 654: 1048-1055.
http://dx.doi.Org/10.1016/i.scitotenv.2018.ll.190.

Xu. C: Shu. Y: Fu. Z: Hu. Y: Mo. X. (2017). Associations between lead concentrations and cardiovascular risk
factors in U.S. adolescents. Sci Rep 7: 9121. http://dx.doi.org/10.1038/s41598-017-Q9701-4.

Xu. HP: Mao. Y: Xu. BC: Hu. YN. (2021). Low-level environmental lead and cadmium exposures and dyslipidemia
in adults: Findings from the NHANES 2005-2016. J Trace Elem Med Biol 63: 126651.
http://dx.doi.Org/10.1016/i.itemb.2020.126651.

Yu. DY: Li. WF: Deng. B: Mao. XF. (2008). Effects of lead on hepatic antioxidant status and transcription of

superoxide dismutase gene in pigs. Biol Trace Elem Res 126: 121-128. http://dx.doi.org/10.10Q7/sl2011-
008-8198-4.

Zeng. X: Xu. XJ: Boezen. HM: Vonk. JM: Wu. WD: Huo. X. (2017). Decreased lung function with mediation of
blood parameters linked to e-waste lead and cadmium exposure in preschool children. Environ Pollut 230:
838-848. http://dx.doi.Org/10.1016/i.envpol.2017.07.014.

Zeng. X: Xu. XJ: Zheng. XB: Reponen. T: Chen. AM: Huo. X. (2016). Heavy metals in PM2.5 and in blood, and
children's respiratory symptoms and asthma from an e-waste recycling area. Environ Pollut 210: 346-353.
http://dx.doi.Org/10.1016/i.envpol.2016.01.025.

Zhai. HL: Chen. C: Wang. NJ: Chen. Y: Nie. XM: Han. B: Li. O: Xia. FZ: Lu. YL. (2017). Blood lead level is

associated with non-alcoholic fatty liver disease in the Yangtze River Delta region of China in the context
of rapid urbanization. Environ Health 16: 93. http://dx.doi.org/10.1186/sl2940-017-03Q4-7.

External Review Draft

9-156

DRAFT: Do not cite or quote

-------
Zhang. Y: Zhou. L: Li. S: Liu. J: Sun. S: Ji. X: Yan. C: Xu. J. (2019). Impacts of lead exposure and chelation

therapy on bone metabolism during different developmental stages of rats. Ecotoxicol Environ Saf 183:
109441. http://dx.doi.Org/10.1016/i.ecoenv.2019.109441.

Zhou. CC: Gao. ZY: Wang. J: Wu. MO: Hu. S: Chen. F: Liu. JX: Pan. H: Yan. CH. (2018). Lead exposure induces
Alzheimers's disease (AD)-like pathology and disturbes cholesterol metabolism in the young rat brain.
Toxicol Lett 296: 173-183. http://dx.doi.Org/10.1016/i.toxlet.2018.06.1065.

External Review Draft

9-157

DRAFT: Do not cite or quote

-------