EPA/635/R-23/337

IRIS Assessment Protocol

www.epa.gov/iris

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)
(Preliminary Assessment Materials)

[CASRN 14797-55-8 and 147-65-0]

[Sodium nitrate: CASRN 7631-99-4]

[Sodium nitrite: CASRN 7632-00-0]

[Potassium nitrate: CASRN 7757-79-1]

[Potassium nitrite: CASRN 7758-09-0]

[Ammonium nitrate: CASRN 6484-54-2]

[Calcium nitrate: CASRN 10124-37-5 [anhydrous]; 13477-34-4

[tetrahydrate]]

November 2023

Integrated Risk Information System
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

DISCLAIMER

This document is a public comment draft for review purposes only. This information is
distributed solely for the purpose of public comment. It has not been formally disseminated by EPA. It
does not represent and should not be construed to represent any Agency determination or policy. Mention
of trade names or commercial products does not constitute endorsement or recommendation for use.

This document is a draft for review purposes only and does not constitute Agency policy.

ii	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

CONTENTS

AUTHORS | CONTRIBUTORS | REVIEWERS	ix

1.	INTRODUCTION	1-1

2.	SCOPING AND PROBLEM FORMULATION	2-1

2.1.	BACKGROUND	2-1

2.1.1.	Physical and Chemical Properties	2-1

2.1.2.	Sources, Production, and Use	2-3

2.1.3.	Environmental Fate and Transport	2-4

2.1.4.	Potential for Human Exposure and Populations with Potentially Greater Exposure	2-4

2.2.	SCOPING SUMMARY	2-5

2.3.	PROBLEM FORMULATION	2-7

2.4.	KEY SCIENCE ISSUES	2-9

3.	OVERALL OBJECTIVES AND SPECIFIC AIMS	3-1

3.1. SPECIFIC AIMS	3-1

4.	LITERATURE SEARCH, SCREENING, AND LITERATURE INVENTORY	4-1

4.1.	POPULATIONS, COMPARATORS, EXPOSURES, OUTCOMES CRITERIA FOR THE

SYSTEMATIC EVIDENCE MAP	4-1

4.2.	SUPPLEMENTAL CONTENT SCREENING CRITERIA	4-2

4.3.	LITERATURE SEARCH STRATEGIES	4-7

4.3.1.	Database Search Term Development	4-7

4.3.2.	Database Searches	4-7

4.3.3.	Searching Other Sources	4-8

4.3.4.	Non-Peer-Reviewed Data	4-9

4.4.	LITERATURE SCREENING	4-10

4.4.1.	Title-and-Abstract Screening	4-10

4.4.2.	Full-Text Screening	4-11

4.4.3.	Multiple Publications of the Same Data	4-11

4.4.4.	Literature Flow Diagrams	4-12

4.5.	LITERATURE INVENTORY	4-12

4.5.1.	Studies That Meet Problem Formulation PECO Criteria	4-12

4.5.2.	Organizational Approach for Supplemental Material	4-12

This document is a draft for review purposes only and does not constitute Agency policy.

iii	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

5.	REFINED PROBLEM FORMULATION AND ASSESSMENT APPROACH	5-1

5.1.	ASSESSMENT PECO CRITERIA	5-1

5.1.1. Other Exclusions Based on Full-Text Content	5-7

5.2.	UNITS OF ANALYSES FOR DEVELOPING EVIDENCE SYNTHESIS AND INTEGRATION
JUDGMENTS FOR HEALTH EFFECT CATEGORIES	5-8

5.3.	CONSIDERATION OF SUPPLEMENTAL MATERIAL	5-12

5.3.1.	Noncancer MOA Mechanistic Information	5-12

5.3.2.	ADME and PK/PBPK Model Information	5-12

5.3.3.	Other Supplemental Material Content	5-12

6.	STUDY EVALUATION (RISK OF BIAS AND SENSITIVITY)	6-1

6.1.	STUDY EVALUATION OVERVIEW FOR HEALTH EFFECT STUDIES	6-1

6.2.	EPIDEMIOLOGY STUDY EVALUATION	6-5

6.3.	EXPERIMENTAL ANIMAL STUDY EVALUATION	6-15

6.4.	CONTROLLED HUMAN EXPOSURE STUDY EVALUATION	6-24

6.5.	IN VITRO AND OTHER MECHANISTIC STUDY EVALUATION	6-24

6.6.	PHARMACOKINETIC MODEL EVALUATION	6-33

6.6.1.	Pharmacokinetic (PK)/Physiologically Based Pharmacokinetic (PBPK) Model
Descriptive Summary	6-33

6.6.2.	Pharmacokinetic (PK)/Physiologically Based Pharmacokinetic (PBPK) Model
Evaluation	6-34

6.6.3.	Selection of the Appropriate Dose Metric	6-37

7.	DATA EXTRACTION OF STUDY METHODS AND RESULTS	7-1

8.	EVIDENCE SYNTHESIS AND INTEGRATION	8-1

8.1.	EVIDENCE SYNTHESIS	8-5

8.2.	EVIDENCE INTEGRATION	8-15

9.	DOSE-RESPONSE ASSESSMENT: SELECTING STUDIES AND QUANTITATIVE ANALYSIS	9-1

9.1.	OVERVIEW	9-1

9.2.	SELECTING STUDIES FOR DOSE-RESPONSE ASSESSMENT	9-2

9.2.1. Hazard and MOA Considerations for Dose Response	9-2

9.3.	CONDUCTING DOSE-RESPONSE ASSESSMENTS	9-6

9.3.1.	Dose-Response Analysis in the Range of Observation	9-6

9.3.2.	Extrapolation: Slope Factors and Unit Risk	9-9

9.3.3.	Extrapolation: Reference Values	9-9

APPENDIX A. SYSTEMATIC EVIDENCE MAP FOR HEALTH EFFECTS OF NITRATES AND NITRITES	A-l

This document is a draft for review purposes only and does not constitute Agency policy.

iv	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

A.l. INTRODUCTION	A-l

A.2.METHODS	A-l

A.2.1. Specific Aims	A-l

A.2.2. Literature Search and Screening Strategies	A-2

A.2.3. Literature Inventory	A-3

A.3.RESULTS	A-3

A.3.1. Available Health Values	A-3

A.3.2. Literature Screening Results	A-8

A.3.3. Characterizing Animal and Epidemiological Studies	A-9

A.4. CONCLUSIONS	A-12

APPENDIX B. SURVEY OF EXISTING TOXICITY VALUES	B-l

APPENDIX C. LITERATURE SEARCH STRATEGIES	C-l

APPENDIX D. PROCESS FOR SEARCHING AND COLLECTING EVIDENCE FROM SELECTED OTHER

RESOURCES	D-l

D.l.REVIEW OF REFERENCE LISTS FROM EXISTING ASSESSMENTS (FINAL OR PUBLICLY

AVAILABLE DRAFT), JOURNAL REVIEW ARTICLES, AND STUDIES CONSIDERED RELEVANT
TO PECO BASED ON FULL-TEXT SCREENING	D-l

D.2.EUROPEAN CHEMICALS AGENCY	D-l

D.3.EPA CHEMVIEW	D-l

D.4.NTP CHEMICAL EFFECTS IN BIOLOGICAL SYSTEMS	D-2

D.5.ECOTOX DATABASE	D-2

D.6.EPA COMPTOX CHEMICAL DASHBOARD VERSION TO RETRIEVE A SUMMARY OF ANY

TOXCAST ORTOX21 HIGH THROUGHPUT SCREENING INFORMATION	D-2

D.7.COMPARATIVE TOXICOGENOMICS DATABASE	D-3

REFERENCES	R-l

This document is a draft for review purposes only and does not constitute Agency policy.

v	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

TABLES

Table 2-1. Physicochemical properties of nitrate and nitrite	2-2

Table 2-2. Physicochemical properties of selected nitrate and nitrite compounds	2-2

Table 2-3. EPA program and regional office interest in a reassessment of nitrate and nitrite	2-5

Table 2-4. Nitrate/nitrite compounds considered for assessment	2-6

Table 4-1. Problem formulation populations, exposures, comparators, outcomes (PECO) criteria

for the nitrate and nitrite assessment	4-2

Table 4-2. Categories of potentially relevant supplemental material	4-4

Table 5-1. Assessment PECO criteria for the nitrate/nitrite (oral) assessment	5-6

Table 5-2. Health effect categories and human and animal evidence unit of analysis endpoint

groupings for which evidence integration judgments will be developed	5-9

Table 6-1. Domains, questions, and general considerations to guide the evaluation of

epidemiology studies	6-6

Table 6-2. Domains, questions, and general considerations to guide the evaluation of animal

toxicology studies	6-16

Table 6-3. Domains, questions, and general considerations to guide the evaluation of in vitro

studies	6-25

Table 6-4. Example descriptive summary for a physiologically based pharmacokinetic (PBPK)

model	6-34

Table 6-5. Criteria for evaluation of physiologically based pharmacokinetic (PBPK) models	6-36

Table 8-1. Generalized evidence profile table to show the relationship between evidence
synthesis and evidence integration to reach judgment of the evidence for

hazard	8-3

Table 8-2. Generalized evidence profile table to show the key findings and supporting rationale

from mechanistic analyses	8-4

Table 8-3. Considerations that inform evaluations and judgments of the strength of the evidence

for hazard	8-7

Table 8-4. Framework for strength of evidence judgments from studies in humans	8-12

Table 8-5. Framework for strength of evidence judgments from studies in animals	8-13

Table 8-6. Considerations that inform evidence integration judgments	8-15

Table 8-7. Framework for summary evidence integration judgments in the evidence integration

narrative	8-18

Table 9-1. Attributes used to evaluate studies for derivation of toxicity values	9-4

Table A-l. Details on derivation of the available health effect reference values for oral exposure

to nitrate and nitrite	A-5

Table B-l. Sources searched for existing human health reference values	B-l

Table C-l. Results of initial literature search	C-l

Table D-l. Summary table for other sources search results	D-4

This document is a draft for review purposes only and does not constitute Agency policy.

vi	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

FIGURES

Figure 1-1. Integrated Risk Information System (IRIS) systematic review problem formulation and

method documents	1-2

Figure 2-1. Available health effect reference values for oral exposure to nitrate and nitrite	2-8

Figure 4-1. Visual summary of approach for tagging major categories of supplemental material	4-13

Figure 4-2. Visual summary of overall tagging structure for mechanistic studies	4-14

Figure 4-3. Visual summary of tagging structure for ADME and PK/PBPK studies	4-15

Figure 6-1. Overview of Integrated Risk Information System (IRIS) study evaluation approach	6-2

Figure A-l. Nitrate/ nitrite literature flow diagram	A-9

Figure A-2. Survey of human studies that met PECO criteria summarized by study design,

population, and health systems assessed	A-10

Figure A-3. Survey of animal studies that met PECO criteria by study design, species, and health

systems	A-ll

This document is a draft for review purposes only and does not constitute Agency policy.

vii	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

ABBREVIATIONS

ADME

absorption, distribution, metabolism, or elimination

BMD

benchmark dose

BMDL

benchmark dose lower confidence limit

BW3/4

body-weight scaling to the 3/4 power

BMDS

Benchmark Dose Software

CAS

Chemical Abstracts Service

CASRN

Chemical Abstracts Service Registry Number

CERCLA

Comprehensive Environmental Response, Compensation, and Liability Act

CI

confidence interval

COI

conflict of interest

CPHEA

Center for Public Health and Environmental Assessment

EPA

Environmental Protection Agency

GLP

good laboratory practices

GRADE

Grading of Recommendations Assessment, Development and Evaluation

HAWC

Health Assessment Workspace Collaborative

HEC

human equivalent concentration

HERO

Health and Environmental Research Online

IAP

IRIS Assessment Plan

IPCS

International Programme on Chemical Safety

IRIS

Integrated Risk Information System

ITER

International Toxicity Estimates for Risk

IUR

inhalation unit risk

LOAEL

lowest-observed-adverse-effect level

LOEL

lowest-observed-effect level

MeSH

Medical Subject Headings

MOA

mode of action

NMD

normalized mean difference

NOEL

no-observed-effect level

NTP

National Toxicology Program

NOAEL

no-observed-adverse-effect level

OCHP

Office of Children's Health Protection

OLEM

Office of Land and Emergency Management

ORD

Office of Research and Development

ORAU

Oak Ridge Associated Universities

OSF

oral slope factor

OW

Office of Water

PBPK

physiologically based pharmacokinetic

PECO

populations, exposures, comparators, outcomes

PK

pharmacokinetic

POD

point of departure

RfC

reference concentration

RfD

reference dose

ROBINS-I

Risk of Bias in Non-Randomized Studies of Interventions

SD

standard deviation

SE

standard error

SEM

systematic evidence map

UF

uncertainty factor

This document is a draft for review purposes only and does not constitute Agency policy.

viii	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

AUTHORS | CONTRIBUTORS | REVIEWERS

Assessment Team

Krista Christensen. Ph.D. (co-Assessment Manager) EPA/ORD/CPHEA/CPAD

Alexandra Lee. Ph.D. (co-Assessment Manager)

Roman Mezencev. Ph.D.

Amanda Persad. Ph.D.

Martha Powers. Ph.D.

Rachel Shaffer. Ph.D.

Todd Zurlinden. Ph.D.

Contributors

Brittany Schulz
Nicholas Barnett
Jennifer Lees

Executive Direction

Wayne Cascio, M.D. (CPHEA Director)	EPA/ORD/CPHEA

Kay Holt, M.S. (CPHEA Deputy Director)

Samantha Jones, Ph.D. (CPHEA Associate Director)

Kristina Thayer, Ph.D. (CPAD Director)

Steve Dutton, Ph.D. (HEEAD Director)

Andrew Kraft, Ph.D. (CPAD Associate Director)

Ravi Subramaniam, Ph.D (CPAD Senior Science
Advisor)

Paul White, (CPAD Senior Science Advisor)

Elizabeth Radke, Ph.D. (Branch Chief)

Andrew Hotchkiss, Ph.D. (Branch Chief)

Janice Lee, Ph.D. (Branch Chief)

Kathleen Newhouse, M.S. (Acting Branch Chief)

Glenn Rice, Ph.D. (Branch Chief)

Viktor Morozov, Ph.D. (Branch Chief)

Vicki Soto, B.S. (Branch Chief)

Production Team

Maureen Johnson (CPHEA Webmaster)	EPA/ORD/CPHEA

Ryan Jones (HERO Director)

Dahnish Shams (Production Team)

Jessica Soto-Hernandez (Production Team)

Samuel Thacker (HERO Team)

Garland Waleko (Production Team)

Oak Ridge Associated Universities (ORAU)
Contractor

This document is a draft for review purposes only and does not constitute Agency policy.

ix	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1. INTRODUCTION

The Integrated Risk Information System (IRIS) Program is undertaking a reassessment of
the health effects of nitrate and nitrite via the oral (ingestion) route of exposure. IRIS assessments
provide high quality, publicly available hazard identification and dose-response analyses on
chemicals to which the public might be exposed. These assessments are not regulations but provide
an important source of toxicity information used by the Environmental Protection Agency (EPA),
state and local health agencies, tribes, other federal agencies, and international health
organizations.

An IRIS Assessment Plan (IAP) was presented at a public science meeting on September
27-28, 2017 rhttps://sab.epa.gov/ords/sab/f?p=100:19:3574465722633:::19:P19 ID:9041 to seek
input on the problem formulation components of the assessment plan. The 2017 IAP specified the
EPA need for an assessment of nitrate/nitrite, described the objectives and specific aims of the
assessment, provided draft PECO (populations, exposures, comparators, and outcomes) criteria,
and described areas of scientific complexity. However, in April 2019 the nitrate/nitrite assessment
was suspended due to changes in EPA leadership priorities for the IRIS Program fApril 2019 IRIS
Program Outlook], During the last nomination cycle, EPA's Office of Water (OW), Office of Children's
Health Protection (OCHP), and Region 5 prioritized nitrate and nitrite for assessment by the IRIS
Program. In June 2023, the assessment was added to the IRIS Program Outlook to address
assessment needs of EPA's Offices and Regions. This assessment may also be used to support
actions in other EPA program and regional offices and can inform efforts to address nitrate/nitrite
by tribes, states, and international health agencies (see Section 2.2).

The Protocol document includes the IAP content, revised in response to public input and
updated EPA scoping needs and presents the methods for conducting the systematic review and
dose-response analysis for the assessment. While the IAP described what the assessment will cover,
this Protocol describes how the assessment will be conducted (see Figure 1-1).

The systematic review methods described in this Protocol are based on the Office of
Research and Development (ORD) Staff Standard Operating Procedures for Developing Integrated
Risk Information System (IRIS) Assessments (Version 2.0, referred to as the "IRIS Handbook") fU.S.
EPA. 20221. The methods presented in this Protocol reflect the information provided in the IRIS
Handbook which incorporates adjustments made based on a November 2021 National Academy of
Sciences, Engineering, and Medicine (NASEM) committee review of that version of the IRIS
Handbook fNASEM. 2021: U.S. EPA. 2020al.

This document is a draft for review purposes only and does not constitute Agency policy.

1-1	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Scoping

Literahire
Search

i	Jieqrch

* • 4

Derive Toxicity
Values

ft

Problem
Formulation

Literature
Inventory

Assessment Plans:
What the assessment

will cover

Refine	Dala	Evidence	Derive Toxi

Analysis Plan	Extraction	Integration	Values

Study Evaluation Evidence	Select and

Synthesis	Model Studies

Protocols: How the assessment will be
conducted.

Assessment
Developed

Figure 1-1. Integrated Risk Information System (IRIS) systematic review
problem formulation and method documents.

This document is a draft for review purposes only and does not constitute Agency policy.

1-2	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

2.SCOPING AND PROBLEM FORMULATION

2.1. BACKGROUND

Below is a brief overview of aspects of the physiochemical properties, human exposure, and
environmental fate characteristics of nitrate and nitrite (Chemical Abstract Services Registry
Number [CASRN] 14797-55-8 and 147-65-0). This overview provides a summary of background
information for contextual purposes only and is not intended to be comprehensive descriptions of
the available information.

2.1.1. Physical and Chemical Properties

Inorganic nitrate (NO3-) and nitrite (NO2-) are naturally occurring anions formed by
fixation of nitrogen and oxygen. Nitrate is a more stable form compared to nitrite although
conversion between the two forms can readily occur through biological and chemical processes.
Nitrite can also be converted to a class of compounds called N-nitrosamines. There are many
organic and inorganic nitrate and nitrite compounds; for the purposes of this assessment the focus
is on the following forms: potassium nitrate, potassium nitrite, sodium nitrate, sodium nitrite,
ammonium nitrate, and calcium nitrate. Calcium nitrate was not included in the 2017 IAP but has
been added to this Protocol based upon recommendation from EPA's OW. This group of inorganic
compounds are highly water-soluble and readily dissociate. Selected chemical and physical
properties of nitrate and nitrite are listed in Table 2-1 below, while properties of the nitrate and
nitrite compounds of interest are listed in Table 2-2.

This document is a draft for review purposes only and does not constitute Agency policy.

2-1	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 2-1. Physicochemical properties of nitrate and nitrite

Characteristic or
property (unit)

Nitrate

Nitrite

Chemical structure

0

N+

0 cr

O — z
\\
o

CASRN

14797-55-8

14797-65-0

EPA Chemicals Dashboard
DTXSID

DTXSID5024217

DTXSID5024219

Synonyms

Nitrate; Nitric acid, ion(l-)

Nitrite; Nitrite ion; Nitrous acid, ion(l-)

Color/form

Varies by specific compound

Varies by specific compound

Molecular formula

NObh

N02()

Molecular weight (g/mol)

62.005

46.006

Log Kow

4.05 x 10"2

-5 x 10"3

aU.S. EPA (2021) Chemicals Dashboard: https://comptox.epa.gov/dashboard/chemical/details/DTXSID5024217 and
https://comptox.epa.gov/dashboard/chemical/details/DTXSID5024219 (accessed date October 14, 2022).
Synonyms are those categorized as "valid" or "good" in the CompTox Chemicals Dashboard excluding foreign
language synonyms and United Nation (UN) numbers. Median or average experimental values are used when
available; otherwise, median, or average predicted values are used.

Table 2-2. Physicochemical properties of selected nitrate and nitrite
compounds

Characteristic
or property
(unit)

Calcium
nitrate

Ammonium
nitrate

Sodium
nitrate

Sodium
nitrite

Potassium
nitrate

Potassium
nitrite

Chemical
structure



0
//

0-—N NH.,'
\

0"

0

//

Na cr—n'

\
cr

0 0

Na+

O

!

¦c ox

CASRN

13477-34-4
(anhydrous)
10124-37-5
(tetrahydrate)

6484-52-2

7631-99-4

7632-00-0

7757-79-1

7758-09-0

EPA Chemicals

Dashboard

DTXSID

DTXSID 1039719

DTXSID2029688

DTXSID6020937

DTXSID0020941

DTXSID4029692

DTXSID5042320

Synonyms

Nitric acid,
calcium salt

Nitric acid,
ammonium salt

Nitric acid,
sodium salt

Nitrous acid,
sodium salt

Nitric acid,
potassium salt

Nitrous acid,
potassium salt

This document is a draft for review purposes only and does not constitute Agency policy.

2-2	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Characteristic
or property
(unit)

Calcium
nitrate

Ammonium
nitrate

Sodium
nitrate

Sodium
nitrite

Potassium
nitrate

Potassium
nitrite



Alternative
names: Calcium
Dinitrate; Lime
nitrate; Norge
saltpeter;
Norwegian
saltpeter;
Calcium
saltpeter

Alternative

Names:

Ammonium

nitrate;

Emulite; EXP

200; German

saltpeter;

Norway

saltpeter;

Norge

saltpeter;

Norwegian

saltpeter;

Plenco 12203;

Varioform 1;

ZhVK

Alternative
Names: Chile
saltpeter; Niter;
Nitric acid
sodium salt;
Saltpeter; Soda
niter; Nitrate of
soda; Cubic
niter; Nitratine

Alternative
Names: Nitrous
acid soda;
Nitrous acid
sodium salt

Alternative
Names: Niter;
Nitre; Nitric
acid potassium
salt; Saltpeter;
Saltpetre;
Nitrate of
potash

Alternative
Names: Chile
saltpeter; Niter;
Nitric acid
sodium salt;
Salpeter; Soda
niter

Color/form

White to light
gray; Solid

White,

colorless, gray,
or brown; Solid

White or
colorless; Solid

White to pale
yellow; Solid

Colorless; Solid

Pale yellow;
Solid

Molecular
formula

Ca(N03)2

NH4NO3

NaNOs

NaNCh

KNOb

KNO2

Molecular
weight (g/mol)

164.09

80.04

84.99

68.99

101.10

85.10

Boiling point
(°C)

142

210

380

320

400

537

Melting point
(°C)

43

(tetrahydrate)
561 (anhydrous)

169.7

306

271

334

440

aU.S. EPA (2021) Chemicals Dashboard:
https://comptox.epa.gov/dashboard/chemical/details/DTXSID6020937 (sodium nitrate);
https://comptox.epa.gov/dashboard/chemical/hazard/DTXSID0020941 (sodium nitrite);
https://comptox.epa.gov/dashboard/chemical/details/DTXSID4029692 (potassium nitrate);
https://comptox.epa.gov/dashboard/chemical/hazard/DTXSID5042320 (potassium nitrite);
https://comptox.epa.gov/dashboard/chemical/hazard/DTXSID2029668 (ammonium nitrate);
https://comptox.epa.gov/dashboard/chemical/hazard/DTXSID1039719 (calcium nitrate) (accessed date October
14, 2022).

Synonyms are those categorized as "valid" or "good" in the CompTox Chemicals Dashboard excluding foreign
language synonyms and United Nation (UN) numbers. Median or average experimental values are used when
available; otherwise, median, or average predicted values are used.

2.1.2. Sources, Production, and Use

1	Nitrate and nitrite play an essential role in Earth's nitrogen cycle. Since 1950, human

2	sources of reactive nitrogen into the environment—released either intentionally (e.g., through

3	fertilizer application) or unintentionally (e.g., as a byproduct of fossil fuel combustion)—have

This document is a draft for review purposes only and does not constitute Agency policy.

2-3	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

increased substantially (Fields. 20041. Nitrate salts are mainly used as nitrogen fertilizers and in
industrial explosives, fireworks, and glass making; nitrites are largely used as preservatives for
meat and fish curing and as color fixatives flARC. 2010: Pokornv L. 20061. Nonpoint and point
sources of nitrate/nitrite include animal waste, urban and agricultural runoff, landfill leachate,
storm sewer overflow, vehicle exhaust, septic-system effluent, industrial processes, and industrial
or mining wastewater fATSDR. 2017: Bryan and Loscalzo. 2011: IARC. 2010: Pokornv L. 20061.

2.1.3.	Environmental Fate and Transport

Nitrates account for most of the available total nitrogen in both ground and surface waters;
nitrite levels are generally low in both fDesimone. 20091. According to monitoring data obtained
during EPA's third Six-Year Review of National Primary Drinking Water Regulations fU.S. EPA.
20161. nitrate and nitrite were detected in approximately 63.8% and 11.7% of drinking water
systems, respectively. The 5th to 95th percentile ranges of detected concentrations for nitrate and
nitrite were 84-8,339 ng/L, and 2-1,150 ng/L, respectively (See exhibit 6-1 in the Occurrence
Support Document, (U.S. EPA. 201611. Human activities are responsible for increased levels of
nitrate in drinking water sources; (Desimone. 20091 reported that nitrate concentrations greater
than 1 mg/L (as N) are levels "considered to result from the effects of human activities in many
parts of the United States" and that this level was exceeded in 41.4% of wells surveyed. Populations
served by private well water, especially shallow wells in agricultural areas, may be exposed to
nitrate at levels several times higher than those served by public water systems (Desimone. 2009:
Ward. 20091.

2.1.4.	Potential for Human Exposure and Populations with Potentially Greater Exposure

The general population is exposed to nitrate in both drinking water and food. Vegetables
are the main source of ingested nitrate, with leafy vegetables comprising nearly 80% of nitrate
exposure in an average person's diet Other sources of dietary nitrate include cured meats/fish,
cereal grains, dairy products, and beer (ATSDR. 2017: IARC. 20101. In contrast to nitrates,
endogenous sources account for approximately 80% of all nitrites in the human body, as 5%-8% of
the total nitrate intake is converted into nitrite (WHO. 2016: Mensinga etal.. 20031. Almost all
exogenous exposure to nitrite comes from food, with relatively higher nitrite concentrations found
in cured meats flARC. 20101. Drinking water is generally a minor source of exposure to nitrite
HARC. 20101.

Populations with potentially greater than average exposures include those living in
agricultural areas, users of private well water systems, and those with diets high in concentrations
of nitrate/nitrite. Agricultural areas have some of the highest concentrations of nitrates/nitrites in
soil, surface, and groundwater in the United States. Populations using private wells tend to be those
living in and around these more rural, agricultural areas, where nitrate levels in well water are
several times higher than those found in public water systems fWard. 20091. According to the U.S.
Geological Survey (USGS), in 2015 approximately 42.5 million people, or 13% of the U.S. population,

This document is a draft for review purposes only and does not constitute Agency policy.

2-4	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

depended on private wells as their main source of drinking water (Hutson. 20041. According to a
study of sampled private wells across the United States conducted by the U.S. Geological Survey
(USGS) from 1991 to 2004, approximately 4% of all private wells and 25% of private wells in
agricultural areas contained levels above the maximum contaminant level (MCL) for nitrates
fDesimone. 20091.

2.2. SCOPING SUMMARY

During scoping, the IRIS program meets with EPA program and regional offices that have
interest in an IRIS assessment for nitrate and nitrite to discuss specific assessment needs. Table 2-3
provides a summary of input from this outreach.

Table 2-3. EPA program and regional office interest in a reassessment of
nitrate and nitrite

EPA program or
regional office

Oral

Inhalation

Statutes/regulations

Anticipated uses/interest

Office of Water

~



Safe Drinking Water
Act (SDWA) -
Section 1412

Six-year review of
the National Primary
Drinking Water
regulations.

Region 5a

~





Evaluation of special

provision of the

NPDW regulation [40

CFR 141.11(d)]

allowing, at the

discretion of the

state, noncommunity water

systems to exceed the nitrate MCL

Office of
Children's Health
Protection

~



Executive Order 13045—Protection of Children from Environmental
Health Risks and Safety Risks: Policy on Evaluating Health Risks to
Children.

aRegion 5 serves Illinois, Indiana, Michigan, Minnesota, Ohio, Wisconsin, and 35 tribes.

The EPA OW regulates nitrates and nitrites under the National Primary Drinking Water
Regulations (40 CFR141,142); the current MCLs for nitrate and nitrite, promulgated in 1991, are
10 mg/L and 1 mg/L (as nitrogen), respectively (40 CFR 141.62; 56 FR 3594, January 30,1991). An
updated health assessment of nitrate and nitrite is being considered in the ongoing Six-Year Review
cycle for National Primary Drinking Water Regulations. A provision of the current regulation [40
CFR 141.11(d)] allows, at the discretion of the state, noncommunity water systems to exceed the
nitrate MCL up to 20 mg/L if the supplier can demonstrate that the water will not be available to
children under 6 months of age and that no adverse health effects will result The availability of
more recent health effects literature published since 1991 raises questions about whether the

This document is a draft for review purposes only and does not constitute Agency policy.

2-5	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

current MCLs for nitrate and nitrite and the provision allowing exceedance of the nitrate MCL up to
20 mg/L provide adequate health protection for the general population (all life stages).

As described above, this assessment will address inorganic forms of nitrate and nitrite and
will specifically consider health effect information for the compounds included in Table 2-4. These
salts are highly soluble in water and dissociate under environmental conditions; in solution, they
exist as ions fATSDR. 20171. Because the cations are not expected to introduce significant
differences in the toxicity of the different salts, toxicity findings from all five compounds are
considered relevant to an assessment of nitrate and nitrite toxicity ((EFSA). 2017b). These six
compounds listed in Table 2-4 are the most common nitrate and nitrite salts in the environment
fATSDR. 20171. These compounds (except for calcium nitrate) were also the subject of two recent
health assessments of nitrate and nitrite fATSDR. 2017: IARC. 201011. The decision to develop the
assessment of nitrate/nitrite using health effect information for these six compounds was also
based on known general population exposure to these six compounds and availability of
epidemiological or toxicological information. Specifically, ammonium nitrate is a leading nitrogen
fertilizer, and for this reason, has been used in toxicological studies as a component of "California
mixture" and "Iowa mixture." These two mixtures are representative of groundwater
contamination by fertilizers and pesticides and used for simulations of environmental exposures to
pesticides mixtures. Calcium nitrate is similarly used as a fertilizer fSellars and Nunes. 20211.
Sodium nitrate, sodium nitrite, potassium nitrate, and potassium nitrite are used as food additives
to cure meats. The National Toxicology Program (NTP) has assessed the toxicities of n-nitroso
compounds (NTP. 20211 nitrate and sodium nitrite (NTP. 2001b) in animal toxicology and
carcinogenicity studies.

Table 2-4. Nitrate/nitrite compounds considered for assessment

Compound

Chemical formula

CAS Registry Number

Ammonium nitrate

NH4NO3

6484-52-2

Calcium nitrate

Ca(NOs)2

10124-37-5 (anhydrous);
13477-34-4 (tetrahydrate)

Sodium nitrate

NaNOs

7631-99-4

Sodium nitrite

NaN02

7632-00-0

Potassium nitrate

KNOb

7757-79-1

Potassium nitrite

KNO2

7758-09-0

Assessment of the health effects of nitrate and nitrite following inhalation and dermal
routes of exposure will not be included in the scope of this assessment Inhalation and dermal
exposures to nitrate or nitrite in the general population (i.e., populations not exposed
occupationally, such as factory and fertilizer workers) are expected to be negligible compared to

This document is a draft for review purposes only and does not constitute Agency policy.

2-6	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

36

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

oral exposure fATSDR. 20171. Focusing on the health effects associated with oral exposure to
nitrate and nitrite is consistent with the needs of EPA programs and regional offices.

Given input received during scoping, the IRIS assessment will include evaluation of
noncancer and cancer human health hazards associated with ingested nitrate and nitrite. Although
all health effects will be considered for hazard identification, the assessment will take a different
approach for hematological outcomes. A hematological hazard has already been established
through the known association between methemoglobinemia and nitrate/nitrite (Ward etal.. 2005:
Walton. 19511. Therefore, EPA will not re-consider the hematological domain during hazard
identification. Instead, any new studies identified for methemoglobinemia and supporting
hematological endpoints will be examined for information on the quantitative relationship with
nitrate/nitrite and the potential to support dose-response analysis. For cancer, EPA will develop a
qualitative assessment of the carcinogenic potential of nitrate and nitrite and will explore the
feasibility of developing a quantitative assessment (for details, see Sections 8 and 9). EPA
anticipates that a quantitative cancer assessment will be particularly challenging, given the co-
occurrence of nitrosatable compounds and antioxidants in dietary sources, conflicting results
across studies, and design limitations of epidemiological studies investigating the association
between cancer and nitrate/nitrite exposure at different sites.

2.3. PROBLEM FORMULATION

The IRIS program currently does not include cancer risk values for nitrate or nitrite. The
International Agency for Research on Cancer (IARC) has determined that there is "inadequate"
evidence of carcinogenicity of nitrate in food or drinking water, "limited" evidence for the
carcinogenicity of nitrite in food, and "sufficient" evidence for the carcinogenicity of nitrite in
combination with amines or amides. IARC concludes that "ingested nitrate and nitrite under
conditions that result in endogenous nitrosation is probably carcinogenic to humans (Group 2A)"
flARC. 20101.

The IRIS program lists reference dose (RfD) values of 1.6 mg/kg-day for nitrate and
0.1 mg/kd-g-day for nitrite, based on a critical effect of methemoglobinemia. Agency for Toxic
Substances and Disease Registry (ATSDR) has determined minimal risk levels of 4 mg/kg-day for
nitrate and 0.1 mg/kg-day for nitrite (applicable for acute, intermediate, and chronic durations of
oral exposure) based upon the same health endpoint fATSDR. 20171. The Joint FAO/WHO Expert
Committee on Food Additives (JECFA) has also determined acceptable daily intake values of
3.7 mg/kg-day for nitrate and 0.07 mg/kg-day for nitrite (based on heart and lung effects in rats)
CWHO. 2003: TECFA. 19951.

EPA's MCLs for nitrate and nitrite are 10 mg/L (or ppm) and 1 mg/L (or ppm), respectively.
These are equivalent to ~44 mg nitrate/L as nitrate-nitrogen and ~3.3 mg nitrite/L as nitrite-
nitrogen. California's Office of Environmental Health Hazard Assessment lists public health goals
(PHGs) of 45 mg/L and 3 mg/L for nitrate and nitrite, respectively (the joint nitrate/nitrite PHG is

This document is a draft for review purposes only and does not constitute Agency policy.

2-7	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

10 mg/L) fCalEPA. 20181. The FDA uses these same values for allowable levels in bottled water
(FDA. 20211. and these are also the same values that Health Canada has determined for maximum
allowable concentration values fWater and Air Quality Bureau. 20131.

Federal agencies (OSHA, NIOSH, ATSDR, EPA) have not set legal or recommended limits for
nitrate or nitrite in air, largely due to lack of adequate data. A May 2023 summary of existing
human health reference values for oral exposure to nitrates/nitrites is provided in Figure 2-1. See
Appendix A (Table A-l) for a tabular summary, including derivation details of the displayed values.

Nitrate & Nitrite Oral Reference Values

10

>
as
T3

i

CuO

no
£

(D

oc

0)

o

o 0.1

0.01

Acute

aS-

Short Term

Subchronic

Chronic

EPA/OW RfD (N03)
EPA/IRiS RfD (N03) ndl

SCF ADI (NO,)

ATSDR-MRL(NOs)

ij£r

JECFAADI(N03)



EPA/OW RfD (N02;

0 EPA/HEASTp-RfD(NQ2) ^ ___



ATSDRMRL(N02)

EPA/IRIS RfD (N02]

efsaadi(N02)

JECFAADI (N02) XJ3

SCF ADI(N02)



¦M-

¦M-

May 2023

XATSDR-MRL

~ EPA/IRIS RfD

~ EPA/OW RfD

~ EPA/HEAST p-RfD

XJECFA ADI

~ SCF ADI

~ EFSA ADI

10	100	1,000

Duration (Days)

10,000

100,000

Figure 2-1. Available health effect reference values for oral exposure to nitrate
and nitrite.

To identify noncancer and cancer health outcomes for which possible association with
exposure to nitrate/nitrite has been investigated, a preliminary literature survey was performed
using health assessments produced by other federal, state, and international health agencies
fCalEPA. 2018: ATSDR. 2017: WHO. 2016: Water and Air Quality Bureau. 2013: IARC. 2010: IPCS.
20051. In particular, EPA relied on the ATSDR Toxicological Profile for Nitrate and Nitrite f AT SDR.
20171. as the most recent authoritative health agency assessment, to identify the pertinent health
effect literature through 2016. ATSDR fATSDR. 20171 updated the comprehensive review of the

This document is a draft for review purposes only and does not constitute Agency policy.

2-8	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

cancer epidemiological literature provided in IARC (IARC. 20101 (i.e., literature published up to
approximately 2007), and the IARC monograph also was used to identify the cancer literature. To
identify studies published since the end of the period covered by the ATSDR Toxicological Profile
(i.e., from 2016 to 2022), a literature search update was performed by EPA. The search strategy and
literature screening are described further in Sections 3, 4 and 5 and Appendices B through D. The
details of the preliminary literature survey, also referred to as a systematic evidence map (SEM),
are described in Appendix A.

The SEM revealed many randomized, controlled trial human studies reporting potential
association between controlled nitrate/nitrite exposure and beneficial cardiovascular outcomes.
Because IRIS assessments focus on the adverse effects associated with exposure to environmental
chemicals, a systematic review of the potential beneficial outcomes to the cardiovascular system
associated with the intake of nitrate or nitrite will not be included in this assessment but will be
identified as potentially relevant supplementary material.

2.4. KEY SCIENCE ISSUES

The SEM identified the following key scientific issues and potential mode-of-action
hypotheses as warranting evaluation in this assessment

•	Nitrate and nitrite are generated endogenously as part of the nitrate-nitrite-nitric oxide
cycle that controls the availability of nitric oxide, which is a ubiquitous signaling molecule
involved in the regulation of numerous physiological and pathological processes, including
vasodilation, platelet activation, metabolic regulation, neurotransmission, and host defense
(inflammation). The roles of endogenous versus exogenous nitrate and nitrite in toxicity,
particularly methemoglobinemia in infants, have been debated in the scientific literature.

•	Several susceptible populations and life stages have been identified for
methemoglobinemia. These include infants under 6 months of age; individuals with higher-
than-normal gastric pH; individuals with glucose-6-phosphate dehydrogenase or NADH
(nicotinamide adenine dinucleotide (NAD) + hydrogen)-dependentmethemoglobin
reductase deficiency; individuals with diseases such as anemia, cardiovascular disease, lung
disease, and sepsis; individuals with abnormal hemoglobin species including
carboxyhemoglobin, sulfhemoglobin, and sickle hemoglobin.

•	A physiologically based pharmacokinetic (PBPK) model structure for simulating the kinetics
of methemoglobinemia formation after oral exposure to nitrate in adults is available
fZeilmaker etal.. 2010: Zeilmaker et al.. 19961. An updated parameterization of this model
using recent human data fLin etal.. 20201 needs to be evaluated against the original model
fit (Zeilmaker et al.. 20101 for its potential to inform human variability in the dose-response
assessment

•	Previously published assessments by Health Canada (Water and Air Quality Bureau. 20131.
ATSDR f ATSDR. 20171. IARC CIARC. 20101. the California EPA fCalEPA. 20181 and the WHO
fWHO. 20161 and newer animal and epidemiological studies published after 2014 raise the
following issues related to cancer risk:

This document is a draft for review purposes only and does not constitute Agency policy.

2-9	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

•	Risk associated with intake of nitrates, nitrites, or both from cured meats, vegetables, and
drinking water could differ because of co-occurrence with antioxidants (e.g., vitamin C,
vitamin E) in vegetables, amines in fish and meats, and calcium in drinking water.
Consequently, risks associated with dietary intake, intake through drinking water, and total
intake may need to be assessed separately.

•	There may be susceptible populations with increased cancer risk associated with intake of
nitrate/nitrite due to increased exposure or intrinsic factors.

•	Populations vary significantly in the ability to reduce salivary nitrate by oral bacteria (e.g.,
actinomyces and veilonella) (Bryan and Petrosino. 20171. For example, patients with
migraines were shown to have higher abundance of nitrate, nitrite, and nitric oxide
reductase genes in their oral bacterial metagenome fGonzalez etal.. 20161. In contrast, the
use of antiseptic mouthwashes appears to deplete nitrate-reducing oral bacteria and affect
some nitrite-mediated biological processes (Kapil etal.. 20131. Individuals from some
subgroups may be able to convert more nitrate to nitrite and consequently produce more
carcinogenic n-nitroso derivatives.

This document is a draft for review purposes only and does not constitute Agency policy.

2-10	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

3.0VERALL OBJECTIVES AND SPECIFIC AIMS

The overall objective of this assessment is to identify adverse health effects of nitrate and
nitrite ingestion exposure and characterize exposure-response relationships for these effects to
support development of toxicity values. This assessment will use systematic review methods to
evaluate the epidemiological and toxicological literature, including consideration of relevant
mechanistic evidence, for the specified forms of nitrate/nitrite. The assessment methods described
in this Protocol utilize EPA guidelines.1

3.1. SPECIFIC AIMS

•	To aid problem formulation, develop a SEM to identify epidemiological (i.e., human),
toxicological (i.e., experimental animal), and supplemental literature pertinent to
characterizing the health effects of ingestion exposure to nitrate and nitrite.

° Epidemiological studies, toxicological studies, and PBPK models are identified for
inclusion based on predefined PECO criteria. The problem formulation PECO used to
develop the SEM is intended to identify the amount and type of evidence available to
address a particular topic and is a useful scoping tool for health effects assessments
fNASEM. 2021: Wolffe etal.. 20191.

° Supplemental material content includes mechanistic studies, including in vivo, in vitro,
ex vivo, or in silico models; nonmammalian model systems; pharmacokinetic and
absorption, distribution, metabolism, and excretion (ADME) studies; human exposure
characteristics (no health outcome); human biomarker studies with a health outcome;
mixture studies; non-ingestion routes of exposure; case studies or case series; records
with no original data; and conference abstracts.

•	Use the results of the SEM to (1) develop refined PECO criteria for the assessment (referred
to as "assessment PECO"); (2) define the unit(s) of analysis at the level of endpoint or health
outcome for hazard characterization; and (3) identify priority analyses of supplemental
material to address the specific aims, uncertainties in hazard characterization,
susceptibility, and dose-response analysis.

•	Conduct study evaluations (risk of bias and sensitivity) for individual epidemiological and
toxicological studies that meet assessment PECO criteria.

•	Conduct a scientific and technical review for PBPK models considered for use in the
assessment If a PBPK or PK model is selected for use, the most reliable dose metric will be
applied based on analyses of the available dose metrics and the outcomes to which they are
being applied.

'EPA guidance documents: http://www.epa.gov/iris/basic-information-about-integrated-risk-information-
svstem#guidance/.

This document is a draft for review purposes only and does not constitute Agency policy.

3-1	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

23

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

•	Conduct data extraction (summarizing study methods and results) from epidemiological
and animal toxicological studies that meet the assessment PECO criteria.

•	For each evidence stream, and for each unit of analysis, use a structured framework to
develop and describe the strength of evidence across studies and the supporting rationale
("evidence synthesis"). Depending on the specific health endpoint or outcome, mechanistic
information and precursor events may be included in a unit of analysis.

•	For each health effect category, use a structured framework to develop and describe weight
of evidence judgments across evidence streams and the supporting rationale for those
judgments ("evidence integration"). The evidence integration analysis presents inferences
and conclusions on human relevance of findings in animals, cross-evidence stream
coherence, potentially susceptible populations and lifestages, and other critical inferences
supported by mechanistic, ADME, or PK/PBPK data (e.g., biological plausibility).

•	For each health effect category, summarize evidence synthesis (strength of evidence) and
evidence integration (weight of evidence) conclusions in an evidence profile table.

•	As supported by the currently available evidence, derive chronic and subchronic oral
reference doses (RfDs) and organ- or system-specific RfDs, and cancer oral slope factors
(OSFs). Apply pharmacokinetic and dosimetry modeling (possibly including PBPK
modeling) to account for interspecies differences, as appropriate. Characterize confidence
in any toxicity values that are derived.

•	Characterize uncertainties and identify key data gaps and research needs, such as
limitations of the evidence base, and consideration of dose relevance and pharmacokinetic
differences when extrapolating findings from higher dose animal studies to lower levels of
human exposure.

This document is a draft for review purposes only and does not constitute Agency policy.

3-2	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

4. LITERATURE SEARCH, SCREENING, AND
LITERATURE INVENTORY

The literature search and screening processes described in this Section were used to
develop an SEM using the problem formulation PECO (see Section 4.1) and supplemental screening
criteria (see Section 4.2) to guide the inclusion of studies. The resulting inventory of studies
identified in the SEM was used to develop the assessment PECO criteria and identify priority
analyses of supplemental material (described in Section 5). The initial literature search as well as
all subsequent literature search updates use the same literature search and screening process, and
therefore the literature inventory is continually updated with new studies as the assessment
progresses.

4.1. POPULATIONS, COMPARATORS, EXPOSURES, OUTCOMES CRITERIA
FOR THE SYSTEMATIC EVIDENCE MAP

PECO criteria are used to focus the assessment question(s), search terms, and inclusion
criteria. To meet the PECO criteria a study must meet all PECO elements. The problem formulation
PECO criteria used to develop the SEM were intentionally broad (see Table 4-1) to identify all the
available evidence in humans and animal models.

During problem formulation, exposure to nitrates/nitrites from routes other than ingestion,
were determined to be out of scope for this assessment. Studies of beneficial health effects were
identified but not included in the study evaluation process since the focus of the assessment is on
hazard identification and dose-response analysis for adverse health effects.

This document is a draft for review purposes only and does not constitute Agency policy.

4-1	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 4-1. Problem formulation populations, exposures, comparators,
outcomes (PECO) criteria for the nitrate and nitrite assessment

PECO
element

Evidence

Populations

Human: Anv population and lifestage (occupational or general population, including children and
other sensitive populations).

Animal: Nonhuman mammalian animal species (whole organism) of anv lifestage (including fetal,
early postnatal, adolescents and adults) that are informative for human health risk assessment.

Exposures

Human: Anv exposure to the nitrate/nitrite forms below via the oral route for anv duration.
Studies will also be included if biomarkers of exposure are evaluated (e.g., measured chemical or
metabolite levels in tissues or bodily fluids) AND there is additional information to allow
estimation/attribution of nitrate/nitrite ingestion (e.g., measures of nitrate/nitrite in
environmental media). If there is no additional information, but the exposure route is unclear or
likely from multiple routes, the study will be tagged as "potentially relevant supplemental
material." Other exposure routes, such as those that are clearly inhalation or dermal, will be
tracked during title and abstract screening and tagged as "potentially relevant supplemental
material."

Animal: Anv exposure to the nitrate/nitrite forms below. Studies involving exposures to mixtures
will be included only if they include an experimental arm with exposure to the nitrate/nitrite
forms below, alone. Other exposure routes, including inhalation or dermal, will be tracked during
title and abstract as "potentially relevant supplemental material."

Relevant forms of nitrate/nitrite: Calcium nitrate, Ammonium nitrate, Potassium nitrate,
Potassium nitrite, Sodium nitrate, Sodium nitrite.

Comparators

Human: A comparison or referent population with exposure to lower levels, no exposure, or
exposure below detection limits; exposure for shorter periods of time; or cases versus controls; or
a repeated measures design. Worker surveillance studies are considered to meet PECO criteria
even if no statistical analyses using a referent group is presented. Case reports or case series of >3
people will be considered to meet PECO criteria, while case reports describing findings in 1-3
people will be tracked as "potentially relevant supplemental material."

Animal: A concurrent control group exposed to vehicle-onlv treatment and/or untreated control.
The control could be a baseline measurement (e.g., acute toxicity studies of mortality) or a
repeated measure design.

Outcomes

All health outcomes are considered relevant (i.e., both cancer and noncancer). In general,
endpoints related to clinical diagnostic criteria, disease outcomes, biochemical, histopathological
examination, or other apical/phenotypic outcomes are considered to meet PECO criteria. We
continue to include relevant studies of methemoglobinemia even though, for this outcome, the
hazard is established. However, the focus is on studies that inform quantitative dose-response
relationships.

4.2. SUPPLEMENTAL CONTENT SCREENING CRITERIA

1	During the literature screening process, studies containing information that may be

2	potentially relevant to the specific aims of the assessment are tagged as supplemental material by

3	category. Some studies could emerge as being critically important to the assessment and may need

This document is a draft for review purposes only and does not constitute Agency policy.

4-2	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	to be evaluated and summarized at the individual study level (e.g., certain cancer MOA or ADME

2	studies), or might be helpful to provide context (e.g., provide hazard evidence from routes or

3	durations of exposure not meeting the PECO), or might not be cited at all in the assessment

4	(e.g., individual studies that contribute to a well-established scientific conclusion). Because it is

5	often difficult to assess the impact of individual studies tagged as supplemental material on

6	assessment conclusions at the screening stage, the tagging structure, described in Table 4-2, allows

7	for easy retrieval later in the assessment process.

This document is a draft for review purposes only and does not constitute Agency policy.

4-3	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 4-2. Categories of potentially relevant supplemental material

Category

Evidence

Classical pharmacokinetic (PK) or
physiologically based
pharmacokinetic (PBPK) model
studies

Classical pharmacokinetic or dosimetry model studies: Classical PK or dosimetry modeling usually divides the body into
just one or two compartments, which are not specified by physiology, wherein movement of a chemical into, between,
and out of the compartments is quantified empirically by fitting model parameters to absorption, distribution,
metabolism, and excretion (ADME) data. This category is for papers that provide detailed descriptions of PK models that
are not physiologically based PK (PBPK) models.

•	The data are typically the concentration time course in blood or plasma after oral and or intravenous exposure,
but other exposure routes can be described.

•	A classical PK model might be elaborated from the basic structure applied in standard PK software, for example to
include dermal or inhalation exposure, or growth of body mass over time, but otherwise does not use specific
tissue volumes or blood flow rates as model parameters.

•	Such models can be used for extrapolation similar to PBPK models, although such use might be more limited.

Note: ADME studies often report classical PK parameters, such as bioavailability (fraction of an oral dose absorbed),
volume of distribution, clearance rate, and/or half-life or half-lives. If a paper provides such results only in tables with
minimal description of the underlying model or software (i.e., uses standard PK software without elaboration), including
"noncompartmental analysis," it should only be listed as a supplemental material ADME study.

Physiologically based pharmacokinetic or mechanistic dosimetry model studies: PBPK models represent the body as
various compartments (e.g., liver, lung, slowly perfused tissue, richly perfused tissue) to quantify the movement of
chemicals or particles into and out of the body (compartments) by defined routes of exposure, metabolism, and
elimination, and thereby estimate concentrations in blood or target tissues.

•	Usually specific to humans or defined animal species; often a single model structure is calibrated for multiple
species.

•	Some mechanistic dosimetry models might not be compartmental PBPK models but predict dose to the body or
specific regions or tissues based on mechanistic data, such as ventilation rate and airway geometry.

•	A defining characteristic is that key parameters are determined from a substance's physicochemical parameters
(e.g., particle size and distribution, octanol-water partition coefficient) and physiological parameters

(e.g., ventilation rate, tissue volumes); that is, data that are independent of in vivo ADME data that are otherwise
used to estimate model parameters.

This document is a draft for review purposes only and does not constitute Agency policy.

4-4	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Category

Evidence



• Chemical-specific information on metabolism (e.g., l/max, Km) or other molecular processes (e.g., protein binding)
might be obtained by fitting the model to in vivo ADME data or determined from in vitro experiments and
extrapolated to in vivo predictions.

Allow extrapolation between species, routes of exposure, or exposure durations and levels; that is, they do not just
quantify ADME for specific experiments to which they have been fitted.

Pharmacokinetic (ADME)

Pharmacokinetic (ADME) studies are primarily controlled experiments in which defined exposures usually occur by
intravenous, oral, inhalation, or dermal routes, and the concentration of particles, a chemical, or its metabolites in blood
or serum, other body tissues, or excreta are then measured.

•	These data are used to estimate the amount absorbed (A), distributed to different organs (D), metabolized (M),
and/or excreted (E) through urine, breath, or feces.

•	The most informative studies involve measurements over time such that the initial increase and subsequent
concentration decline is observed, preferably at multiple exposure levels.

•	Data collected from multiple tissues or excreta at a single time point also inform distribution.

•	ADME data can also be collected from human subjects who have had environmental or workplace exposures that
are not quantified or fully defined. However, to be useful such data must involve either repeated measurements
over a time period when exposure is known (e.g., is zero because previous exposure ended) or time- and
subject-matched tissue or excreta concentrations (e.g., plasma and urine, or maternal and cord blood).

•	ADME data, especially metabolism and tissue partition coefficient information, can be generated using in vitro
model systems. Although in vitro data may not be as definitive as in vivo data, these studies should also be
tracked as ADME. For large evidence bases it may be appropriate to separately track the in vitro ADME studies.

Note: Studies describing environmental fate and transport or metabolism in bacteria or model systems not applicable to
humans or animals should not be tagged.

Mechanistic

Studies reporting measurements related to a health outcome that inform the biological or chemical events associated with
phenotypic effects, in both mammalian and nonmammalian model systems, including in vitro, in vivo (by various routes of
exposure), ex vivo, and in silico studies. Studies in which the chemical is used as a laboratory reagent generally do not
need to be tagged (e.g., as a chemical probe used to measure antibody response).

Non-PECO animal model (i.e.,
nonmammalian systems)

Studies reporting outcomes in animal models that meet the outcome criteria but do not meet the "P" in the PECO criteria.
Depending on the endpoints measured in these studies, they can also provide mechanistic information (in these cases
studies should also be tagged "mechanistic endpoints").

Non-PECO route of exposure

Studies using routes of exposure that fall outside the PECO scope.

This document is a draft for review purposes only and does not constitute Agency policy.

4-5	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Category

Evidence

Human exposure and
biomonitoring (no health
outcome)

Exposure characteristic studies include data that are unrelated to toxicological endpoints, but which provide information
on exposure sources or measurement properties of the environmental agent (e.g., demonstrate a biomarker of exposure).

Biomarker studies for which
exposure route is unknown and
cannot be inferred

Studies evaluate health effects in relation to biomarkers of nitrate and/or nitrite exposure (e.g., urinary, or salivary levels)
without additional information to inform exposure via ingestion.

Mixture studies

Mixture studies that are not considered PECO relevant because they do not contain an exposure or treatment group
assessing only the chemical of interest. This categorization generally does not apply to epidemiological studies in which
the exposure source might be unclear.

Case reports or case series

Case reports describing health outcomes after exposure are tracked as potentially relevant supplemental information
when the number of subjects is <3.

Records with no original data

Records that do not contain original data, such as other agency assessments, informative scientific literature reviews,
editorials, or commentaries.

Posters, conference abstracts,
abstract-only

Records that do not contain sufficient documentation to support study evaluation and data extraction.

This document is a draft for review purposes only and does not constitute Agency policy.

4-6	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

4.3. LITERATURE SEARCH STRATEGIES

4.3.1.	Database Search Term Development

The database search terms focused only on the chemical names and CASRNs, limited to
publication years 2016-2022 with the exception that no year limit was placed on the search for
calcium nitrate as it was not considered in the earlier (2017) IAP.

4.3.2.	Database Searches

The literature search focused on studies published after the period covered by the ATSDR
Toxicological Profile fATSDR. 20171. namely 1/1/2016 onward. This literature search was initially
conducted in August 2022 and regular updates performed with the most recent update occurring in
August 2022. The databases listed below are searched by an EPA information specialist and stored
in the Health and Environmental Research Online (HERO)2 database.

•	PubMed fNational Library of Medicine 1

•	Web of Science (WoS; Thomson Reuters!: given the number of records identified from an
initial WoS search, a more targeted WoS search strategy was used to identify the records
most likely to be applicable to human health (see Appendix A)

After deduplication in HERO, records are imported into SWIFT Review software (Howard et
al.. 2016) to identify those references most likely to be applicable to a human health assessment. In
brief, SWIFT Review has preset literature search strategies ("filters") developed and applied by
information specialists to identify studies more likely to be useful for identifying human health
content from those that likely are not (e.g., analytical methods). The filters function like a typical
search strategy in which studies are tagged as belonging to a certain filter if the terms in the filter
literature search strategy appear in title, abstract, keyword or medical subject headings (MeSH)
fields content The applied SWIFT Review filters focused on lines of evidence: human, animal
models for human health, and in vitro studies. The details of the search strategies that underlie the
filters are available online. Studies not retrieved using these filters are not considered further.
Studies that included one or more of the search terms in the title, abstract, keyword, or MeSH fields
are exported as a RIS (Research Information System) file for further screening as described below.
The impact of application of the SWIFT evidence stream filters on the number of studies for title
and abstract screening is presented in Appendix A.

The literature searches are updated throughout the assessment's development and review
process to identify newly published literature. During this period the literature search terms do not
change from that used in the initial search and studies are screened according to both the problem
formulation and assessment PECO criteria. Thus, the literature inventory is updated during the

2Health and Environmental Research Online: https: //hero.epa.gov/hero/.

This document is a draft for review purposes only and does not constitute Agency policy.

4-7	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

34

35

36

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

process of developing the draft assessment The last full literature search update is conducted
several months prior to the planned release of the draft document for public comment. Studies
identified after peer review begins are only considered for inclusion if they are directly relevant to
the assessment PECO criteria and are expected to fundamentally alter the draft assessment
conclusions.

4.3.3. Searching Other Sources

For this assessment, the starting point is the 2017 ATSDR Toxicological Profile, thus the
literature search aimed to identify studies published in 01/2016 or later. The literature search will
be expanded in subsequent stages of assessment development, to identify any potentially missed
studies from previous years (or published after the end of the current literature search timeframe)
for the health effect categories selected for hazard characterization.

The literature search strategy described above was designed to be broad, but like any
search strategy, studies can be missed [e.g., cases in which the specific chemical is not mentioned in
title, abstract, or keyword content; ability to capture "gray" literature (studies not reported in the
peer-reviewed literature) that is not indexed in the databases listed above]. Thus, in addition to the
database searches, the sources below are used to identify studies that could have been missed
based on the database search. Searching of these resources occurs during preparation of the initial
literature inventory when assembling the SEM. After preparation of the initial literature inventory,
references can be identified during public comment periods, by technical consultants, and during
peer review. Records that appear to meet the initial PECO criteria are uploaded into DistillerSR,
annotated with respect to source of the record, and screened using the methods described in
Section 4.4. Appendix D.l describes the specific methods and results for searching the sources
below. Searching of these sources is summarized to include the source type or name, the search
string (when applicable), number of results present within the resource, and the URL (uniform
resource locator, when available and applicable). The list of other sources consulted includes:

•	Manual review (at the title level) of the reference list from other publicly available final or
draft assessments from other non-EPA Agencies including studies published after 2015
(e.g., ATSDR Toxicological Profile) or published journal review specifically focused on
human health. Reviews can be identified from the database search or from the resources
listed in Appendix D.

•	European Chemicals Agency (ECHA) registration dossiers to identify data submitted by
registrants fhttp://echa.europa.eu/information-on-chemicals/information-from-existing-
substances-regulation).

•	EPA ChemView database fU.S. EPA. 2019al to identify unpublished studies, information
submitted to EPA under Toxic Substances Control Act Section 4 (chemical testing results),
Section 8(d) (health and safety studies), Section 8(e) (substantial risk of injury to health or
the environment notices), and FYI (for your information, voluntary documents). Other

This document is a draft for review purposes only and does not constitute Agency policy.

4-8	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	databases accessible via ChemView include EPA's High Production Volume Challenge

2	database and the Toxic Release Inventory database.

3	• The NTP database of study results and research projects f https: //ntp.niehs.nih.gov/data].

4	• The Organization for Economic Cooperation and Development Screening Information

5	DataSet (SIDS) High Production Volume Chemicals

6	f https://www.echemportal.org/echemportal/].

7	• The EPA CompTox (Computational Toxicology Program) Chemical Dashboard fU.S. EPA.

8	2019b) to retrieve a summary of any ToxCast or Tox21 high throughput screening

9	information. This data will be evaluated and, if amenable, used to generate mechanistic

10	insight, predict adverse outcome, and potentially inform dose-response modeling. Their

11	importance for outcome prediction and dose-response modeling depends on the context,

12	size and quality, and information value of retrieved results and the lack of availability of

13	other data typically used for these purposes.

14	• The National Institute of Health Gene Expression Omnibus (GEO)

15	fhttp://ncbi.nlm.nih.gov/geo/) and the European Bioinformatics Institute (EMBL-EBI)

16	Array Express fhttp: //ebi.ac.uk/biostudies/arrayexpress] repositories to retrieve

17	functional genomics data from appropriate in vitro and in vivo studies. If available, this data

18	will be evaluated and potentially used to generate mechanistic insight, predict adverse

19	outcomes, and inform dose-response assessment

20	• Review of the list of references in the ECOTOX database for the chemical(s) of interest.

21	• Comparative Toxicogenomics Database (CTDB). available at http://ctdbase.org/.

22	• References identified during public comment periods, by technical consultants, and during

23	peer review.

4.3.4. Non-Peer-Reviewed Data

24	IRIS assessments rely mainly on publicly accessible, peer-reviewed studies. However, it is

25	possible that unpublished data directly relevant to the PECO may be identified during assessment

26	development. In these instances, the EPA will try to get permission to make the data publicly

27	available (e.g., in HERO); data that cannot be made publicly available are not used in IRIS

28	assessments. In addition, on rare occasions when unpublished data would be used to support key

29	assessment decisions (e.g., deriving a toxicity value), EPA may obtain external peer review if the

30	owners of the data are willing to have the study details and results made publicly accessible, or if an

31	unpublished report is publicly accessible (or submitted to EPA in a non-confidential manner) fU.S.

32	EPA. 2015). This independent, contractor driven, peer review would include an evaluation of the

33	study similar to that for peer review of a journal publication. The contractor would identify and

34	typically select three scientists knowledgeable in scientific disciplines relevant to the topic as

35	potential peer reviewers. Persons invited to serve as peer reviewers would be screened for conflict

36	of interest. In most instances, the peer review would be conducted by letter review. The study and

37	its related information, if used in the IRIS assessment, would become publicly available. In the

This document is a draft for review purposes only and does not constitute Agency policy.

4-9	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

assessment, EPA would acknowledge that the document underwent external peer review managed
by the EPA, and the names of the peer reviewers would be identified. In certain cases, IRIS will
assess the utility of a data analysis of accessible raw data (with descriptive methods) that has
undergone rigorous quality assurance/quality control review (e.g., ToxCast/Tox21 data, results of
NTP studies not yet published) but that have not yet undergone external peer review.

Unpublished data from personal author communication can supplement a peer-reviewed
study if the information is made publicly available. If such ancillary information is acquired, it is
documented in the Health Assessment Workspace Collaborative (HAWC) or HERO project page
(depending on the nature of the information received).

4.4. LITERATURE SCREENING

Records identified from the literature searches are housed in the HERO system and
imported into SWIFT-Active Screener fhttps://www.sciome.com/swift-activescreener/) for an
initial title abstract screen using machine learning followed by import into DistillerSR (Evidence
Partners; https: //distillercer.com/products/distillersr-systematic-review-software/) for full-text
screening. Both title-and-abstract (TIAB) and full-text screening are conducted by two independent
reviewers.

4.4.1. Title-and-Abstract Screening

The studies identified from the searches described above are imported into SWIFT-Active
Screener fhttps://www.sciome.com/swift-activescreener/) for TIAB screening. SWIFT-Active
Screener is a web-based collaborative software application that utilizes active machine learning
approaches to reduce the screening effort fHoward et al.. 20201. Following a pilot phase to calibrate
screening guidance, two screeners independently perform a TIAB screen using a structured form.
Studies considered "relevant" or "unclear" based on meeting all problem formulation PECO criteria
at the TIAB level are considered for inclusion and advanced to full-text screening. TIAB screening is
conducted by two independent reviewers and any screening conflicts are resolved by discussion
between the primary screeners with consultation by a third reviewer, if needed. For citations with
no abstract, articles are initially screened based on the following: title relevance (title should
indicate clear relevance), and page length (articles two pages in length or less are assumed to be
conference reports, editorials, or letters). Eligibility status of non-English studies is assessed using
the same approach with online translation tools or engagement with a native speaker.

The machine learning screening process is designed to prioritize references that appear to
meet the problem formulation PECO criteria or supplemental material content for manual review
(i.e., both types of references are screened as "include" for machine learning purposes). Screening
continues until SWIFT-Active Screener indicates that it was likely at least 95% of the relevant
studies are identified, a percent identification often used to evaluate the performance of machine
learning applications and considered comparable to human error rates fBannach-Brown et al..

This document is a draft for review purposes only and does not constitute Agency policy.

4-10	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

2018: Howard etal.. 2016: Cohen etal.. 20061. Any studies with "partially screened" status at the
time of reaching the 95% threshold are then fully screened. Studies identified as meeting the
problem formulation PECO criteria, unclear, or supplemental material during TIAB screening are
then imported into DistillerSR software fhttps: //www.evidencepartners.com/products/distillersr-
svstematic-review-software/I either for conflict resolution or for an additional round of more
specific TIAB tagging (i.e., to separate studies meeting PECO criteria versus supplemental content
and to tag the evidence stream or specific type of supplemental content). In DistillerSR, TIAB
screening is conducted by two independent reviewers and any screening conflicts resolved by
discussion between the primary screeners with consultation by a third reviewer, if needed.

Conflicts between screeners in applying the supplemental tags, which primarily occur at the TIAB
level, are resolved similarly, erring on the side of over-tagging based on TIAB content Note that
more granular sub-tagging of supplemental material occurs during preparation of the literature
inventory as described in Section 4.5.2.

4.4.2.	Full-Text Screening

Full-text references are sought through the EPA's HERO database for studies screened as
meeting the problem formulation PECO criteria or "unclear" based on the TIAB screening. Full-text
screening occurs in Distiller SR. Full-text copies of these records are retrieved, stored in the HERO
database, and independently assessed by two screeners using a structured form in DistillerSR to
confirm eligibility. Screening conflicts are resolved by discussion among the primary screeners with
consultation by a third reviewer or technical advisor (as needed to resolve any remaining
disagreements). Rationales for excluding studies are documented, e.g., study did not meet PECO,
full-text not available. Approaches for language translation include online translation tools or
engagement of a native speaker. Fee-based translation services for non-English studies are typically
reserved for studies that are anticipated as being useful for toxicity value derivation.

4.4.3.	Multiple Publications of the Same Data

When there are multiple publications using the same or overlapping data, all publications
are included, with one selected for use as the primary study; the others are considered as
secondary publications with annotation in HAWC and HERO indicating their relationship to the
primary record during data extraction. For epidemiology studies, the primary publication is
generally the one with the longest follow-up, the largest number of cases, or the most recent
publication date. For animal studies, the primary publication is typically the one with the longest
duration of exposure, the largest sample size, or with the outcome(s) most informative to the initial
PECO. For both epidemiology and animal studies, the assessments include relevant data from all
publications of the study, although if the same data are reported in more than one study, the data
are extracted only once (see Section 7). For corrections, retractions, and other companion
documents to the included publications, a similar approach to annotation is taken and the most
recently published data are incorporated into the assessments.

This document is a draft for review purposes only and does not constitute Agency policy.

4-11	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

4.4.4. Literature Flow Diagrams

The results of the screening process are posted on the project page for the assessment in
the HERO database fhttps://heronetepa.gov/heronet/index.cfm/proiect/page/project id/2367).
Results are also summarized in a literature flow diagram and interactive HAWC literature trees
(where additional sub-tagging beyond what is presented in HERO is documented and visualized,
e.g., more details on the nature of mechanistic or ADME studies).

4.5. LITERATURE INVENTORY

During TIAB or full-text level screening, studies that meet the problem formulation PECO
criteria are categorized by evidence type (human or animal) or category of supplemental
information (e.g., mechanistic, PB the ADME properties are dynamic). Next, study design details for
studies that meet the problem formulation PECO criteria are summarized as described in Section
4.5.1. A more granular tagging of supplemental material is conducted as described in Section 4.5.2.
The results of this categorization and tagging are referred to as the literature inventory and is the
key analysis output of the SEM.

4.5.1.	Studies That Meet Problem Formulation PECO Criteria

Human and animal studies that met the problem formulation PECO criteria after TIAB and
full-text review are briefly summarized using data extraction forms in HAWC (hawc.epa.gov: see
Figure 4-1). The literature inventories are used to inform the assessment PECO criteria and
assessment approach. More detail on the process of summarizing studies is presented in Section 7
(Data Extraction of Study Methods and Results).

4.5.2.	Organizational Approach for Supplemental Material

The results of the supplemental material tagging conducted in DistillerSR are imported into
the literature review module in HAWC, where more granular sub-tagging within a type of
supplemental material content category can be conducted. A single study can have multiple tags.
The degree of sub-tagging depends on the extent of content for a given type of supplemental
material and needs of the assessment with respect to developing human health hazard conclusions
and derivation of toxicity values. Tagging judgments in HAWC are made by one assessment member
and confirmed during preparation of draft assessment by another member of the assessment team.
The overall tagging structure for supplemental material content is presented in Figure 4-1, with
details on sub-tagging presented in the following sections under the specific type of supplemental
content (i.e., mechanistic, ADME and PK/PBPK).

This document is a draft for review purposes only and does not constitute Agency policy.

4-12	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

. , Has additional sub-tagging I
No additional sub-tagging |

Nitrate/Nitrite Literature Tag Tree {2016-2022)

Mechanistic

Animal Study

Biomarker Study (With Health
Outcome)

Other Animal Model {Not PECO
Relevant)

Other Route of Exposure (Not
Ingestion)

No Original Data

Case Study or Case Series

ADME

Human Randomized Controlled

Figure 4-1. Visual summary of approach for tagging major categories of
supplemental material.

Organization of Mechanistic Information

The literature inventory of mechanistic information is used to develop the assessment
approach (see Section 5), in particular to help assess whether any units of analysis should be
defined to include mechanistic information and to identify prioritized analyses. The sub-tagging
structure applied to mechanistic evidence is based on 10 specific mechanism pathways or events,
listed below:

Mitochondrial function.

Inflammation

Oxidative and nitrosative stress
Genotoxicity

Nitrosation of amines/production of nitrosamines
S-Nitrosation

Generation of methemoglobin
Endothelial function

This document is a draft for review purposes only and does not constitute Agency policy.

4-13	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

9)	NO-mediated cell signaling

10)	Modulation of enzyme activity

Figure 4-2 illustrates how the categories are represented in the overall tagging approach for
mechanistic information.

Q Has additional sub-tagging
Q No additional sub-tagging

Nitrate/Nitrite Literature Tag Tree (2016-2022)

Nitrates/Nitrites (2022)

Not PECO Relevant
Supplemental Material

Mechanistic

Human Exposure (No Health
Ouieome)

0

Biomarker Study (With Health
Outcome)

®

Other Animal Model (Not PECO
Relgsiant)

©

Other Route of Exposure (Not

No Original Data

©

Abstract Only

©

Case Study or Case Series

®

ADME

©

Mixture Study

©

Mitochondrial Function

©

Inflammation

@

Oxidative and Nltrosative
Stress. Hypoxia

0

Genotoxicity

_	©

rosatlor
;/Produ<
itro^pfjiim
S-Nltrosatlon

©

Generation of Methemoglobln
©

Endothelial Function
©

Nitric Oxide-Mediated Cell
Signaling

@

Modulation of Enzyme Activity
*

Endoplasmic Reticulum Stress

©

Figure 4-2. Visual summary of overall tagging structure for mechanistic
studies.

Organization of ADME and PK/PBPK Model Information

ADME and PK/PBPK model evidence are tagged as supplemental material in DistillerSR as
outlined in Table 4-2. Tagged ADME studies and PK/PBPK models were imported into the HAWC
Literature Review module and underwent more detailed tagging by disciplinary experts. Primary
data ADME studies are tagged as absorption, distribution, metabolism, or elimination (using a tag
all that apply approach). PK/PBPK models are tagged according to species applicability, i.e., animal,
human, or multiple species (to include human). See Figure 4-3 for organizational structure.

This document is a draft for review purposes only and does not constitute Agency policy,

4-14	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Has additional sub-tagging
Q No additional sub-tagging

Nitrate/Nitrite Literature Tag Tree (2016-2022)



Mechanistic

Nitrates/Nitrites (2022)

©i

Not PECO Relevant
1248

Supplemental Material

©

Other Animal Model (Not PECO
Relevant)

®

No Original Data

—¦©

Abstract Only

"—©

Case Study or Case Series

0

©
Mixture Study

PK/PBPK Model

- ©
Absorption

®
Distribution

©
Metabolism

©

©

Review

Figure 4-3. Visual summary of tagging structure for ADME and PK/PBPK
studies.

This document is a draft for review purposes only and does not constitute Agency policy.

4-15	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

5. REFINED PROBLEM FORMULATION AND
ASSESSMENT APPROACH

5.1. ASSESSMENT PECO CRITERIA

The primary purpose of this step is to provide further specification to the assessment
methods based on characterization of the extent and nature of the evidence identified from the
literature inventory. This includes refinements to PECO criteria and defining the unit(s) of analysis
for health endpoints/outcomes during evidence synthesis, and presenting analysis approaches for
mechanistic, ADME or other types of supplemental material content. A unit of analysis is an
outcome or group of related outcomes within a health effect category that are considered together
during evidence synthesis (see Section 8). In some assessments, the units of analysis may include
predefined categories of mechanistic evidence (e.g., biomarkers or precursors relating to other
outcomes within the unit of analysis, evidence that provides support for grouping together
biologically linked endpoints into a unit of analysis).

Based on the results of the initial literature inventories for the SEM, the problem
formulation PECO criteria were refined to the "assessment" PECO criteria (see Table 5-1). The
assessment PECO criteria reflect the subset of studies that will be the focus of the systematic review
and will move forward for study evaluation and evidence synthesis. The literature search identified
178 human randomized control trials; since these studies concerned protective health effects of
nitrate/nitrite exposure, they will be considered as supplemental material. Note that there were no
studies identified during the primary literature search that evaluated hazards of oral exposure to
calcium nitrate.

The systematic review will focus on the health outcome categories that appear to have
sufficient information available to support hazard identification, based upon the availability of
animal and human studies as cited in ATSDR (ATSDR. 2017) and IARC (IARC. 2010). and the
updated literature search conducted by EPA. EPA anticipates conducting a systematic review for
the following health effect categories, for which the available epidemiology and experimental
animal studies are likely to be sufficient for drawing conclusions about human hazard:

Cancer

ATSDR concluded that "In general outcomes of cohort and case-control studies have found no
or weak associations between nitrate intakes and cancer in humans¦, with stronger associations for
exposures to nitrite or intake of high-nitrite foods such as cured meat" and that "Associations between
intake of nitrite and a variety of cancer types has been studied; however¦, the strongest and most
consistent evidence for carcinogenicity of nitrite derives from studies of gastrointestinal cancers and

This document is a draft for review purposes only and does not constitute Agency policy.

5-1	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

23

24

25

26

27

28

29

30

31

32

33

34

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

in particular¦, gastric cancer (Buiatti et al. 1990; Engel et al 2003; La Vecchia et al. 1994,1997; Mayne
et al. 2001; Palli et al. 2001; Risch et al. 1985; Rogers et al. 1995; Ward et al. 2007,2008). In general,
these studies have found significant positive trends for cancer risk (risk increases with increasing
intake), and three studies found elevated cancer risk (Engel et al. 2003; Kim et al. 2007; Risch et al.
1985)." Since the conclusion of the ATSDR literature search period, 28 human epidemiology studies
have been published that evaluate associations between nitrate/nitrite in water and diet with
cancer at various sites (including eight studies of colorectal cancer, with smaller numbers of studies
evaluating other cancer types). The human cancer studies are supported by two animal studies
evaluating neoplasms in rodent models (one study concerns colorectal cancer, the other multiple
cancers).

Cardiovascular effects

ATSDR found few studies evaluating risk of cardiovascular effects, reflected by the
conclusion in the 2017 IAP. However, EPA's updated literature search identified 13 total new
human epidemiology studies evaluating endpoints including cardiovascular disease mortality (5
studies) as well as cardiovascular and cerebrovascular disease (6 studies) and blood pressure (6
studies). In addition, a large number (48 studies) of toxicology studies were found to have
evaluated cardiovascular endpoints.

Developmental effects

ATSDR identified several studies of developmental effects following nitrate/nitrite
exposure in early life (including in utero), with many of the human studies focusing on risk of
congenital malformation. However, they note that "Several population-based, case-control studies
evaluated possible associations between developmental end points and exposure to nitrate from
drinking water sources. The results are not adequate for quantitative risk assessment because
estimations of nitrate intakes were typically based on measurements of nitrate levels in drinking
water sources at selected time points and self-reported estimates of water consumption, possible
confounding by other potential toxicants was not evaluated, and most studies did not account for
dietary nitrate or nitrite intake which is typically the major source of ingested nitrate and nitrite.
Statistically significant associations between nitrate in the drinking water and selected developmental
end points (e.g., birth defects, spontaneous abortions) were reported by some investigators, but were
not observed by others." EPA found that since the conclusion of the ATSDR search, six new human
studies have been published including two studies evaluating birth defects, three evaluating
measures of early life size and growth and one evaluating offspring mortality. The new body of
human studies is complemented by three new toxicology studies.

This document is a draft for review purposes only and does not constitute Agency policy.

5-2	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	Endocrine effects

2	Nitrate is a competitive inhibitor of the sodium iodide symporter, therefore endocrine

3	effects due to nitrate/nitrite exposure are of concern. ATSDR noted that "Available human data

4	provide suggestive evidence that elevated levels of nitrate in drinking water and/or nitrate-rich diets

5	may be associated with signs of thyroid dysfunction. However¦, limitations of these studies include lack

6	of individual dose-response data, quantification of iodine intake, and control for other potential

7	substances that may affect the thyroid; one study relied on self-reported thyroid status and self-

8	reported dietary nitrate intake." Since then, a small number of new human and animal studies (four

9	human and two animal) have been published evaluating hypothyroidism and thyroid abnormalities

10	in humans, and thyroid hormone levels and function in animals.

11	Hematological effects

12	All existing toxicity values have been based upon methemoglobinemia. EPA identified one

13	new human study with this endpoint, which will be evaluated for its potential to support dose-

14	response characterization. Additionally, 11 new animal studies have been published that evaluate

15	both methemoglobin levels and other hematologic endpoints. While the hazard for hematological

16	endpoints is considered well-established and will not be revisited, new studies evaluating

17	methemoglobinemia and related endpoints will be considered for their potential to support dose-

18	response evaluation.

19	Hepatic effects

20	ATSDR did not identify any human studies of hepatic effects and noted that the five animal

21	studies identified did not show associations with nitrate/nitrite exposure. However, since that time,

22	22 new animal studies have been published evaluating a variety of endpoints including liver

23	function biomarkers (such as alanine aminotransferase and aspartate aminotransferase) and liver

24	histopathology. In addition, there has been one new human study evaluating mortality due to

25	chronic liver disease.

26	Metabolic effects

27	ATSDR identified a number of human studies (but few animal studies) evaluating metabolic

28	effects, noting that "Possible associations between nitrate and/or nitrite in drinking water and/or

29	food sources and risk of type 1 diabetes have been investigated in a number of epidemiological studies

30	(Casu et al. 2000; Dahlquist et al. 1990; Kostraba et al. 1992; Moltchanova et al. 2004; Parslow et al.

31	1997; van Maanen et al. 2000; Zhao et al. 2001). Statistically significant associations between

32	estimated nitrate and/or nitrite intake were reported by some investigators but were not observed by

33	others. Limitations of studies include the lack of quantitative dose-response data and the likelihood of

34	confounding by other potential toxicants. Therefore, there is considerable uncertainty regarding

This document is a draft for review purposes only and does not constitute Agency policy.

5-3	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	nitrate or nitrite intake and risk of type 1 childhood diabetes." EPA identified one new study in

2	humans evaluating type 1 diabetes; two other human studies evaluating type 2 diabetes; and

3	metabolic dysfunction and one study evaluating mortality due to diabetes mellitus. However, a

4	large number (50) of new animal studies have measured a variety of endpoints related to lipid

5	levels, insulin, and glucose homeostasis that may inform human health risk for endocrine outcomes.

6	Nervous system effects

7	ATSDR identified few studies of nervous system effects (two human studies reporting

8	headache, and three animal studies). However, since that time three new human studies have been

9	published that evaluated nervous system effects in both adolescents (depressive symptoms) as well

10	as in middle-aged and older adults (cognitive function, mortality due to Alzheimer's disease. In

11	addition, seven new animal studies have evaluated endpoints, including tremor, sensory endpoints,

12	learning, and memory.

13	Reproductive effects

14	Much of the evidence identified in the ATSDR Toxicological Profile is described under

15	developmental effects. However, EPA identified new human studies that evaluated time to

16	pregnancy (one study) or gestation duration (five studies), and 12 new animal studies that

17	evaluated reproductive endpoints in both male and female animals including reproductive

18	hormone levels, reproductive organ histopathology and fertility.

19	Urinary effects

20	The ATSDR Toxicological Profile did not find any human studies, and only one animal study

21	evaluating urinary system effects. However, EPA identified one new human study that evaluated

22	risk of chronic kidney disease and one new human study evaluating mortality due to kidney

23	disease. A larger number (14) of new animal studies have evaluated urinary system effects, mainly

24	kidney function and histopathology.

25	Other health effect categories (not considered further)

26	The health effect categories listed in this section are those for which the ATSDR

27	Toxicological Profile found limited or no epidemiological or toxicological evidence. Further, EPA's

28	updated literature search identified no new substantial evidence. Primarily, these studies

29	investigated health protective effects, which is outside the scope of this assessment Several also

30	only administered nitrate/nitrite with the purpose of inducing toxicity, using doses high enough to

31	be of limited use to generalizing dose-response analysis to target populations. Therefore, none of

32	the following categories will be carried forward for hazard evaluation based on the literature

33	available in 2022, although new evidence may be identified with future literature search updates:

This document is a draft for review purposes only and does not constitute Agency policy.

5-4	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Dermal effects: ATSDR only identified one case study; there were no new human studies and
only one new animal study. This study only administered nitrate/nitrite with the purpose of
inducing toxicity and had limited results reporting.

Gastrointestinal effects: ATSDR identified one human study of acid reflux and a few animal
studies of forestomach epithelial hyperplasia. There was one new human study evaluating risk of
diarrheal disease, and 11 new animal studies. The human study was ecological in design,
correlating diarrheal disease case counts with nitrate levels in water samples (no associations
found (Kulinkina et al.. 201611. Most of the animal studies examined protective effects of
nitrate/nitrite and one only administered nitrate/nitrite with the purpose of inducing toxicity.

Immune effects: ATSDR did not identify any animal or human studies. There was one new
human study evaluating risk of diarrheal disease (discussed above (Kulinkina etal.. 201611. one
new human study evaluating type 1 diabetes and islet autoimmunity ((Mattila etal.. 20201 included
under metabolic effects), and one human study evaluating mortality due to infection. The bulk of
the new animal studies (n = 25) evaluated cytokine levels, markers of inflammation and oxidative
stress, or white blood cell counts (classified as hematological endpoints). One new animal study
examined metabolic islet autoimmunity (included under metabolic effects) and one animal study
evaluated beneficial effect of nitrate in an animal model of colitis (included under gastrointestinal
effects).

Musculoskeletal effects: ATSDR did not identify any animal or human studies. There were
two new human studies evaluating muscle function (EPA inventoried these as 'whole body' effects),
and 10 new animal studies—however, these focused on identifying beneficial effects of exposure to
nitrate/nitrite rather than hazard. Both human studies found protective effects. Three of the animal
studies found non-conclusive or adverse effects, but the rest provided evidence of therapeutic
effects of nitrate/nitrite.

Ocular effects: ATSDR did not identify any animal or human studies. There were three new
human studies evaluating risk of glaucoma (two studies) and retinal microvasculature (one study),
and one animal study evaluating features of macular degeneration. Each human study investigated
and found protective effects of nitrate/nitrite against ocular disease. The animal study looked for
any impact of nitrate/nitrite and found evidence of adverse effects, though at relatively high doses
of exposure.

Respiratory effects: ATSDR did not identify any animal or human studies. One new human
study evaluated mortality from respiratory disease, and three new animal studies evaluated
pulmonary function or histopathology. Of the animal studies, two investigated and found health
protective effects. The third only administered nitrate/nitrite at a very high dose with the purpose
of inducing toxicity.

This document is a draft for review purposes only and does not constitute Agency policy.

5-5	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	As noted in Section 4, the literature inventory initially developed using the problem

2	formulation. PECO is continually updated as the assessment progresses in part to ensure that

3	emerging areas of potential health concern are monitored. The only adjustment made to the

4	approaches used to tag supplemental material presented in Table 4-2 was the addition of human

5	randomized controlled trials as a supplemental material category.

Table 5-1. Assessment PECO criteria for the nitrate/nitrite (oral) assessment

PECO
element

Evidence

Populations

Human: Anv population and lifestage (occupational or general population, including children and
other sensitive populations).

Animal: Nonhuman mammalian animal species (whole organism) of anv lifestage (including fetal,
early postnatal, adolescents and adults) that are informative for human health risk assessment.

Examples:

•	PECO-relevant: humans and laboratory animals, such as mice, rats, guinea pigs, monkeys,
hamsters, dogs, etc.

•	Supplemental: zebrafish in developmental studies, hens in neurotoxicology studies, frog
embryos for teratogenicity; in vitro assays will be tagged as "mechanistic."

•	Not PECO-relevant: birds, trout, salmon, algae, seedlings, hens in feather growth; farm
animals (especially multi-stomach animals) like cattle, sheep, pigs, etc.

Exposures

Human: Anv exposure to the nitrate/nitrite forms below via the oral route for anv duration.
Studies will also be included if biomarkers of exposure are evaluated (e.g., measured chemical or
metabolite levels in tissues or bodily fluids) AND there is additional information to allow
estimation/attribution of nitrate/nitrite ingestion. If there is no additional information, but the
exposure route is unclear or likely from multiple routes, the study will be tagged as "potentially
relevant supplemental material." Other exposure routes, such as those that are clearly inhalation
or dermal, will be tracked during title and abstract screening and tagged as "potentially relevant
supplemental material."

Animal: Anv exposure to the nitrate/nitrite forms below. Studies involving exposures to mixtures
will be included only if they include an experimental arm with exposure to the nitrate/nitrite
forms below, alone. Other exposure routes, including inhalation or dermal, will be tracked during
title and abstract as "potentially relevant supplemental material."

Relevant forms of nitrate/nitrite: Calcium nitrate, Ammonium nitrate, Potassium nitrate,
Potassium nitrite, Sodium nitrate, Sodium nitrite.

This document is a draft for review purposes only and does not constitute Agency policy.

5-6	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

PECO
element

Evidence

Comparators

Human: A comparison or referent population with exposure to lower levels, no exposure, or
exposure below detection limits; exposure for shorter periods of time; or cases versus controls; or
a repeated measures design. Worker surveillance studies are considered to meet PECO criteria
even if no statistical analyses using a referent group is presented. Case reports or case series of >3
people will be considered to meet PECO criteria, while case reports describing findings in 1-3
people will be tracked as "potentially relevant supplemental material."

Animal: A concurrent control group exposed to vehicle-onlv treatment and/or untreated control.
The control could be a baseline measurement (e.g., acute toxicity studies of mortality) or a
repeated measure design.

Outcomes

All health endpoints for the following health effect categories are considered relevant: cancer;

cardiovascular; developmental; endocrine; hematopoietic; hepatic; metabolic; nervous;

reproductive; urinarv. In general, endpoints related to clinical diagnostic criteria, disease
outcomes, biochemical, histopathological examination, or other apical/phenotypic outcomes are
considered to meet PECO criteria. We continue to include relevant studies of methemoglobinemia
even though, for this outcome, the hazard is established. However, the focus is on studies that
inform quantitative dose-response relationships. Human randomized controlled trials examining
the protective effects of nitrate/nitrite exposure will be considered "potentially relevant

supplemental material".

Underlined text shows changes made to the assessment PECO criteria compared to the initial PECO criteria.

5.1.1. Other Exclusions Based on Full-Text Content

1	In addition to failure to meet PECO criteria (described above), epidemiological and

2	toxicological studies may be excluded at the full-text level due to critical reporting limitations.

3	Reporting limitations can be identified during full-text screening but are more commonly identified

4	during subsequent phases of the assessment (e.g., literature inventory, study evaluation).

5	Regardless of when the limitation is identified, exclusions based on full-text content are

6	documented at the level of full-text exclusions in literature flow diagrams with a rationale of

7	"critical reporting limitation."

8	A similar approach is taken for in vitro studies that are prioritized for focused analysis

9	during assessment development (i.e., the critical reporting deficiency may preclude them from

10	consideration). Critical reporting information for different study types are summarized below. For

11	each piece of information, if the information can be inferred (when not directly stated) for an

12	exposure/endpoint combination, the study should be included.

13	Epidemiology studies

14	• Sample size

15	• Exposure characterization and/or measurement method

16	• Outcome ascertainment method

This document is a draft for review purposes only and does not constitute Agency policy.

5-7	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	•	Study design

2	Animal studies

3	•	Species

4	•	Test article name

5	•	Levels and duration of exposure

6	•	Route of exposure

7	•	Quantitative or qualitative (e.g., photomicrographs; author-reported lack of an effect on the

8	outcome) results for at least one endpoint of interest

9	In vitro studies prioritized for focused analysis

10	•	Cell/tissue type(s) or test system

11	•	Test article name

12	•	Concentration and duration of treatment

13	•	Quantitative or qualitative results for at least one endpoint of interest

5.2. UNITS OF ANALYSES FOR DEVELOPING EVIDENCE SYNTHESIS AND
INTEGRATION JUDGMENTS FOR HEALTH EFFECT CATEGORIES

14	The planned units of analysis based on outcomes identified in the assessment PECO are

15	summarized in Table 5-2. General considerations for defining the units of analysis are presented in

16	the IRIS Handbook. Each unit of analysis is initially synthesized and judged separately within an

17	evidence stream (see Section 8.1). Depending on the specific health endpoint or outcome, PK data,

18	mechanistic information, and other supporting evidence (e.g., from studies of non-PECO routes of

19	exposure) may be included in a unit of analysis.

20	The units of analysis can also include or be framed to focus on precursor events (e.g.,

21	biomarkers). Evidence integration judgments focus on the stronger within evidence stream

22	synthesis conclusions when multiple units of analysis are synthesized. The evidence synthesis

23	judgments are used alongside other key considerations (i.e., human relevance of findings in animal

24	evidence, coherence across evidence streams, information on susceptible populations or lifestages,

25	and other critical inferences that draw on mechanistic evidence) to draw an overall evidence

26	integration judgment for each health effect category or more granular health outcome grouping

27	(see Section 8.2).

This document is a draft for review purposes only and does not constitute Agency policy.

5-8	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 5-2. Health effect categories and human and animal evidence unit of analysis endpoint groupings for which
evidence integration judgments will be developed

Health effect categories
for evidence integration

Units of analysis for evidence synthesis that inform evidence integration
(each bullet represents a unit of analysis)



Human evidence

Animal evidence

Cardiovascular

•	Cardiovascular disease and mortality;
cerebrovascular disease

•	Blood pressure

•	Blood pressure and other measures of vascular function

•	Heart and vessel morphology

•	Heart function

Developmental

•	Fetal viability/pregnancy outcomes
(spontaneous abortion)

•	Congenital malformations

•	Size and weight in early life

•	Fetal viability/survival or other birth parameters (e.g.,
resorptions, number of pups per litter)

•	Fetal growth (e.g., weight or length)

(Note: An analysis of dam health (e.g., weight gain, food consumption) is
also conducted to support conclusions of specificity of the effects as
being developmental (versus derivative of maternal toxicity).)

Endocrine

• Thyroid hormones and antibodies; goiter

•	Thyroid hormones

•	Thyroid morphology/histopathology

Hematopoietic (focus on
studies to support dose-
response)

• Methemoglobin

• Methemoglobin

Hepatic

• (None identified)

•	Liver function biomarkers (including liver enzymes)

•	Liver histopathology

Metabolic

• Metabolic dysfunction, including diabetes

• Serum lipid measures (e.g., triglycerides; cholesterol)

This document is a draft for review purposes only and does not constitute Agency policy.

5-9	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Health effect categories
for evidence integration

Units of analysis for evidence synthesis that inform evidence integration
(each bullet represents a unit of analysis)



Human evidence

Animal evidence





•	Indicators of insulin production and glucose homeostasis

•	Adiposity

Nervous

•	Cognitive function in adulthood

•	Depressive symptoms

•	Neurodegenerative disease

•	Learning/memory

•	Brain morphology/histopathology

•	Neurodegenerative disease

•	Sensory processing

Reproductive

• Gestational length (e.g., preterm birth)

•	Reproductive hormone levels

•	Sperm parameters

•	Reproductive organ morphology/histopathology

•	Fertility

Urinary

• Kidney disease

•	Kidney function biomarkers

•	Kidney morphology/histopathology

This document is a draft for review purposes only and does not constitute Agency policy.

5-10	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Health effect categories
for evidence integration

Units of analysis for evidence synthesis that inform evidence integration
(each bullet represents a unit of analysis)



Human evidence

Animal evidence

Carcinogenicity

• Colorectal cancer

• Colorectal cancer precursors



• Breast cancer

• All other cancer endpoints observed as part of general toxicity



• Gastrointestinal tract cancer

assays



• Bladder cancer





• Kidney cancer





• Central nervous system cancer





• Thyroid cancer





• Liver cancer





• Cancer of reproductive organs





• Reticuloendothelial cancer





• Cancer mortality



This document is a draft for review purposes only and does not constitute Agency policy.

5-11	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

5.3. CONSIDERATION OF SUPPLEMENTAL MATERIAL

5.3.1.	Noncancer MOA Mechanistic Information

The non-carcinogenesis mechanistic studies were screened and tagged according to the
relevant target organ/health system as described in Section 4.5.2. Findings from newly identified
studies will be briefly summarized in tabular format. Nonmammalian model systems were included
in this analysis. These summary conclusions regarding mechanisms of toxicity for nitrate and
nitrite will be used to support evidence integration conclusions for specific health system hazard
analyses as well as describe general features of mechanisms of toxicity.

5.3.2.	ADME and PK/PBPK Model Information

Studies containing ADME and PK/PBPK content were screened and tagged as described in
Section 4.5.2. Oral pharmacokinetics of nitrates and nitrites are the primary focus since the current
assessment focuses on the derivation of oral toxicity values. However, pharmacokinetic studies
from alternate routes of exposure can still inform various aspects of ADME and are also considered.

For supplemental material studies categorized as PK/PBPK models, only three such models
were identified: fZeilmaker etal.. 20101 (an application of the Zeilmaker, 1996, 3859914 model),
fLin etal.. 20201 (an updated parametrization of the Zeilmaker, 1996, 3859914 model), and
(Coggan and Thies. 20201. With the limited number of studies, an initial scoping process is not
needed, and all three models will be evaluated for their suitability for deriving toxicity values for
the nitrates/nitrites assessment (for more detail, see Section 6.6, Pharmacokinetic Model
Evaluation). Model determination will include the evaluation of underlying pharmacokinetic data
for training, model assumptions relative to known ADME, and the ability to predict internal dose
metrics of interest.

5.3.3.	Other Supplemental Material Content

Structured approaches to organize evidence like those presented for genotoxicity
mechanistic studies, noncancer MOA, and ADME/PK/PBPK were not developed for other types of
supplemental material. Instead, the tagged material was reviewed during preparation of the draft to
see if studies were available to address specific uncertainties of the health study evidence base,
inform susceptibility conclusions, and ensure completeness of identifying primary data papers
most pertinent to the assessment.

•	Titles of studies tagged as exposure-only are reviewed to see if they provided information
pertinent to establish study evaluation considerations for the exposure domain.

•	Titles of review articles are reviewed to identify those that are directly pertinent to the
scope of the assessment The reference lists of such reviews are scanned to identify primary
data studies that might have been missed from database search queries. The reviews may

This document is a draft for review purposes only and does not constitute Agency policy.

5-12	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

also be used to provide perspective on interpretation of foundational science cited in the
assessment.

Other types of supplemental material did not undergo additional analysis because the
information was not considered likely to impact toxicity value development (including
application of uncertainty factors). The specific categories are case reports, mixtures, or
conference abstracts.

This document is a draft for review purposes only and does not constitute Agency policy.

5-13	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

6. STUDY EVALUATION (RISK OF BIAS AND
SENSITIVITY)

The general approach for evaluating primary health effect studies that meet PECO is
described in Section 6.1 Instructional and informational materials for study evaluations are
available at https: //hawc.epa.gOv/assessment/l00000039 /. The approach is conceptually the
same for epidemiology, controlled human exposure, animal toxicology, and in vitro studies but the
application specifics differ; thus, they are described separately in Sections 6.2, 6.3 and 6.4,
respectively. Any physiologically based PBPK models used in the assessment are evaluated using
methods described in the Quality Assurance Project Plan for PBPK models fU.S. EPA. 20181. which
is summarized below (see Section 6.6).

6.1. STUDY EVALUATION OVERVIEW FOR HEALTH EFFECT STUDIES

The IRIS program uses a domain-based approach to evaluate studies. Key concerns for the
review of epidemiology and animal toxicology studies are potential bias (factors that affect the
magnitude or direction of an effect in either direction) and insensitivity (factors that limit the
ability of a study to detect a true effect; low sensitivity is a bias toward the null when an effect
exists). The study evaluations are aimed at discerning the expected magnitude of any identified
limitations (focusing on limitations that could substantively change a result), considering the
expected direction of the bias. The study evaluation approach is designed to address a range of
study designs, health effects, and chemicals. The general approach for reaching an overall judgment
regarding confidence in the reliability of the results is illustrated in Figure 6-1.

This document is a draft for review purposes only and does not constitute Agency policy.

6-1	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

(a) Individual evaluation domains

Epidemiology

Animal

In vitro

•	Exposure measurement

¦	Outcome ascertainment

¦	Participant selection

•	Confounding

•	Analysis

•	Selective reporting

•	Sensitivity

¦ Allocation

•	Observational bias/blinding

•	Confounding

•	Attrition

•	Chemical administration and
characterization

•	Endpoint measurement

•	Results presentation

•	Selective reporting

•	Sensitivity

¦	Observational bias/blinding

¦	Variable control

¦	Selective reporting

•	Chemical administration and
characterization

•	Endpoint measurement

•	Results presentation

•	Sensitivity

(b) Domain level judgements and overall study rating

Domain judgments

Judgment

Interpretation

0 Good
Adequate

Deficient

0 Critically
Deficient

Appropriate study conduct relating to the domain and minor
deficiencies not expected to influence results.

A study that may have some limitations relating to the
domain, but they are not likely to be severe or to have a
notable impact on results.

Identified biases or deficiencies interpreted as likely to have
had a notable impact on the results or prevent reliable
interpretation of study findings.

A serious flaw identified that makes the observed effect(s)
uninterpretable. Studies with a critical deficiency are
considered "uninformative" overall.

Overall study rating for an outcome

Rating

Interpretation

High

Medium
Low

Uninformative

No notable deficiencies or concerns identified; potential for
bias unlikely or minimal; sensitive methodology.

Possible deficiencies or concerns noted but they are unlikely
to have a significant impact on results.

Deficiencies or concerns were noted, and the potential for
substantive bias or inadequate sensitivity could have a
significant impact on the study results or their interpretation.
Serious flaw(s) makes study results uninterpretable but may
be used to highlight possible research gaps.

Figure 6-1. Overview of Integrated Risk Information System (IRIS) study
evaluation approach, (a) individual evaluation domains organized by evidence
type, and (b) individual evaluation domains judgments and definitions for overall
ratings (i.e., domain and overall judgments are performed on an outcome-specific
basis).

1	To calibrate the assessment-specific considerations, the study evaluation process includes a

2	pilot phase to assess and refine the evaluation process. Following this pilot, at least two reviewers

3	independently evaluate studies to identify characteristics that bear on the informativeness

This document is a draft for review purposes only and does not constitute Agency policy,

6-2	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

40

41

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

(i.e., validity and sensitivity) of the results. The independent reviewers use structured web-forms
for study evaluation housed within the EPA's version of HAWC

fhttps: //hawc.epa.gov/assessment/100500308/1 to record separate judgments for each domain
and the overall study for each outcome and unit of analysis, to reach consensus between reviewers,
and when necessary, resolve differences by discussion between the reviewers or consultation with
additional independent reviewers. As reviewers examine a group of studies, additional chemical-
specific knowledge or methodological concerns could emerge, and a second pass of all pertinent
studies might become necessary.

In general, considerations for reviewing a study with regard to its conduct for specific
health outcomes are based on considerations presented in the IRIS Handbook fU.S. EPA. 20221 and
use of existing guideline documents when available, including EPA guidelines for carcinogenicity,
neurotoxicity, reproductive toxicity, and developmental toxicity fU.S. EPA. 2005a. 1998.1996.
1991a).

Study authors may be queried for information, especially if manuscripts are missing key
information on study design or relevant results. Queries may also be made to inquire about
additional analyses that could address major study limitations. During study evaluation, the
decision on whether to seek missing information focuses on information that could result in a
reevaluation of the overall study confidence for an outcome. Outreach to study authors is
documented in HAWC and considered unsuccessful if researchers do not respond to an email or
phone request within one month of the attempt to contact. Only information or data that can be
made publicly available (e.g., within HAWC or HERO) will be considered.

When evaluating studies that examine more than one outcome, the evaluation process is
explicitly conducted at the individual outcome level within the study. Thus, the same study may
have different outcome domain judgments for different outcomes. These measures could still be
grouped for evidence synthesis.

During review, for each evaluation domain, reviewers reach a consensus judgment of good,
adequate, deficient, not reported, or critically deficient. If a consensus is not reached, a third
reviewer performs conflict resolution. It is important to emphasize that evaluations are performed
in the context of the study's utility for identifying individual hazards. Limitations specific to the
usability of the study for dose-response analysis are useful to note and applicable to selecting
studies for that purpose (see Section 9), but they do not contribute to the study confidence
classifications. These four categories are applied to each evaluation domain for each outcome
considered within a study, as follows:

•	Good represents a judgment that the study was conducted appropriately in relation to
the evaluation domain, and any minor deficiencies noted are not expected to influence
the study results or interpretation of the study findings.

•	Adequate indicates a judgment that methodological limitations related to the evaluation
domain are (or are likely to be) present, but those limitations are unlikely to be severe
or to notably impact the study results or interpretation of the study findings.

•	Deficient denotes identified biases or deficiencies interpreted as likely to have had a
notable impact on the results, or that limit interpretation of the study findings.

This document is a draft for review purposes only and does not constitute Agency policy.

6-3	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

40

41

42

43

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

•	Not reported indicates the information necessary to evaluate the domain question was
not available in the study. Depending on the expected impact, the domain may be
interpreted as adequate or deficient for the purposes of the study confidence rating.

•	Critically deficient reflects a judgment that the study conduct relating to the evaluation
domain introduced a serious flaw that is interpreted to be the primary driver of any
observed effect(s) or makes the study uninterpretable. Studies with critically deficient
judgments in any evaluation domain are almost always classified as overall
uninformative for the relevant outcome(s).

Once the evaluation domains are rated, the identified strengths and limitations are
considered collectively to reach a study confidence classification of high, medium, or low confidence,
or uninformative for each specific health outcome (s). This classification is based on the reviewer
judgments across the evaluation domains and considers the likely impact that the noted
deficiencies in bias and sensitivity have on the outcome-specific results. There are no predefined
weights for the domains, and the reviewers are responsible for applying expert judgment to make
this determination. The study confidence classifications, which reflect a consensus judgment
between reviewers, are defined as follows:

•	High confidence: No notable deficiencies or concerns were identified; the potential for
bias is unlikely or minimal, and the study used sensitive methodology. High confidence
studies generally reflect judgments of good across all or most evaluation domains.

•	Medium confidence: Possible deficiencies or concerns were identified, but the
limitations are unlikely to have a significant impact on the study results or their
interpretation. Generally, medium confidence studies include adequate or good
judgments across most domains, with the impact of any identified limitation not being
judged as severe.

•	Low confidence: Deficiencies or concerns are identified, and the potential for bias or
inadequate sensitivity is expected to have a significant impact on the study results or
their interpretation. Typically, low confidence studies have a deficient evaluation for one
or more domains, although some medium confidence studies might have a deficient
rating in domain(s) considered to have less influence on the magnitude or direction of
effect estimates. Low confidence results are given less weight compared to high or
medium confidence results during evidence synthesis and integration (see Sections 7
and 8) and are generally not used as the primary sources of information for hazard
identification or derivation of toxicity values unless they are the only studies available
(in which case, this significant uncertainty would be emphasized during dose-response
analysis). Studies rated low confidence only because of sensitivity concerns are
asterisked or otherwise noted because they often require additional consideration
during evidence synthesis. Effects observed in studies that are biased toward the null
may increase confidence in the results, assuming the study is otherwise well conducted
(see Section 8).

•	Uninformative: Serious flaw(s) are judged to make the study results uninterpretable for
use in the assessment. Studies with critically deficient judgments in any evaluation
domain are almost always rated uninformative. Studies with multiple deficient
judgments across domains may also be considered uninformative. Given that the

This document is a draft for review purposes only and does not constitute Agency policy.

6-4	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

findings of interest are considered uninterpretable based on the identified flaws (see
above definition of critically deficient) and do not provide information of use to
assessment interpretations, these studies have no impact on evidence synthesis or
integration judgments and are not usable for dose-response analyses but may be used
to highlight research gaps.

As previously noted, study evaluation determinations reached by each reviewer and the
consensus judgment between reviewers are recorded in HAWC. Final study evaluations housed in
HAWC are made available when the draft is publicly released. The study confidence classifications
and their rationales are carried forward and considered as part of evidence synthesis (see
Section 11) to help interpret the results across studies.

6.2. EPIDEMIOLOGY STUDY EVALUATION

Evaluation of epidemiology studies of health effects to assess risk of bias and study
sensitivity will be conducted for the following domains: exposure measurement, outcome
ascertainment, participant selection, potential confounding, analysis, study sensitivity, and selective
reporting. Bias can result in false positives and negatives (i.e., Types I and II errors), while study
sensitivity is typically concerned with identifying the latter.

The principles and framework used for evaluating epidemiology studies are based on the
Cochrane Risk of Bias in Nonrandomized Studies of Interventions [ROBINS-I; fSterne etal.. 20161]
but modified to address environmental and occupational exposures. Core and prompting questions,
shown in Table 6-1, are used to collect information to guide evaluation of each domain. Core
questions represent key concepts, while the prompting questions help the reviewer focus on
relevant details under each key domain. Table 6-1 also includes criteria that apply to all exposures
and outcomes.

This document is a draft for review purposes only and does not constitute Agency policy.

6-5	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 6-1. Domains, questions, and general considerations to guide the evaluation of epidemiology studies

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes

Exposure
measurement

Does the exposure
measure reliably
distinguish
between levels of
exposure in a time
window

considered most
relevant for a
causal effect with
respect to the
development of
the outcome?

For all:

•	Does the exposure measure
capture the variability in exposure
among the participants,
considering intensity, frequency,
and duration of exposure?

•	Does the exposure measure reflect
a relevant time window? If not, can
the relationship between measures
in this time and the relevant time
window be estimated reliably?

•	Was the exposure measurement
likely to be affected by knowledge
of the outcome?

•	Was the exposure measurement
likely to be affected by the
presence of the outcome

(i.e., reverse causality)?

For case-control studies of occupational
exposures:

•	Is exposure based on a
comprehensive job history
describing tasks, setting, period,
and use of specific materials?

Is the degree of
exposure
misclassifi cation
likely to vary by
exposure level?

If the correlation
between exposure
measurements is
moderate, is there
an adequate
statistical approach
to ameliorate
variability in
measurements?

If potential for bias
is a concern, is the
predicted direction
or distortion of the
bias on the effect
estimate (if there is
enough
information)?

Good

Valid exposure assessment methods used, which represent
the etiologically relevant period of interest.

Exposure misclassification is expected to be minimal.

Adequate

Valid exposure assessment methods used, which represent
the etiologically relevant period of interest.

Exposure misclassification could exist but is not expected to
greatly change the effect estimate.

Deficient

•	Valid exposure assessment methods used, which represent
the etiologically relevant time period of interest. Specific
knowledge about the exposure and outcome raises
concerns about reverse causality, but whether it is
influencing the effect estimate is uncertain.

•	Exposed groups are expected to contain a notable
proportion of unexposed or minimally exposed individuals,
the method did not capture important temporal or spatial
variation, or other evidence of exposure misclassification
would be expected to notably change the effect estimate.

Critically deficient

•	Exposure measurement does not characterize the
etiologically relevant period of exposure or is not valid.

•	Evidence exists that reverse causality is very likely to
account for the observed association.

This document is a draft for review purposes only and does not constitute Agency policy.

6-6	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes



For biomarkers of exposure, general
population:

•	Is a standard assay used? What are
the intra- and inter-assay
coefficients of variation? Is the
assay likely to be affected by
contamination? Are values less
than the limit of detection dealt
with adequately?

•	What exposure period is reflected
by the biomarker? If the half-life is
short, what is the correlation
between serial measurements of
exposure?



• Exposure measurement was not independent of outcome
status.

Outcome
ascertainment
Does the outcome
measure reliably
distinguish the
presence or
absence (or degree
of severity) of the
outcome?

For all:

•	Is outcome ascertainment likely
affected by knowledge, or
presence, of exposure

(e.g., consider access to health
care, if based on self-reported
history of diagnosis)?

For case-control studies:

•	Is the comparison group without
the outcome (e.g., controls in a
case-control study) based on
objective criteria with little or no
likelihood of inclusion of people
with the disease?

For mortality measures:

Is there a concern
that any outcome
misclassification is
nondifferential,
differential, or
both?

What is the
predicted direction
or distortion of the
bias on the effect
estimate (if there is
enough
information)?

Good

•	High certainty in the outcome definition (i.e., specificity and
sensitivity), minimal concerns with respect to
misclassification.

•	Assessment instrument was validated in a population
comparable to the one from which the study group was
selected.

Adequate

•	Moderate confidence that outcome definition was specific
and sensitive, some uncertainty with respect to
misclassification but not expected to greatly change the
effect estimate.

•	Assessment instrument was validated but not necessarily in
a population comparable to the study group.

This document is a draft for review purposes only and does not constitute Agency policy.

6-7	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes



•	How well does cause-of-death data
reflect occurrence of the disease in
an individual? How well do
mortality data reflect incidence of
the disease?

For diagnosis of disease measures:

•	Is the diagnosis based on standard
clinical criteria? If it is based on
self-report of the diagnosis, what is
the validity of this measure?

For laboratory-based measures
(e.g., hormone levels):

•	Is a standard assay used? Does the
assay have an acceptable level of
inter-assay variability? Is the
sensitivity of the assay appropriate
for the outcome measure in this
study population?



Deficient

•	Outcome definition was not specific or sensitive.

•	Uncertainty regarding validity of assessment instrument.
Critically deficient

•	Invalid/insensitive marker of outcome.

•	Outcome ascertainment is very likely to be affected by
knowledge of, or presence of, exposure.

Note: Lack of blinding should not be automatically construed to be
critically deficient.

Participant
selection
Is there evidence
that selection into
or out of the study
(or analysis
sample) was jointly
related to
exposure and to
outcome?

For longitudinal cohort:

•	Did participants volunteer for the
cohort on the basis of knowledge
of exposure or preclinical disease
symptoms? Was entry into, or
continuation in, the cohort related
to exposure and outcome?

For occupational cohort:

•	Did entry into the cohort begin
with the start of the exposure?

Were differences
in participant
enrollment and
follow-up

evaluated to assess
bias?

If potential for bias
is a concern, what
is the predicted
direction or
distortion of the
bias on the effect

Good

•	Minimal concern for selection bias based on description of
recruitment process and follow-up (e.g., selection of
comparison population, population-based random sample
selection, recruitment from sampling frame including
current and previous employees).

•	Exclusion and inclusion criteria specified and would not
induce bias.

•	Participation rate is reported at all steps of study (e.g., initial
enrollment, follow-up, selection into analysis sample). If
rate is not high, appropriate rationale is given for why it is

This document is a draft for review purposes only and does not constitute Agency policy.

6-8	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes



•	Was follow-up or outcome
assessment incomplete, and if so,
was follow-up related to both
exposure and outcome status?

•	Could exposure produce symptoms
that would result in a change in
work assignment/work status
("healthy worker survivor effect")?

For case-control study:

•	Were controls representative of
population and periods from which
cases were drawn?

•	Are hospital controls selected from
a group whose reason for
admission is independent of
exposure?

•	Could recruitment strategies,
eligibility criteria, or participation
rates result in differential
participation relating to both
disease and exposure?

For population-based survey:

•	Was recruitment based on
advertisement to people with
knowledge of exposure, outcome,
and hypothesis?

estimate (if there is

enough

information)?

Were appropriate
analyses
performed to
address changing
exposures over
time relative to
symptoms?

Is there a
comparison of
participants and
nonparticipants to
address whether
differential
selection or study
retention/continua
tion is likely?

unlikely to be related to exposure (e.g., comparison
between participants and nonparticipants or other available
information indicates differential selection is not likely).

Adequate

•	Enough of a description of the recruitment process to be
comfortable that there is no serious risk of bias.

•	Inclusion and exclusion criteria specified and would not
induce bias.

•	Participation rate is incompletely reported but available
information indicates participation is unlikely to be related
to exposure.

Deficient

•	Little information on recruitment process, selection
strategy, sampling framework, and participation OR aspects
of these processes raises the potential for bias (e.g., healthy
worker effect, survivor bias).

Critically deficient

•	Aspects of the processes for recruitment, selection strategy,
sampling framework, or participation result in concern that
selection bias is likely to have had a large impact on effect
estimates (e.g., convenience sample with no information
about recruitment and selection, cases and controls are
recruited from different sources with different likelihood of
exposure, recruitment materials stated outcome of interest
and potential participants are aware of or are concerned
about specific exposures).

Confounding

Is confounding adequately addressed by
considerations in:

If potential for bias
is a concern, what
is the predicted

Good

• Conveys strategy for identifying key confounders, including
co-exposures. This may include a priori biological

This document is a draft for review purposes only and does not constitute Agency policy.

6-9	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes

Is confounding of
the effect of the
exposure likely?

•	Participant selection (matching or
restriction)?

•	Accurate information on potential
confounders and statistical
adjustment procedures?

•	Lack of association between
confounder and outcome, or
confounder and exposure in the
study?

• Information from other sources?

•	Is the assessment of confounders
based on a thoughtful review of
published literature, potential
relationships (e.g., as can be gained
through directed acyclic graphing),
and minimizing potential
overcontrol (e.g., inclusion of a
variable on the pathway between
exposure and outcome)?

direction or
distortion of the
bias on the effect
estimate (if there is
enough
information)?

consideration, published literature, causal diagrams, or
statistical analyses, with the recognition that not all "risk
factors" are confounders.

•	Inclusion of potential confounders in statistical models not
based solely on statistical significance criteria (e.g., p < 0.05
from stepwise regression).

•	Does not include variables in the models likely to be
influential colliders or intermediates on the causal pathway.

•	Key confounders are evaluated appropriately and
considered unlikely sources of substantial confounding. This
often will include:

o Presenting the distribution of potential confounders by
levels of the exposure of interest or the outcomes of
interest (with amount of missing data noted);

o Consideration that potential confounders were rare
among the study population, or were expected to be
poorly correlated with exposure of interest;

o Consideration of the most relevant functional forms of
potential confounders;

o Examination of the potential impact of measurement
error or missing data on confounder adjustment; or

o Presenting a progression of model results with
adjustments for different potential confounders, if
warranted.

Adequate

•	Similar to good but might not have included all key
confounders, or less detail might be available on the
evaluation of confounders (e.g., sub-bullets in good). That

This document is a draft for review purposes only and does not constitute Agency policy.

6-10	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes







residual confounding could explain part of the observed
effect is possible, but concern is minimal.

Deficient

•	Does not include variables in the models shown to be
influential colliders or intermediates on the causal pathway.

•	And any of the following:

o The potential for bias to explain some results is high
based on an inability to rule out residual confounding,
such as a lack of demonstration that key confounders of
the exposure-outcome relationships were considered;

o Descriptive information on key confounders (e.g., their
relationship relative to the outcomes and exposure
levels) are not presented; or

o Strategy of evaluating confounding is unclear or is not
recommended (e.g., only based on statistical significance
criteria or stepwise regression [forward or backward
elimination]).

Critically deficient

•	Includes variables in the models that are colliders or
intermediates in the causal pathway, indicating that
substantial bias is likely from this adjustment; or

•	Confounding is likely present and not accounted for,
indicating that all results were most likely due to bias.

Analysis

Does the analysis
strategy and
presentation
convey the
necessary

• Are missing outcome, exposure,
and covariate data recognized, and
if necessary, accounted for in the
analysis?

If potential for bias
is a concern, what
is the predicted
direction or
distortion of the
bias on the effect

Good

• Use of an optimal characterization of the outcome variable,
including presentation of subgroup- or lifestage-specific
comparisons (as appropriate for the outcome).

This document is a draft for review purposes only and does not constitute Agency policy.

6-11	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes

familiarity with the
data and
assumptions?

•	Does the analysis appropriately
consider variable distributions and
modeling assumptions?

•	Does the analysis appropriately
consider subgroups or lifestages of
interest (e.g., based on variability in
exposure level or duration or
susceptibility)?

•	Is an appropriate analysis used for
the study design?

•	Is effect modification considered,
based on considerations developed
a priori?

•	Does the study include additional
analyses addressing potential
biases or limitations (i.e., sensitivity
analyses)?

estimate (if there is

enough

information)?

•	Quantitative results presented (effect estimates and
confidence limits or variability in estimates) (i.e., not
presented only as a p-value or "significant"/"not
significant").

•	Descriptive information about outcome and exposure
provided (where applicable).

•	Amount of missing data noted and addressed appropriately
(discussion of selection issues—missing at random vs.
differential).

•	Where applicable, for exposure, includes Limit of detection
LOD (and percentage below the LOD), and decision to use
log transformation.

•	Includes analyses that address robustness of findings,
e.g., examination of exposure-response (explicit
consideration of nonlinear possibilities, quadratic, spline, or
threshold/ceiling effects included, when feasible); relevant
sensitivity analyses; effect modification examined based
only on a priori rationale with sufficient numbers.

•	No deficiencies in analysis evident. Discussion of some
details might be absent (e.g., examination of outliers).

Adequate

•	Same as 'Good/ except:

•	Descriptive information about exposure provided (where
applicable) but might be incomplete; might not have
discussed missing data, cut-points, or shape of
distribution(s).

•	Includes analyses that address robustness of findings
(examples in 'Good'), but some important analyses are not
performed.

This document is a draft for review purposes only and does not constitute Agency policy.

6-12	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes







Deficient

•	Does not conduct analysis using optimal characterization of
the outcome variable.

•	Descriptive information about exposure levels not provided
(where applicable).

•	Effect estimate and p-value presented, without standard
error or confidence interval.

•	Results presented as statistically "significant"/"not
significant."

Critically deficient

•	Analysis methods are not appropriate for design or data of
the study.

Selective reporting
Is there reason to
be concerned
about selective
reporting?

•	Were results provided for all the
primary analyses described in the
methods section?

•	Is appropriate justification given for
restricting the amount and type of
results shown?

•	Are only statistically significant
results presented?

If potential for bias
is a concern, what
is the predicted
direction or
distortion of the
bias on the effect
estimate (if there is
enough
information)?

Good

•	The results reported by study authors are consistent with
the primary and secondary analyses described in a
registered protocol or methods paper.

Adequate

•	The authors described their primary (and secondary)
analyses in the methods section and results were reported
for all primary analyses.

Deficient

•	Concerns were raised based on previous publications, a
methods paper, or a registered protocol indicating that
analyses were planned or conducted that were not
reported, or that hypotheses originally considered to be
secondary were represented as primary in the reviewed
paper.

This document is a draft for review purposes only and does not constitute Agency policy.

6-13	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

Follow-up
questions

Criteria that apply to most exposures and outcomes







•	Only subgroup analyses were reported, suggesting that
results for the entire group were omitted.

•	Only statistically significant results were reported.

Sensitivity
Is there a concern
that sensitivity of
the study is not
adequate to detect
an effect?

•	Is the exposure contrast adequate
to detect associations and
exposure-response relationships?

•	Was the appropriate population or
lifestage included?

•	Was the length of follow-up
adequate? Is the time/age of
outcome ascertainment optimal
given the interval of exposure and
the health outcome?

•	Do other aspects related to risk of
bias or otherwise raise concerns
about sensitivity?



Good

•	There is sufficient variability/contrast in exposure to
evaluate primary hypotheses.

•	The study population was sensitive to the development of
the outcomes of interest (e.g., ages, lifestage, sex).

•	The timing of outcome ascertainment was appropriate given
expected latency for outcome development (i.e., adequate
follow-up interval).

•	The study was adequately powered to observe an effect.

•	No other concerns raised regarding study sensitivity.

Adequate

Same considerations as Good, except:

•	There may be issues identified that could reduce sensitivity,
but they are considered unlikely to substantially impact the
overall findings of the study.

Deficient

•	Concerns were raised about the considerations described
for Good that are expected to notably decrease the
sensitivity of the study to detect associations for the
outcome.

Critically deficient

•	Severe concerns were raised about the sensitivity of the
study such that any observed associations are likely to be
explained by bias.

This document is a draft for review purposes only and does not constitute Agency policy.

6-14	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

6.3. EXPERIMENTAL ANIMAL STUDY EVALUATION

1	Using the principles described in Section 6.1, the animal studies of health effects are

2	evaluated for the following domains to assess risk of bias and sensitivity: allocation, observational

3	bias/blinding, confounding, selective reporting, attrition, chemical administration and

4	characterization, endpoint measurement and validity, results presentation and comparisons, and

5	sensitivity (see Table 6-2).

6	The rationale for judgments is documented at the outcome level. The evaluation

7	documentation in HAWC includes the identified limitations and their expected impact on the overall

8	confidence level. To the extent possible, the rationale will reflect an interpretation of the potential

9	influence on the outcome-specific results, including the direction or magnitude of influence
10	(or both).

This document is a draft for review purposes only and does not constitute Agency policy.

6-15	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 6-2. Domains, questions, and general considerations to guide the evaluation of animal toxicology studies

Domain and core
question

Prompting questions

General considerations

Allocation

Were animals assigned to
experimental groups using a
method that minimizes
selection bias?

For each study:

Did each animal or litter have an equal
chance of being assigned to any
experimental group (i.e., random
allocation)?3

Is the allocation method described?
Aside from randomization, were any
steps taken to balance variables across
experimental groups during allocation?

These considerations typically do not need to be refined by assessment teams.
A judgment and rationale for this domain should be given for each cohort or
experiment in the study.

Good: Experimental groups were randomized, and any specific randomization
procedure was described or inferable (e.g., computer-generated scheme. Note that
normalization is not the same as randomization [see response for adequate]).
Adequate: Authors report that groups were randomized but do not describe the
specific procedure used (e.g.," animals were randomized"). Alternatively, authors
used a nonrandom method to control for important modifying factors across
experimental groups (e.g., body-weight normalization).

Not reported (interpreted as deficient): No indication of randomization of groups or
other methods (e.g., normalization) to control for important modifying factors across
experimental groups.

Critically deficient: Bias in the animal allocations was reported or inferable.

Observational bias/blinding

Did the study implement
measures to reduce
observational bias?

For each endpoint/outcome or grouping
of endpoints/outcomes in a study:

Does the study report blinding or other
procedures for reducing observational
bias?

If not, did the study use a design or
approach for which such procedures can
be inferred?

What is the expected impact of failure
to implement (or report
implementation) of these procedures on
results?

These considerations typically do not need to be refined by the assessment teams.
(Note that it can be useful for teams to identify highly subjective measures of
endpoints/outcomes when observational bias may strongly influence results prior to
performing evaluations.)

A judgment and rationale for this domain should be given for each endpoint/outcome
or group of endpoints/outcomes investigated in the study.

Good: Measures to reduce observational bias were described (e.g., blinding to
conceal treatment groups during endpoint evaluation; consensus-based evaluations
of histopathology-lesions).b

Adequate: Methods for reducing observational bias (e.g., blinding) can be inferred or
were reported but described incompletely.

Not reported: Measures to reduce observational bias were not described.
(Interpreted as adequate) The potential concern for bias was mitigated based on use
of automated/computer driven systems, standard laboratory kits, relatively simple,
objective measures (e.g., body or tissue weight), or screening-level evaluations of
histopathology.

(Interpreted as deficient) The potential impact on the results is major (e.g., outcome
measures are highly subjective).

This document is a draft for review purposes only and does not constitute Agency policy.

6-16	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





Critically deficient: Strong evidence for observational bias that impacted the results.

Confounding

Are variables with the
potential to confound or
modify results controlled
for and consistent across
experimental groups?

Note:

Consideration of overt
toxicity (possibly masking
more specific effects) is
addressed under endpoint
measurement reliability.

For each study:

Are there difference across the
treatment groups, considering both
differences related to the exposure
(e.g., co-exposures, vehicle, diet,
palatability) and other aspects of the
study design or animal groups (e.g.,
animal source, husbandry, or health
status), that could bias the results?
If differences are identified, to what
extent are they expected, based on a
specific scientific understanding, to
impact the results?

These considerations may need to be refined by assessment teams, as the specific
variables of concern can vary by experiment or chemical.

A judgment and rationale for this domain should be given for each cohort or
experiment in the study, noting when the potential for confounding is restricted to
specific endpoints/outcomes.

Good: Outside of the exposure of interest, variables that are likely to confound or
modify results appear to be controlled for and consistent across experimental groups.
Adequate: Some concern that variables that were likely to confound or modify results
were uncontrolled or inconsistent across groups but are expected to have a minimal
impact on the results.

Deficient: Notable concern that potentially confounding variables were uncontrolled
or inconsistent across groups and are expected based on to substantially impact the
results.

Critically deficient: Confounding variables were presumed to be uncontrolled or
inconsistent across groups and are expected to be a primary driver of the results.

Attrition

Did the study report results
for all tested animals?

For each study:

Are all animals accounted for in the
results?

If there is attrition, do authors provide
an explanation (e.g., death or
unscheduled sacrifice during the study)?
If unexplained attrition of animals for
outcome assessment is identified, what
is the expected impact on the
interpretation of the results?

These considerations typically do not need to be refined by assessment teams.
A judgment and rationale for this domain should be given for each cohort or
experiment in the study.

Good: Results were reported for all animals. If animal attrition is identified, the
authors provide an explanation, and these are not expected to impact the
interpretation of the results.

Adequate: Results are reported for most animals. Attrition is not explained but this is
not expected to significantly impact the interpretation of the results.

Deficient: Moderate-to-high level of animal attrition that is not explained and may
significantly impact the interpretation of the results.

Critically deficient: Extensive animal attrition that prevents comparisons of results
across treatment groups.

Chemical administration
and characterization

Did the study adequately
characterize exposure to
the chemical of interest and

For each study:

Are there concerns [specific to this
chemical] regarding the source and
purity and/or composition (e.g., identity

It is essential that these considerations are considered, and potentially refined, by
assessment teams, as the specific variables of concern can vary by chemical
(e.g., stability may be an issue for one chemical but not another).

A judgment and rationale for this domain should be given for each cohort or
experiment in the study.

This document is a draft for review purposes only and does not constitute Agency policy.

6-17	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations

the exposure administration

methods?

Note:

Consideration of the
appropriateness of the
route of exposure (not the
administration method) is
not a risk of bias
consideration. Relevance
and utility of the routes of
exposure are considered in
the PECO criteria for study
inclusion and during
evidence synthesis.
Relatedly, consideration of
exposure level selection
(e.g., were levels sufficiently
high to elicit effects) is
addressed during evidence
synthesis and is not a risk of
bias consideration.

and percent distribution of different
isomers) of the chemical?
Was independent analytical verification
of the test article (e.g., composition,
homogeneity, and purity) performed?
Were nominal exposure levels verified
analytically? Are there concerns about
the methods used to administer the
chemical (e.g., inhalation chamber type,
gavage volume)?

Good: Chemical administration and characterization is complete (i.e., source and
purity are provided or can be obtained from the supplier and test article is analytically
verified). There are no notable concerns about the composition, stability, or purity of
the administered chemical, or the specific methods of administration. Exposure levels
are verified using reliable analytical methods.

Adequate: Some uncertainties in the chemical administration and characterization
are identified but these are expected to have minimal impact on interpretation of the
results (e.g., purity of the test article is suboptimal but interpreted as unlikely to have
a significant impact; analytical verification of exposure levels is not reported or
verified with non-preferred methods).

Deficient: Uncertainties in the exposure characterization are identified and expected
to substantially impact the results (e.g., source of the test article is not reported, and
composition is not independently verified; impurities are substantial or concerning;
administration methods are considered likely to introduce confounders, such as use
of static inhalation chambers or a gavage volume considered too large for the species
or lifestage at exposure).

Critically deficient: Uncertainties in the exposure characterization are identified and
there is reasonable certainty that the study results are largely attributable to factors
other than exposure to the chemical of interest (e.g., identified impurities are
expected to be a primary driver of the results).

Endpoint measurement

Are the selected
procedures, protocols and
animal models adequately
described and appropriate
for the

endpoint(s)/outcome(s) of

interest?

Notes:

Considerations related to
the sensitivity of the animal
model and timing of
endpoint measurement are

For each endpoint/outcome or grouping
of endpoints/outcomes in a study:
Are the evaluation methods and animal
model adequately described and
appropriate?

Are there concerns regarding the
methodology selected for endpoint
evaluation?

Are there concerns about the specificity
of the experimental design?

Are there serious concerns regarding
the sample size or how endpoints were
sampled?

Considerations for this domain are highly variable depending on the
endpoint(s)/outcome(s) of interest and typically must be refined by assessment
teams.

A judgment and rationale for this domain should be given for each endpoint/outcome
or group of endpoints/outcomes investigated in the study.

Some considerations include the following:

Good:

•	Adequate description of methods and animal models.

•	Use of generally accepted and reliable endpoint methods.

This document is a draft for review purposes only and does not constitute Agency policy.

6-18	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations

evaluated under Sensitivity
Considerations related to
adjustments/corrections to
endpoint measurements
(e.g., organ weight
corrected for body weight)
are addressed under results
presentation.

Are appropriate control groups for the
study/assay type included?

*	Sample sizes are generally considered adequate for the assay or protocol of
interest and there are no notable concerns about sampling in the context of
the endpoint protocol (e.g., sampling procedures for histological analysis).

*	Includes appropriate control groups and any use of nonconcurrent or
historical control data (e.g., for evaluation of rare tumors) is justified (e.g.,
authors or evaluators considered the similarity between current
experimental animals and laboratory conditions to historical controls).

Ratings of Adequate, Deficient, and Critically Deficient are generally defined as
follows:

Adequate: Issues are identified that may affect endpoint measurement but are
considered unlikely to substantially impact the overall findings or the ability to
reliably interpret those findings.

Deficient: Concerns are raised that are expected to notably affect endpoint
measurement and reduce the reliability of the study findings

Critically deficient: Severe concerns are raised about endpoint measurement and any
findings are likely to be largely explained by these limitations

• The following specific examples of relevant concerns are typically associated
with a Deficient rating, but Adequate or Critically Deficient might be applied
depending on the expected impact of limitations on the reliability and interpretation
of the results:

*	Study report lacks important details that are necessary to evaluate the
appropriateness of the study design (e.g., description of the assays or
protocols; information on the strain, sex, or lifestage of the animals).

*	Selection of protocols that are nonpreferred or lack specificity for
investigating the endpoint of interest. This includes omission of additional
experimental criteria (e.g., inclusion of a positive control or dosing up to
levels causing minimal toxicity) when required by specific testing
guidelines/protocols.3

*	Overt toxicity (e.g., mortality, extreme weight loss) is observed or expected
based on findings from similarly designed studies and may mask
interpretation of outcome(s) of interest.

This document is a draft for review purposes only and does not constitute Agency policy.

6-19	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





•	Sample sizes are smaller than is generally considered adequate for the assay
or protocol of interest. Inadequate sampling can also be raised within the
context of the endpoint protocol (e.g., in a pathology study, bias that is
introduced by only sampling a single tissue depth or an inadequate number
of slides per animal).b

•	Control groups are not included, considered inappropriate, or comparisons
to non-concurrent or historical controls are not adequately justified.

Results presentation

Are the results presented
and compared in a way that
is appropriate and
transparent?

For each endpoint/outcome or grouping
of endpoints/outcomes in a study:

Does the level of detail allow for an
informed interpretation of the results?
Are the data compared, or presented, in
a way that is inappropriate or
misleading?

Considerations for this domain are highly variable depending on the outcomes of
interest and typically must be refined by assessment teams.

A judgment and rationale for this domain should be given for each endpoint/outcome
or group of endpoints/outcomes investigated in the study.

Some considerations include the following:

Good:

•	No concerns with how the data are presented.

•	Results are quantified or otherwise presented in a manner that allows for an
independent consideration of the data (assessments do not rely on author
interpretations).

•	No concerns with completeness of the results reporting.0

•	Ratings of Adequate, Deficient, and Critically Deficient are generally defined
as follows:

Adequate: Concerns are identified that may affect results presentation but are
considered unlikely to substantially impact the overall findings or the ability to
reliably interpret those findings.

Deficient: Concerns with results presentation are identified and expected to
substantially impact results interpretation and reduce the reliability of the study
findings.

Critically deficient: Severe concerns about results presentation were identified and
study findings are likely to be largely explained by these limitations.

•	The following specific examples of relevant concerns are typically associated
with a Deficient rating but Adequate or Critically Deficient might be applied

This document is a draft for review purposes only and does not constitute Agency policy.

6-20	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





depending on expected impact of limitations on the reliability and interpretation of
the results:

•	Nonpreferred presentation of data (e.g., developmental toxicity data
averaged across pups in a treatment group, when litter responses are more
appropriate; presentation of only absolute organ weight data when relative
weights are more appropriate).

•	Pooling data when responses are known or expected to differ substantially
(e.g., across sexes or ages).

•	Incomplete presentation of the data0 (e.g., presentation of mean without
variance data; concurrent control data are not presented; dichotomizing or
truncating continuous data).

Selective reporting

Did the study report results
for all prespecified
outcomes?

Note:

This domain does not
consider the
appropriateness of the
analysis/results
presentation. This aspect of
study quality is evaluated in
another domain.

For each study:

Are results presented for all
endpoints/outcomes described in the
methods (see note)?

If unexplained results omissions are
identified, what is the expected impact
on the interpretation of the results?

These considerations typically do not need to be refined by assessment teams.
A judgment and rationale for this domain should be given for each cohort or
experiment in the study.

Good: Quantitative or qualitative results were reported for all prespecified outcomes
(explicitly stated or inferred), exposure groups and evaluation time points. Data not
reported in the primary article is available from supplemental material. If results
omissions are identified, the authors provide an explanation, and these are not
expected to impact the interpretation of the results.

Adequate: Quantitative or qualitative results are reported for most prespecified
outcomes (explicitly stated or inferred) and evaluation time points. Omissions and are
not explained but are not expected to significantly impact the interpretation of the
results.

Deficient: Quantitative or qualitative results are missing for many prespecified
outcomes (explicitly stated or inferred), omissions are not explained and may
significantly impact the interpretation of the results.

Critically deficient: Extensive results omission is identified and prevents comparisons
of results across treatment groups.

Sensitivity

Are there concerns that
sensitivity in the study is not

Was the exposure period, timing (e.g.,
lifestage), frequency, and duration
sensitive for the outcome(s) of interest?

These considerations may require customization to the specific exposure and
outcomes. Some study design features that affect study sensitivity may have already
been included in the other evaluation domains; these should be noted in this domain,

This document is a draft for review purposes only and does not constitute Agency policy.

6-21	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations

adequate to detect an

effect?

Note:

Consideration of exposure
level selection (e.g., were
levels sufficiently high to
elicit effects) is addressed
during evidence synthesis
and is not a study sensitivity
consideration.

Based on knowledge of the health
hazard of concern, did the selection of
species, strain, and/or sex of the animal
model reduce study sensitivity?
Are there concerns regarding the timing
(e.g., lifestage) of the outcome
evaluation?

Are there aspects related to risk of bias
domains that raise concerns about
insensitivity (e.g., selection of protocols
that are known to be insensitive or
nonspecific for the outcome(s) of
interest)

along with any features that have not been addressed elsewhere. Some

considerations include:

Good

•	The experimental design (considering exposure period, timing, frequency,
and duration) is appropriate and sensitive for evaluating the outcome(s) of
interest.

•	The selected animal model (considering species, strain, sex, and/or lifestage)
is known or assumed to be appropriate and sensitive for evaluating the
outcome(s) of interest.

•	No significant concerns with the ability of the experimental design to detect
the specific outcome(s) of interest, (e.g., outcomes evaluated at the
appropriate lifestage; study designed to address known endpoint variability
that is unrelated to treatment, such as estrous cyclicity or time of day).

•	Timing of endpoint measurement in relation to the chemical exposure is
appropriate and sensitive (e.g., behavioral testing is not performed during a
transient period of test chemical-induced depressant or irritant effects;
endpoint testing does not occur only after a prolonged period, such as weeks
or months, of nonexposure).

•	Potential sources of bias toward the null are not a substantial concern.
Adequate

Same considerations as Good, except:

•	The duration and frequency of the exposure was appropriate, and the
exposure covered most of the critical window (if known) for the outcome(s)
of interest.

•	Potential issues are identified that could reduce sensitivity, but they are
unlikely to impact the overall findings of the study.

This document is a draft for review purposes only and does not constitute Agency policy.

6-22	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





Deficient

•	Concerns were raised about the considerations described for Good or
Adequate that are expected to notably decrease the sensitivity of the study
to detect a response in the exposed group(s).

Critically deficient

•	Severe concerns were raised about the sensitivity of the study and
experimental design such that any observed associations are likely to be
explained by bias. The rationale should indicate the specific concern(s).

Overall confidence

Considering the identified
strengths and limitations,
what is the overall
confidence rating for the
endpoint(s)/outcome(s) of
interest?

For each endpoint/outcome or grouping
of endpoints/outcomes in a study:

Were concerns (i.e., limitations or
uncertainties) related to the risk of bias
or sensitivity identified?

If yes, what is their expected impact on
the overall interpretation of the
reliability and validity of the study
results, including (when possible)
interpretations of impacts on the
magnitude or direction of the reported
effects?

The overall confidence rating considers the likely impact of the noted concerns
(i.e., limitations or uncertainties) in reporting, bias, and sensitivity on the results.
Reviewers should mark studies that are rated lower than high confidence only due to
low sensitivity (i.e., bias toward the null) for additional consideration during evidence
synthesis. If the study is otherwise well conducted and an effect is observed, it may
increase the strength of evidence judgment.

A confidence rating and rationale should be given for each endpoint/outcome or
group of endpoints/outcomes investigated in the study. Confidence ratings are
described above (see Section 6.1).

aThese limitations typically also raise a concern for insensitivity.

bSample size alone is not a reason to conclude an individual study is critically deficient.

cFailure to describe any findings for assessed outcomes (i.e., report lacks any qualitative or quantitative description of the results in tables, figures, or text) is
addressed under Selective Reporting.

This document is a draft for review purposes only and does not constitute Agency policy.

6-23	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

6.4.	CONTROLLED HUMAN EXPOSURE STUDY EVALUATION

Controlled human health studies are seldom available for IRIS assessments. In the case of
nitrate/nitrite there is a substantial body of literature evaluating potential beneficial health effects
(namely, cardiovascular benefits) of controlled exposure to nitrate/nitrite, but no controlled human
exposure studies evaluated risk of adverse health effects. However, if any such studies are identified
during literature search updates, evaluation criteria will be developed incorporating aspects of the
approaches used for epidemiology studies and experimental animal studies, as well as the Cochrane
risk of bias tools for randomized trials (R0B2) (Sterne etal.. 20191 and the ROBINS-I tool discussed
in Section 6.2 f Sterne etal.. 20161. Controlled human exposure studies will be evaluated for
important attributes of experimental studies, including randomization of exposure assignments,
blinding of subjects and investigators, exposure generation, inclusion of a clean air control
exposure (if applicable), outcome ascertainment, missing data, deviations from the intended
intervention, study sensitivity, and other aspects of the exposure protocol. Evaluation will also
include confirmation that the study protocol was approved by an institutional review board.

6.5.	IN VITRO AND OTHER MECHANISTIC STUDY EVALUATION

Mechanistic studies will be evaluated using the considerations presented in Table 6-3 for
the following domains: risk of bias (observational bias/blinding, variable control, specificity,
selective reporting, chemical administration and characterization, endpoint measurement validity,
and results presentation and comparisons) and study sensitivity. Mechanistically relevant
endpoints reported in human and in vivo animal studies are evaluated using the domains for
epidemiology and experimental animal studies presented in the previous sections. Assay-specific
considerations are applied when evaluating the sensitivity domains and will be recorded in their
evaluations in the HAWC database.

This document is a draft for review purposes only and does not constitute Agency policy.

6-24	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 6-3. Domains, questions, and general considerations to guide the evaluation of in vitro studies

Domain and core
question

Prompting questions

General considerations

Observational bias/blinding

Did the study implement
measures, where possible,
to reduce observational
bias?

Considerations will vary
depending on the specific
assay/model system being
used and may not be
applicable to some analyses.

For each assay or endpoint in a study:

Did the study report steps taken to minimize
observational bias during analysis (e.g.,
blinding/coding of slides or plates for analysis;
collection of data from randomly selected
fields; positive controls that are not
immediately identifiable)?

If not, did the study use a design or approach
for which such procedures can be inferred, or
which would not be possible to implement?
Were the assays evaluated using automated
approaches (e.g., microplate readers) that
reduce concern for observational bias?

What is the expected impact of failure to
implement (or report implementation) of these
methods/procedures on results?

These considerations typically do not need to be refined by the assessment
teams. Prior to performing evaluations, teams should consider the specific
assay to identify highly subjective measures of endpoints where observational
bias may strongly influence results.

A judgment and rationale for this domain should be given for each assay or
endpoint or group of endpoints investigated in the study.

Good: Measures to reduce observational bias were described (e.g., specific
mention of blinding and/or coding of slides for analysis), or observational bias
is not a concern because of use of automated/computer driven systems
and/or standard laboratory kits.

Not reported, interpreted as adequate: Measures to reduce observational
bias were not described, but the potential concern for bias was mitigated
because protocol cited includes a description of requirements for
blinding/coding, or the impact on results is expected to be minor because the
specific measurement is more objective.

Not reported, interpreted as deficient: No protocol cited; the potential
impact on the results is major because the endpoint measures are highly
subjective (e.g., counting plaques or live vs. dead cells).

Critically deficient: Strong evidence for observational bias that could have
impacted the results.

Variable control

Are all introduced variables
with the potential to affect
the results of interest
controlled for and consistent
across experimental groups?

For each study:

Are there any known or presumed differences
across treatment groups (e.g., co-exposures,
culture conditions, cell passages, variations in
reagent production lots, mycoplasma
infections) that could bias the results? If
differences are identified, to what extent are
they expected to impact the results?

Did the study address features inherent to the
physico-chemical properties of the test
substance(s) that have the potential to bias the

These considerations will need to be refined by assessment teams as the
specific variables of concern can vary by the experimental test system and
chemical.

A judgment and rationale for this domain should be given for each
experiment in the study, noting when the potential to affect results is
restricted to specific assays or endpoints.

Good: Outside of the exposure of interest, variables or features of the test
system and/or chemical properties that are likely to impact results appear to
be controlled for and consistent across experimental groups.

Adequate: Some concern that variables or features of the test system and/or
chemical properties that are likely to modify or interfere with results were

This document is a draft for review purposes only and does not constitute Agency policy.

6-25	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations



results away from the null? For example, could
the test article interfere with a given assay
(e.g., auto-fluoresces or inhibits enzymatic
processes necessary for assay signals),
potentially leading to an erroneous positive
signal? (Note that concerns related to dose are
addressed in chemical administration and
ch aracterizati on.)

Are there known variations in cellular signaling
unique to the model system that could
influence the possibility of detecting the
effect(s) of interest?

Are there concerns regarding the negative
(untreated and/or vehicle) controls used? Were
negative controls run concurrently?

uncontrolled or inconsistent across groups but are expected to have a
minimal impact on the results.

Deficient: Notable concern that important study variables and/or features of
the test system lacked specificity or were uncontrolled or inconsistent across
groups and are expected to substantially impact the results.

Critically deficient: Features of the test system are known to be nonspecific
for this endpoint, and/or influential study variables were presumed to be
uncontrolled or inconsistent across groups and are expected to be a primary
driver of the results.

Selective reporting

Did the study present
results, quantitatively or
qualitatively, for all
prespecified assays or
endpoints and replicates
described in the methods?
Note: The appropriateness
of the analysis or results
presentation is considered
under results presentation.

For each study:

Are results presented for all
endpoints/outcomes described in the
methods?

Did the study clearly indicate the number of
replicate experiments performed? Were the
replicates technical (from the same sample) or
independent (from separate, distinct
exposures)?

If unexplained results omissions are identified,
what is the expected impact on the
interpretation of the results?

These considerations typically do not need to be refined by assessment
teams.

A judgment and rationale for this domain should be given for each assay or
endpoint in the study.

Good: Quantitative or qualitative results were reported for all prespecified
assays or endpoints (explicitly stated or inferred), exposure groups and
evaluation timepoints. Data not reported in the primary article is available
from supplemental material. If results omissions are identified, the authors
provide an explanation, and these are not expected to impact the
interpretation of the results.

Adequate: Quantitative or qualitative results are reported for most
prespecified assays or endpoints (explicitly stated or inferred), exposure
groups and evaluation timepoints. Omissions are not explained but are not
expected to significantly impact the interpretation of the results.

Deficient: Quantitative or qualitative results are missing for many
prespecified assays or endpoints (explicitly stated or inferred), exposure
groups and evaluation timepoints; omissions are not explained and may
significantly impact the interpretation of the results.

This document is a draft for review purposes only and does not constitute Agency policy.

6-26	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





Critically Deficient: Extensive results omissions are identified, preventing
comparisons of results across treatment groups.

Chemical administration
and characterization

Did the study adequately
characterize exposure to the
chemical of interest and the
exposure administration
methods?

For each study:

Are there concerns regarding the purity and/or
composition (e.g., identity and percent
distribution of different isomers) of the test
material/chemical? If so, can the purity and/or
composition be obtained from the supplier
(e.g., as reported on the website)?
Was independent analytical verification of the
test article purity and composition performed?
If not, is this a significant concern for this
substance?

Are there concerns about the stability of the
test chemical in the vehicle and/or culture
media (e.g., pH, solubility, volatility, adhesion
to plastics) that were not corrected for, leading
to potential bias away from the null (e.g.,
observed precipitate formation at high
concentrations) or toward the null (e.g.,
enclosed chambers not used for testing volatile
chemicals)?

Are there concerns about the preparation or
storage conditions of the test substance?
Are there concerns about the methods used to
administer the chemical?

It is essential that these criteria are considered, and potentially refined, by
assessment teams, as the specific variables of concern can vary by chemical
(e.g., stability may be an issue for one chemical but not another).

A judgment and rationale for this domain should be given for each
experiment in the study.

Good: Chemical administration and characterization is complete (i.e., source,
purity, and analytical verification of the test article are provided). There are
no concerns about the composition, stability, or purity of the administered
chemical, or the specific methods of administration.

Adequate: Some uncertainties in the chemical administration and
characterization are identified but these are expected to have minimal impact
on interpretation of the results (e.g., source and vendor-reported purity are
presented but not independently verified; purity of the test article is
suboptimal but not concerning).

Deficient: Uncertainties in the exposure characterization are identified and
expected to substantially impact the results (e.g., the source and purity of the
test article are not reported, and no independent verification of the test
article was conducted; levels of impurities are substantial or concerning;
deficient administration methods were used).

Critically deficient: Uncertainties in the exposure characterization are
identified and there is reasonable certainty that the results are largely
attributable to factors other than exposure to the chemical of interest (e.g.,
identified impurities are expected to be a primary driver of the results).

This document is a draft for review purposes only and does not constitute Agency policy.

6-27	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations

Endpoint measurement

Are the selected protocols,
procedures, and test
systems adequately
described and appropriate
for evaluating the
endpoint(s) of interest?
Notes:

Considerations related to
adjustments or corrections
to endpoint measurements
are addressed under results
presentation.

Considerations related to the
sensitivity of the animal
model and timing of
endpoint measurement are
evaluated under sensitivity.

For each endpoint or grouping of endpoints in a
study:

Are the evaluation methods and test systems
adequately described and appropriate?
Are there concerns regarding the methodology
selected (e.g., accepted guidelines, established
criteria) for endpoint evaluation?

Are there concerns about the specificity of the
experimental design? Did the study address
features inherent to the test system or
experiment that have the potential to lead to
bias away from the null?

Are there serious concerns about the number
of replicates or sample size in the study?
Are appropriate control groups for the
study/assay type included? Was there a need
for the assay to include specific controls to
reduce potential sources of underlying bias?
Did the test compound induce cytotoxicity
(known, or expected based on other studies of
similar design) to a degree that is expected to
affect interpretation of results?

Considerations for this domain are highly variable depending on the assay or
endpoint(s) of interest and must be refined by assessment teams.

A judgment and rationale for this domain should be given for each assay or
endpoint or group of endpoints investigated in the study.

Some considerations include the following:

Good:

•	Adequate description of methods and test system.

•	Use of generally accepted and reliable endpoint methods that are
consistent with accepted guidelines or established criteria for the
assay(s)/endpoint(s) of interest.

•	Sample sizes are generally considered adequate for the assay or
protocol of interest and there are no notable concerns about
sampling in the context of the endpoint protocol.

•	Includes appropriate control groups (e.g., use of loading controls)
and any use of nonconcurrent or historical control data (e.g., for
comparison to background levels in negative controls) is justified
(e.g., authors or evaluators considered the similarity between
current cell cultures and laboratory conditions to historical controls).

Ratings of Adequate, Deficient, and Critically Deficient are generally defined
as follows:

Adequate: Issues are identified that may affect endpoint measurement but
are considered unlikely to substantially impact the overall findings or the
ability to reliably interpret those findings.

Deficient: Concerns are raised that are expected to notably affect endpoint
measurement and reduce the reliability of the study findings
Critically deficient: Severe concerns are raised about endpoint measurement
and any findings are likely to be largely explained by these limitations.
The following specific examples of relevant concerns are typically associated
with a Deficient rating, but Adequate or Critically Deficient might be applied
depending on the expected impact of limitations on the reliability and
interpretation of the results:

This document is a draft for review purposes only and does not constitute Agency policy.

6-28	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





•	Study report lacks important details that are necessary to evaluate
the appropriateness of the study design (e.g., description of the
assays or protocols; information on the cell line, passage number).

•	Selection of protocols that are nonpreferred or lack specificity for
investigating the endpoint of interest. This includes omission of
additional experimental criteria (e.g., inclusion of a positive control
or dosing up to levels causing minimal toxicity) when required by
specific testing guidelines/protocols.3

•	Cytotoxicity is observed or expected based on findings from similarly
designed studies and may mask interpretation of outcome(s) of
interest.

•	Sample sizes are smaller than is generally considered adequate for
the assay or protocol of interest. Inadequate sampling can also be
raised within the context of the endpoint protocol (e.g., in a
pathology study, bias that is introduced by only sampling a single
tissue depth or an inadequate number of slides per animal).b

•	Controls are not included or considered inappropriate.

Results presentation

Are the results presented
and compared in a way that
is appropriate and
transparent and makes the
data usable?

For each assay/endpoint or grouping of
endpoints in a study:

Does the level of detail allow for an informed
interpretation of the results?

If applicable, was the assay signal normalized to
account for nonbiological differences across
replicates and exposure groups?

Are the data compared or presented in a way
that is inappropriate or misleading (e.g.,
presenting western blot images without
including numerical values for densitometry
analysis, or vice versa)? Flag potentially
inappropriate statistical comparisons for
further review.

Considerations for this domain are highly variable depending on the
endpoints of interest and must be refined by assessment teams.

A judgment and rationale for this domain should be given for each assay or
endpoint or group of endpoints investigated in the study.

Some considerations include the following:

Good:

•	No concerns with how the data are presented.

•	Results are quantified or otherwise presented in a manner that
allows for an independent consideration of the data (assessments do
not rely on author interpretations).

•	No concerns with completeness of the results reporting.0

This document is a draft for review purposes only and does not constitute Agency policy.

6-29	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





•	Ratings of Adequate, Deficient, and Critically Deficient are generally
defined as follows:

Adequate: Concerns are identified that may affect results presentation but
are considered unlikely to substantially impact the overall findings or the
ability to reliably interpret those findings.

Deficient: Concerns with results presentation are identified and expected to
substantially impact results interpretation and reduce the reliability of the
study findings.

Critically deficient: Severe concerns about results presentation were
identified and study findings are likely to be largely explained by these
limitations.

•	The following specific examples of relevant concerns are typically
associated with a Deficient rating but Adequate or Critically Deficient might
be applied depending on expected impact of limitations on the reliability and
interpretation of the results:

•	Nonpreferred presentation of data (e.g., averaging technical
replicates rather than independent replicates).

•	Failure to present quantitative results.

•	Pooling data when responses are known or expected to differ
substantially (e.g., across cell types or passage number).

•	Incomplete presentation of the data0 (e.g., presentation of mean
without variance data; concurrent control data are not presented;
failure to report or address overt cytotoxicity).

Sensitivity

Are there concerns that
sensitivity in the study is not
adequate to detect an
effect?

Was the exposure period, timing (i.e., cell
passage number, insufficient culture maturity
for the adequate expression of mature cell
markers; insufficient treatment and/or
measurement duration for the production of
protein above the level of detection),
frequency, and duration of exposure sensitive
for the assay/model system of interest,

Are there concerns regarding the need for positive controls (e.g., concerns
that the effects of interest may be inhibited or otherwise poorly manifest in
the test system, for example due to differences from in vivo biology)? If used,
was the selected positive test substance (and dose) reasonable and
appropriate and was the intended positive response induced?

Considerations for this domain are highly variable depending on the specific
assay/model system used or endpoint(s) of interest and must be refined by
assessment teams. Some study design features that affect study sensitivity

This document is a draft for review purposes only and does not constitute Agency policy.

6-30	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations



particularly in the absence of a positive
control?

Assay-specific considerations regarding
sensitivity, specificity, and validity of the
selection of the test methods will be described
here (e.g., metabolic competency, antibody
specificity) (some of these external
considerations may have been applied during
prioritization of studies for evaluation). Are
there aspects related to risk of bias domains
that raise concerns about insensitivity (e.g.,
selection of protocols or methods that are
known to be insensitive or nonspecific for the
outcome(s) of interest)?

Are there concerns regarding the need for
positive controls (e.g., concerns that the effects
of interest may be inhibited or otherwise
poorly manifest in the test system, for example
due to differences from in vivo biology)? If
used, was the selected positive test substance
(and dose) reasonable and appropriate and was
the intended positive response induced?

may have already been included in the other evaluation domains; these
should be noted in this domain, along with any features that have not been
addressed elsewhere.

Some considerations include:

Good

•	The experimental design (considering exposure period, timing,
frequency, and duration) is appropriate and sensitive for evaluating
the outcome(s) of interest.

•	The selected test system is appropriate and sensitive for evaluating
the outcome(s) of interest (e.g., cell line/cell type is appropriate and
routinely used for the selected assay).

•	No significant concerns with the ability of the experimental design to
detect the specific outcome(s) of interest, (e.g., study designed to
address known endpoint variability that is unrelated to treatment,
such as doubling time or confluency).

•	Timing of endpoint measurement in relation to the chemical
exposure is appropriate and sensitive (e.g., cultures adequately
express mature cell markers).

•	Potential sources of bias toward the null is not a substantial concern.
Adequate

•	Potential issues are identified related to the considerations described
for Good that could reduce sensitivity, but they are unlikely to
impact the overall findings of the study.

Deficient

•	Concerns were raised about the considerations described for Good
that are expected to notably decrease the sensitivity of the study to
detect a response in the exposed group(s).

This document is a draft for review purposes only and does not constitute Agency policy.

6-31	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Domain and core
question

Prompting questions

General considerations





Critically deficient

• Severe concerns were raised about the sensitivity of the study and
experimental design such that any observed associations are likely to
be explained by bias. The rationale should indicate the specific
concern(s).

Overall confidence

Considering the identified
strengths and limitations,
what is the overall
confidence rating for the
assay(s) or endpoint(s) of
interest?

Note:

Reviewers should mark
studies for additional
consideration during
evidence synthesis if due to
low sensitivity only (i.e., bias
toward the null), these
studies are rated as lower
than high confidence. If the
study is otherwise well
conducted and an effect is
observed, the confidence
may be increased.

For each assay or endpoint or grouping of
endpoints in a study:

•	Were concerns (i.e., limitations or
uncertainties) related to the risk of
bias or sensitivity identified?

•	If yes, what is their expected impact
on the overall interpretation of the
reliability and validity of the study
results, including (when possible)
interpretations of impacts on the
magnitude or direction of the reported
effects?

The overall confidence rating considers the likely impact of the noted
concerns (i.e., limitations or uncertainties) in reporting, bias, and sensitivity
on the results.

A confidence rating and rationale should be given for each assay or endpoint
or group of endpoints investigated in the study. Confidence rating definitions
are described above (see Section 4.1).

aThese limitations typically also raise a concern for insensitivity.

bSample size alone is not a reason to conclude an individual study is critically deficient.

cFailure to describe any findings for assessed outcomes (i.e., report lacks any qualitative or quantitative description of the results in tables, figu res, or text) will
result in a critically deficient rating for the outcome(s) of interest for results presentation; overall completeness of reporting at the study level is addressed
under Selective Reporting.

This document is a draft for review purposes only and does not constitute Agency policy.

6-32	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

6.6. PHARMACOKINETIC MODEL EVALUATION

PBPK (or classical pharmacokinetic [PK]) models should be used in an assessment when a
validated and applicable one exists and no equal or better alternative for dosimetric extrapolation
is available. Any models used should represent current scientific knowledge and accurately
translate the science into computational code in a reproducible, transparent manner. For a specific
target organ/tissue, it may be possible to employ or adapt an existing PBPK model or develop a new
PBPK model or an alternate quantitative approach. Data for PBPK models may come from studies
across various species and may be in vitro or in vivo in design. Specific details for this evaluation
are provided below and in the Umbrella quality assurance project plan (QAPP) for dosimetry and
mechanism-based models fU.S. EPA. 2020bl

6.6.1. Pharmacokinetic (PK)/Physiologically Based Pharmacokinetic (PBPK) Model

Descriptive Summary

PBPK modeling is the preferred approach for calculating a human equivalent dose
according to the hierarchy of approaches outlined in EPA guidelines (U.S. EPA, 2011a). As PK/PBPK
studies had been evaluated in the 2001 EPA oral assessment, a literature search was conducted for
PK/PBPK studies published since 2000. As described in Section 4.2, PK/PBPK studies identified in
our search were tagged as supplemental material.

Following literature searches, a stepwise approach is taken that includes conducting an
initial scoping of the supplemental material studies categorized as PK/PBPK models. Then, an in-
depth full model evaluation is implemented to identify PBPK models that are potentially suitable
for deriving toxicity values for the nitrate/nitrite assessment

The initial scoping process is distinct from the full model evaluation. The scoping process
provides a rapid assessment and communication of the availability, structure, and potential uses of
PBPK/PK models, but is not a full evaluation. Full model evaluation—the complete and thorough
assessment of the quality and utility of a particular model—is conducted if the initial scoping
identifies one or more models that are available and considered appropriate for one or more
applications in the assessment The model evaluation is then conducted for the selected
application(s). As shown below in Table 6-4 for example, key information from identified PBPK
models during the scoping process is summarized in tabular format for further in-depth model
evaluation following the evaluation approaches summarized in Section 6.6.2.

This document is a draft for review purposes only and does not constitute Agency policy.

6-33	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 6-4. Example descriptive summary for a physiologically based
pharmacokinetic (PBPK) model

Study detail

Description/notes

Author

Smith et al. (2003)

Contact email

xxxxx (Semail.com

Contact phone

xxx-xxx-xxxx

Sponsor

N/A

Model summary

Species

Rat



Strain

F433



Sex

Male and female



Life stage

Adult



Exposure routes

Inhalation

Oral

I.V.

Skin



Tissue dosimetry

Blood

Liver

Kidney

Urine

Lung

Model evaluation

Language

ACSL 11.8

Code available

YES

Effort to recreate model

COMPLETE

Code received

YES

Effort to migrate to open software

SIGNIFICANT

Structure evaluated

YES

Math evaluated

YES

Code evaluated

YES. Issue (minor): Incorrect units listed in comments for liver metabolism (line 233).
Issue (major): Mass balance error in stomach compartment.

Available PK data

Urine (cumulative amount excreted) and blood (concentration) time course data for
oral (gavage) and inhalation (6 hr/day for 4 days) exposure. In vitro skin permeation.

6.6.2. Pharmacokinetic (PK)/Physiologically Based Pharmacokinetic (PBPK) Model
Evaluation

1	Once available PBPK models are summarized, the assessment team undertakes model

2	evaluation in accordance with criteria outlined by U.S. EPA f2020bl Judgments on the suitability of

3	a model are separated into two categories: scientific and technical (see Table 6-5). The scientific

4	criteria focus on whether the biology, chemistry, and other information available for chemical

5	MOA(s) are justified (i.e., preferably with citations to support use) and represented by the model

6	structure and equations. The scientific criteria are judged based on information presented in the

7	publication or report that describes the model and do not require evaluation of the computer code.

8	Preliminary technical criteria include the availability of the computer code and completeness of

9	parameter listing and documentation. Studies that meet the preliminary scientific and technical

This document is a draft for review purposes only and does not constitute Agency policy.

6-34	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

criteria are then subjected to an in-depth technical -evaluation, which includes a thorough review
and testing of the computational code. The in-depth technical- and scientific analyses focus on the
accurate implementation of the conceptual model in the computational code, the use of
scientifically supported and biologically consistent parameters in the model, and the
reproducibility of model results reported in journal publications and other documents. This
approach stresses (1) clarity in the documentation of model purpose, structure, and biological
characterization; (2) validation of mathematical descriptions, parameter values, and computer
implementation; and (3) the ability of the model to predict each plausible dose metric such as
nitrate and nitrite concentrations in the blood and the production of relevant metabolites. The in-
depth analysis is used to evaluate the potential value and cost of developing a new model or
substantially revising an existing one. PBPK models adapted, modified, or developed by EPA during
the assessment will undergo peer review, either as a component of the draft assessment or by
publication in a journal article.

In brief, a major strength of a PBPK model is its capacity to provide quantitative
descriptions of ADME of chemicals by accounting for the dynamic but complex relationships among
physiological, biochemical, and metabolic determinants. When describing a published PBPK model,
two components must be evaluated: 1) the underlying biological assumptions and resulting
mathematical equations giving rise to the model structure and 2) the parameterization of these
mathematical equations using experimental pharmacokinetic data (such as concentration vs. time
data). Taken together, these two components of model structure and model parameters constitute a
unique PBPK model. To this end, three PBPK models exist for nitrates/nitrites: (Zeilmaker etal..
20101. (Lin etal.. 20201. and (Coggan etal.. 20211. Of these models, (Zeilmaker etal.. 20101 and (Lin
etal.. 20201 share the same underlying model structure originally introduced in (Zeilmaker etal..
19961 with different in vivo datasets used to parameterize the model structure.

Biotransformation of nitrate to nitrite through gut and salivary bacteria is thought to be a
major source of dietary nitrate toxicity. Therefore, the PBPK model(s) selected for the
nitrate/nitrite assessment should reflect the underlying mechanisms and anatomical location for
this biotransformation and any additional mechanisms of action for specific toxicological endpoints
when estimating relevant dose metrics (U.S. EPA. 20181. For example, nitrite is known to react with
hemoglobin in the blood to form methemoglobin. Inclusion of this mechanism will be important for
linking exposure-response information for effect of nitrite on risk of methemoglobinemia, to
exogenous nitrate exposure.

The available PBPK models aim to describe the pharmacokinetics of nitrate and nitrite
following nitrate absorption in the stomach and biotransformation to nitrite throughout the body.
Briefly, the (Zeilmaker et al.. 19961 model structure assumes exposure only to nitrate. In this model
structure, nitrate is absorbed into a central compartment and secreted into a salivary compartment
where it undergoes conversion to nitrite. Following this conversion, nitrite is absorbed through the
stomach into the central compartment where it reacts with hemoglobin to create methemoglobin.

This document is a draft for review purposes only and does not constitute Agency policy.

6-35	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

The fZeilmaker et al.. 20101 model then parameterizes this model structure using data from human
volunteers to characterize nitrate and nitrite levels in blood and saliva following a known exposure
to nitrate. Comparatively, the fCoggan etal.. 20211 model structure assumes exogenous exposure to
both nitrate and nitrite in which nitrate is transformed to nitrite in the central compartment
through first order kinetics. Using a cohort of elderly volunteers, this model structure is
parameterized using plasma nitrate and nitrite concentrations. Finally, fLin etal.. 20201 uses the
same model structure as fZeilmaker etal.. 20101 and updates the parameters using an additional
human nitrate dataset. Further evaluation of these models will be conducted according to EPA's
Umbrella QAPP for Dosimetry and Mechanism-Based Models (U.S. EPA. 2020bl. It may be that none
of the existing PBPK models adequately fulfills all assessment applications. In this case, a hybrid
model could be created that merges elements from the existing models to achieve this objective if
needed and feasible under the time constraints for the assessment.

Table 6-5. Criteria for evaluation of physiologically based pharmacokinetic
(PBPK) models

Criteria

Example information

Scientific

Biological basis for the model is accurate.

•	Consistent with mechanisms that significantly impact dosimetry.

•	Predicts dose metrics expected to be relevant.

•	Applicable for relevant route(s) of exposure.



Consideration of model fidelity to the biological system strengthens the scientific basis of the
assessment relative to standard exposure-based extrapolation (default) approaches.

•	Can the model describe critical behavior, such as nonlinear kinetics in a relevant dose
range, better than the default (i.e., BW3/4scaling)?

•	Is the available metric a better predictor of risk than the default? (Specifically, model-
based metrics may correlate better than the applied doses with animal/human dose-
response data.) The degree of certainty in model predictions vs. default is also a factor
(e.g., while target tissue metrics are generally considered better than blood
concentration metrics, lack of data to validate tissue predictions when blood data are
available may lead to choosing the latter metric).



Principle of parsimony

• Model complexity or biological scale, including number and parameterization of
(sub)compartments (e.g., tissue or subcellular levels) should be commensurate with
data available to identify parameters.



Model describes existing PK data reasonably well, both in "shape" (matches curvature,
inflection points, peak concentration time, etc.) and quantitatively (e.g., within factor of 2-3).



Model equations are consistent with biochemical understanding and biological plausibility.



Well-documented model code is readily available to the EPA and the public.

This document is a draft for review purposes only and does not constitute Agency policy.

6-36	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Criteria

Example information

Initial technical

A set of published parameters is clearly identified, including origin/derivation.



Parameters do not vary unpredictably with dose (e.g., any dose dependence in absorption
constants is predictable across the dose ranges relevant for animal and human modeling).



Sensitivity and uncertainty analysis have been conducted for relevant exposure levels (local
sensitivity analysis is sufficient, but a global analysis provides more information).

•	If a sensitivity analysis was not conducted, EPA may decide to independently
conduct this additional work before using the model in the assessment.

•	A sound explanation should be provided when sensitivity of the dose metric to model
parameters differs from what is reasonably expected based on experience.

6.6.3. Selection of the Appropriate Dose Metric

The level of confidence in using a pharmacokinetic (PK) or PBPK model depends on its
ability to provide a reliable estimation of dose metrics based on biological plausibility and MOA
considerations. Thus, one needs to take into consideration mechanism(s) relevant to the
endpoint(s) of interest, data availability and uncertainties in estimating that dose metric. For
nitrates and nitrites, hemoglobin is an established target of toxicity, although other toxicities might
exist Existing noncancer reference values for nitrate are derived from its transformation to nitrite
and resulting risk for methemoglobinemia. An existing model for nitrate exposure fZeilmaker etal..
20101 includes the nitrite-dependent transformation of hemoglobin to methemoglobin mechanism
of action. Therefore, the production of methemoglobin from nitrite will serve as the dose metric for
the methemoglobinemia endpoint

Compared to methemoglobin production, it remains less understood what the appropriate
dose metric for other toxicities should be. N-nitrosamines, formed via N-nitrosation, are considered
strong carcinogens. Absent a model predicting the formation of N-nitrosamines from parent
compounds, surrogate dose metrics such as nitrate/nitrite (average) daily area under the curve will
be evaluated for this toxicity. If required, the addition of an N-nitrosamine pathway could be
included in existing models if the appropriate pharmacokinetic data exists.

This document is a draft for review purposes only and does not constitute Agency policy.

6-37	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

7. DATA EXTRACTION OF STUDY METHODS AND
RESULTS

The process of summarizing study methods and results is referred to as data extraction.
Studies that met initial PECO criteria after full-text review are briefly summarized in data extraction
forms available in HAWC. These study summaries are used to create interactive literature inventory
visualizations to display the extent and nature of the available evidence in HAWC.

For experimental animal studies, which are typically studies in rodents, the following
information is captured: chemical form, study type (acute [<24 hours], short term [<7 days], short
term [7-27 days], subchronic [28-90 days], chronic [>90 days3] and developmental, which includes
multigeneration studies), duration of treatment, route, species, strain, sex, dose or concentration
levels tested, dose units, health system and specific endpoints assessed.

For human studies, the following information is summarized in HAWC data extraction
forms: chemical form, population type (e.g., general population-adult, occupational, pregnant
women, infants, and children), study type (e.g., cross-sectional, cohort, case-control), sex, major
route of exposure (if known), description of how exposure was assessed, health system studied, and
specific endpoints assessed.

For epidemiology and animal studies that met the assessment PECO criteria, HAWC is used
for study evaluation and for full extraction of study methods and results. Compared with the
literature inventory, full data extraction in HAWC includes summarizing more details of study
design and gathering effect size information. Instructions on how to conduct data extraction in
HAWC are available at https: //hawcproiect.org/resources/. An additional resource used to
implement use of a consistent vocabulary to summarize endpoints assessed in animal studies is
available in HAWC (the Environmental Health Vocabulary; https://hawc.epa.gov/vocab/ehv/).

In some cases, EPA may conduct their own statistical analysis of human and animal
toxicology data (assuming the data are amenable to doing so and the study is otherwise well
conducted) during evidence synthesis.

Data extraction for in vivo and in vitro studies prioritized to assess mechanisms of
nitrate/nitrite is conducted in Microsoft Word and presented in tabular format.

All findings are considered for extraction, regardless of statistical significance. The level of
extraction for specific outcomes within a study could differ (i.e., narrative only if the finding was

3EPA considers chronic exposure to be more than approximately 10% of the life span in humans. For typical
laboratory rodent species, this can lead to consideration of exposure durations of approximately 90 days to 2
years. However, studies in duration of 1-2 years are typical of what is considered representative of chronic
exposure rather than durations just over 90 days.

This document is a draft for review purposes only and does not constitute Agency policy.

7-1	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

23

24

25

26

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

qualitative). For quality control, studies were summarized by one member of the evaluation team
and independently verified by at least one other member. Discrepancies were resolved by
discussion or consultation within the evaluation team. Data extraction results are presented via
figures, tables, or interactive web-based graphics in the assessment The information is also made
available for download in Excel format when the draft is publicly released. The literature
inventories are presented in the HAWC Visualization module, with options to link to the native
Tableau application where the underlying information is available for download. Download of full
data extraction for animal studies is done directly in HAWC.

For non-English studies online translation tools (e.g., Google translator) or engagement with
a native speaker can be used to summarize studies at the level of the literature inventory. Fee-based
translation services for non-English studies are typically reserved for studies considered potentially
informative for dose response, a consideration that occurs after preparation of the initial literature
inventory during draft assessment development. Digital rulers, such as WebPlotDigitizer
fhttp://arohatgi.info/WebPlotDigitizer/). are used to extract numerical information from figures,
and their use is be documented during extraction. For studies that evaluate endpoints at multiple
time points (e.g., 7 days, 3 weeks, 3 months) data are generally summarized for the longest duration
in the study report, but other durations may be summarized if they provide important contextual
information for hazard characterization (e.g., an effect was present at an interim time point but did
not appear to persist or the magnitude of the effect diminished). A free text field is available in
HAWC to describe cases when the approach for summarizing results requires explanation.

Author queries may be conducted for studies considered for hazard identification or dose-
response to facilitate study evaluation and quantitative analysis (e.g., information on variability or
availability of individual animal data). Outreach to study authors or designated contact persons is
documented and considered unsuccessful if researchers do not respond to email or phone requests
within 1 month of initial attempt(s) to contact. Only information or data that can be made publicly
available (e.g., within HAWC or HERO) will be considered.

This document is a draft for review purposes only and does not constitute Agency policy.

7-2	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

8.EVIDENCE SYNTHESIS AND INTEGRATION

Within-stream evidence synthesis is conducted separately for human, animal, and
mechanistic evidence to directly inform the integration across the streams of evidence and draw
overall conclusions for each of the assessed human health effects. The phrases "evidence synthesis"
and "evidence integration" used here are analogous to the phrases "strength of evidence" and
"weight of evidence," respectively, used in some other assessment processes fEFSA. 2017: U.S. EPA.
2017: NRC. 2014: U.S. EPA. 2005al A structured framework approach is used to guide both
evidence synthesis and integration. This structured framework includes consideration of
mechanistic information during both evidence synthesis and integration, although the focus of the
analysis differs. Similarly other types of supplemental information (e.g., ADME, non-PECO route of
exposure) can also inform evidence synthesis and integration analyses.

•	Evidence synthesis: Judgment(s) regarding the strength of the evidence for hazard for each
unit of analysis from the available human and animal studies are made in parallel, but
separately. These judgments can incorporate PK, mechanistic, and other supplemental
evidence when the unit of analysis is defined as such (see Section 5.2). The units of analysis
can also include or be framed to focus on precursor events (e.g., biomarkers). In addition,
this includes an evaluation of coherence across units of analysis within an evidence stream.
At this stage, the animal evidence judgment(s) does not yet consider the human relevance of
that evidence.

•	Evidence integration: The animal and human evidence judgments are combined to draw an
overall evidence integration judgment(s) that incorporates inferences drawn based on
information on the human relevance of the animal evidence, coherence across evidence
streams, potential susceptibility, and other critical inferences (e.g., biological plausibility)
informed by mechanistic, ADME, or other supplemental data.

Evidence synthesis and integration judgments are expressed both narratively in the
assessment and summarized in tabular format in evidence profile tables (see Table 8-1). Key
findings and analyses of mechanistic and other supplemental content are also summarized in
narrative and tabular format to inform evidence synthesis and integration judgments (see
Table 8-2). In brief, a synthesis (strength of evidence) judgment is drawn for each unit of analysis
summarized as robust, moderate, slight, indeterminate, or compelling evidence of no effect (see
Section 8.1). Next, evidence synthesis judgments are used to inform evidence integration (weight of
evidence) judgments summarized as evidence demonstrates; evidence indicates; evidence suggests;
evidence inadequate, or strong evidence supports no effect) (see Section 8.2). These summary
judgments are included as part of the evidence synthesis and integration narratives. When multiple

This document is a draft for review purposes only and does not constitute Agency policy.

8-1	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	units of analysis are synthesized, the main evidence integration judgments4 typically focus on the

2	unit of analysis with the strongest evidence synthesis judgments, although exceptions may occur.

3	Structured evidence profile tables are used to summarize these analyses and foster consistency

4	within and across assessments. Instructions for using HAWC to create these tables are available at

5	the HAWC project "IRIS PPRTV SEM Template Figures and Resources" (see "Attachments," then

6	select the "Creating Evidence Profile Tables in HAWC").

4In some cases, as discussed in Section 8.2, it will be appropriate to draw multiple evidence integration
judgments within a given health effect category. This is generally dependent on data availability (i.e., more
narrowly defined categories may be possible with more evidence) and the ability to integrate the different
evidence streams at the level of these more granular categories. More granular categories will generally be
organized by predefined manifestations of potential toxicity. For example, within the health effect category of
immune effects, separate and different evidence integration judgments might be appropriate for
immunosuppression, immunostimulation, and sensitization and allergic response (i.e., the three types of
immunotoxicity described in the WHO guidance [2012]). Likewise, within the category of developmental
effects, it may be appropriate to draw separate judgments for potential effects on fetal death, structural
abnormality, altered growth, and functional deficits (i.e., the four manifestations of developmental toxicity
described in EPA guidelines (1991a)). These separate judgments are particularly important when the
evidence supports that the different manifestations might be based on different toxicological mechanisms. As
described for the evidence synthesis judgments, the strongest evidence integration judgment will typically be
used to reflect certainty in the broader health effect category.

This document is a draft for review purposes only and does not constitute Agency policy.

8-2	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 8-1. Generalized evidence profile table to show the relationship between evidence synthesis and evidence
integration to reach judgment of the evidence for hazard

Evidence synthesis (strength of evidence) judgments
(note that many factors and judgments require elaboration or evidence-based justification; see IRIS

Handbook for details)

Evidence integration
(weight of evidence)
judgment(s)

Studies

Summary of
key findings

Factors that increase
certainty
(applied to each unit of
analysis)

Factors that decrease

certainty
(applied to each unit of
analysis)

Evidence synthesis
judgment(s)

Describe overall evidence
integration judgment(s):

©©© Evidence demonstrates

Evidence from human studies

©OO Evidence suggests
OOO Evidence inadequate

	Strong evidence supports

no effect

Highlight the primary supporting
evidence for each integration
judgment3

Present inferences and
conclusions on:

•	Human relevance of findings in
animals3

•	Cross-stream coherence3

•	Potential susceptibility3

•	Understanding of biological
plausibility and MOA3

•	Other critical inferences3

•	Unit of
analysis #1

•	Studies
considered and
study confidence

• Descrip
tion of the
primary results

•	All/Mostly medium or high
confidence studies

•	Consistency

•	Dose-response gradient

•	Large or concerning
magnitude of effect

•	Coherence3

•	All/Mostly low confidence
studies

•	Unexplained inconsistency

•	Imprecision

•	Concerns about biological
significance3

•	Indirect outcome measures3

•	Lack of expected coherence3

• Judgment
reached for each
unit of analysis3

©0© Robust
©©O Moderate
©OO Slight
OOO

Indeterminate

	Compelling

evidence of no effect

•	Unit of
analysis #2

•	Studies
considered and
study confidence

• Descrip
tion of the
primary results

Evidence from animal studies

•	Unit of
analysis #1

•	Studies
considered and
study confidence

• Descrip
tion of the
primary results

•	All/Mostly medium or high
confidence studies

•	Consistency

•	Dose-response gradient

•	Large or concerning
magnitude of effect

•	Coherence3

•	All/Mostly low confidence
studies

•	Unexplained inconsistency

•	Imprecision

•	Concerns about biological
significance3

•	Indirect outcome measures3

•	Lack of expected coherence3

• Judgment
reached for each
unit of analysis

©©© Robust
©©O Moderate
©OO Slight
OOO

Indeterminate

	Compelling

evidence of no effect

•	Unit of
analysis #2

•	Studies
considered and
study confidence

• Descrip
tion of the
primary results

3Can be informed by key findings from the mechanistic analyses (see Table 8-2).

This document is a draft for review purposes only and does not constitute Agency policy.

8-3	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 8-2. Generalized evidence profile table to show the key findings and supporting rationale from mechanistic
analyses

Mechanistic analyses

Biological events or pathways (or other
relevant evidence grouping)

Summary of key findings and interpretation

Judgment(s) and rationale

Different analyses can be presented separately,
e.g., by exposure route or key uncertainty
addressed.

Each analysis can include multiple rows separated
bv biological events or other feature of the
approach used for the analysis.

•	Generally, will cite mechanistic synthesis
(e.g., for references, for detailed analysis)

•	Does not have to be chemical-specific
(e.g., read-across)

Can include separate summaries, for example bv
studv tvpe (e.g., new approach methods vs. in vivo

Overall summary of expert interpretation across
the assessed set of biological events, potential
mechanisms of toxicity, or other analysis approach
(e.g., adverse outcome pathway).

•	Includes the primary evidence supporting
the interpretation(s)

•	Describes and informs the extent to which
the evidence influences inferences across
evidence streams

•	Characterizes the limitations of the
evaluation and highlights existing data
gaps

•	May have overlap with factors
summarized for other streams

biomarkers), dose, or design.

Interpretation: Summary of expert interpretation
for the body of evidence and supporting rationale

Key findings: Summary of findings across the body
of evidence (may focus on or emphasize highly
informative designs or findings), including key
sources of uncertainty or identified limitations of
the study designs tested (e.g., regarding the
biological event or pathway being examined)

This document is a draft for review purposes only and does not constitute Agency policy.

8-4	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

8.1. EVIDENCE SYNTHESIS

IRIS assessments synthesize the evidence separately for each unit of analysis by focusing on
factors that increase or decrease certainty in the reported findings as evidence for hazard (see
Table 8-1). These factors are adapted from considerations for causality introduced by Austin
Bradford Hill (Hill. 19651 with some expansion and adaptation of how they are applied to facilitate
transparent application to chemical assessments that consider multiple streams of evidence.
Specifically, the factors considered are confidence in study findings (risk of bias [RoB] and
sensitivity), consistency across studies or experiments, dose/exposure-response gradient, strength
(effect magnitude) of the association, directness of outcome or endpoint measures, and coherence
[Table 6-3; see additional discussion in fU.S. EPA. 2022. 2005a. 19941], These factors are similar to
the domains considered in the GRADE (Grading of Recommendations Assessment, Development,
and Evaluation) Quality of Evidence framework (Schunemann etal.. 20131. Each of the considered
factors and the certainty of evidence judgments requires elaboration or evidence-based justification
in the synthesis narrative. Analysis of evidence synthesis considerations is qualitative (i.e.,
numerical scores are not developed, summed, or subtracted).

As previously described, the units of analysis may include predefined categories of
mechanistic evidence or other supplemental information (e.g., from studies of non-PECO routes of
exposure). This may include consideration of biomarkers or precursor events. Biological
understanding (e.g., knowledge of how an effect is manifest or progresses) or mechanistic inference
(e.g., dependency on a conserved key event across outcomes) can also be used to define which
related outcomes are considered as a unit of analysis. These considerations also inform the
evaluation of coherence and adversity within a unit of analysis and coherence with other units of
analyses. Mechanistic analyses outside the context of defining and evaluating the units of analysis
during evidence synthesis are considered as part of across stream evidence integration (see
Section 8.2).

Typically, human and animal evidence synthesis sections are structured similarly across
different units of analysis, health effects, and assessments. In contrast, the presentation, and
analyses of mechanistic and other types of supplemental information often differs within and
across assessments. This is due to the diversity of supplemental data that may be available and the
complexity of conducting supplemental analyses. For example, these data may inform unit of
analysis considerations, evidence integration judgments, or both. Each of the key analyses
informing the synthesis judgments are described in the narrative and summarized in an evidence
profile table.

Five levels of certainty in the evidence for (or against) a hazard are used to summarize
evidence synthesis judgments: robust (©©©, very little uncertainty exists), moderate (©©O,
some uncertainty exists), slight (©OO, large uncertainty exists), indeterminate (OOO), or
compelling evidence of no effect (—, little to no uncertainty exists for lack of hazard) (see Table 8-4

This document is a draft for review purposes only and does not constitute Agency policy.

8-5	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

and Table 8-5 for descriptions). Conceptually, before the evidence synthesis framework is applied,
certainty in the evidence is neutral (i.e., functionally equivalent to indeterminate). Next, the level of
certainty regarding the evidence for (or against) hazard is increased or decreased depending on
interpretations using the factors described in Table 8-3. Observations that increase certainty are
having consistency across high or medium confidence studies or experiments, the presence of
medium or high confidence studies with a strong dose-response gradient or observing a large or
concerning magnitude of effect, and coherent findings across medium or high confidence studies for
closely related endpoints (can include mechanistic endpoints) within the unit of analysis within an
evidence stream. Evidence from low confidence studies can further strengthen observations from
medium or high confidence studies but do not increase certainty on their own. Observations that
decrease certainty are having an evidence base of mostly low confidence studies, unexplained
inconsistency, lack of expected coherence, imprecision, unclear biological significance, null findings
with concerns for insensitivity (which decreases certainty in the lack of an effect), or indirect
measures of outcomes. Table 8-3 provides additional detail on how these factors are considered
when evaluating units of analysis.

This document is a draft for review purposes only and does not constitute Agency policy.

8-6	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 8-3. Considerations that inform evaluations and judgments of the strength of the evidence for hazard

Consideration

Increased evidence certainty
(of the human or animal evidence for hazard3)

Decreased evidence certainty
(of the human or animal evidence for hazard3)

Risk of bias and
sensitivity (across
studies)

•	An evidence base of mostly (or all) high or
medium confidence studies is interpreted as
being only minimally affected by bias and
insensitivity.

•	This factor should not be used if no other
factors would increase or decrease the
confidence for a given unit of analysis.

•	In addition, consideration of risk of bias and
sensitivity should inform how other factors are
evaluated, i.e., can inconsistency be potentially
explained by variation in confidence
judgments?

•	An evidence base of mostly (or all) low confidence studies decreases
strength. An exception to this is an evidence base of studies in which
the issues resulting in low confidence are related to insensitivity. This
may increase evidence certainty in cases in which an association is
identified because the expected impact of study insensitivity is toward
the null.

•	An evidence base of mostly null findings in which insensitivity is a
serious concern decreases certainty that the evidence is sufficient to
support a lack of health effect or association.

•	Decisions to increase certainty for other considerations in this table
should generally not be made if there are serious concerns for risk of
bias.

Consistency

• Similarity of findings for a given outcome

(e.g., of a similar direction) across independent
studies or experiments, especially when
medium or high confidence, increases certainty.
The increase in certainty is larger when
consistency is observed across populations
(e.g., geographical location) or exposure
scenarios in human studies, and across
laboratories, species, or exposure scenarios
(e.g., route; timing) in animal studies. When
seemingly inconsistent findings are identified,
patterns should be further analyzed to discern
if the inconsistencies can potentially be
explained based on study confidence, dose or
exposure levels, population, or experimental
model differences, etc. This factor is typically

• Unexplained inconsistency [i.e., conflicting evidence; see (U.S. EPA,
2005a)l decreases certaintv. Generally, certaintv should not be
decreased if discrepant findings can be reasonably explained by
considerations such as study confidence conclusions (including
sensitivity); variation in population or species, sex, or lifestage
(including understanding of differences in pharmacokinetics); or
exposure patterns (e.g., intermittent versus continuous), levels (low
versus high), or duration. Similar to current recommendations in the
Cochrane Handbook f(Higgins et al., 2022), see Section 7.8.61, clear
conflicts of interest (COI) related to funding source can be considered
as a factor to explain apparent inconsistency. For small evidence bases,
it might be hard to assess consistency. An evidence base of a single or a
few studies in which consistency cannot be accurately assessed does
not, alone, increase or decrease evidence certainty. Similarly, a
reasonable explanation for inconsistency does not necessarily result in
an increase in evidence certainty.

This document is a draft for review purposes only and does not constitute Agency policy.

8-7	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Consideration

Increased evidence certainty
(of the human or animal evidence for hazard3)

Decreased evidence certainty
(of the human or animal evidence for hazard3)



given the most attention during evidence
synthesis.



Effect magnitude
and imprecision

•	Evidence of a large or concerning magnitude of
effect can increase strength (generally only
when observed in medium or high confidence
studies).

•	Judgments on effect magnitude and
imprecision consider the rarity and severity of
the effect.

•	Certainty could be decreased if the findings are considered not likely to
be biologically significant. Effects that are small in magnitude might not
be considered biologically significant (adverse15) based on information
such as historical responses and variability. However, effects that
appear to be of small magnitude could be meaningful at the population
level e.g., IQ shifts); 1 such cases, certainty would not be decreased.

•	Certainty might also be decreased for imprecision, particularly if there
are only a few studies available to evaluate consistency in effect
magnitude across studies.

Dose-response

•	Evidence of dose-response or exposure-
response in high or medium confidence studies
increases certainty. Dose-response can be
demonstrated across studies or within studies
and it can be dose- or duration-dependent. It
could also not be a monotonic dose-response
(monotonicity should not necessarily be
expected as different outcomes might be
expected at low vs. high doses due to factors
such as activation of different mechanistic
pathways, systemic toxicity at high doses or
tolerance/acclimation). Sometimes, grouping
studies by level of exposure is helpful to
identify the dose-response pattern.

•	Decreases in a response (e.g., symptoms of
current asthma) after a documented cessation
of exposure also might increase certainty in a
relationship between exposure and outcome
(this is primarily applicable to epidemiology
studies because of their observational nature).

•	A lack of dose-response when expected on the basis of biological
understanding can decrease certainty in the evidence. If the data are
not adequate to evaluate a dose-response pattern, however, certainty
is neither increased nor decreased.

•	In some cases, duration-dependent patterns in the dose-response can
decrease evidence certainty. Such patterns are generally only
observable in experimental studies. Specifically, the magnitude of
effects at a given exposure level might decrease with longer exposures
(e.g., due to tolerance or acclimation), or effects might rapidly resolve
under certain experimental conditions (e.g., reversibility after removal
of exposure). As many reversible and short-lived effects can be of high
concern, decisions about whether such patterns decrease evidence
certainty depend on considering the pharmacokinetics of the chemical
and the conditions of exposure [see U.S. EPA (1998)], endpoint severity,
judgments regarding the potential for delayed or secondary effects, the
underlying mechanism(s) involved, and the exposure context focus of
the assessment (e.g., addressing intermittent or short-term exposures).

This document is a draft for review purposes only and does not constitute Agency policy.

8-8	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Consideration

Increased evidence certainty
(of the human or animal evidence for hazard3)

Decreased evidence certainty
(of the human or animal evidence for hazard3)

Directness of
outcome/end point
measures

• Not applicable

•	If the evidence base primarily includes outcomes or endpoints that are
indirect measures (e.g., biomarkers) of the unit of analysis, certainty
(for that unit of analysis) is typically decreased. Judgments to decrease
certainty based on indirectness should focus on findings for measures
that have an unclear linkage to an apical or clinical (adverse15) outcome.
Scenarios in which the magnitude of the response is not considered to
reflect a biologically meaningful level of change (i.e., biological
significance; see "effect magnitude and imprecision" row, above) are
not considered under indirectness of outcome measures.

•	Related to indirectness, certainty in the evidence can be decreased
when the findings are determined to be nonspecific to the hazard
under evaluation. This consideration is generally only applicable to
animal evidence and the most common example is effects only with
exposures (level, duration) shown to cause excessive toxicity in that
species and lifestage (including consideration of maternal toxicity in
developmental evaluations). This does not apply when an effect is
viewed as secondary to other changes (e.g., effects on pulmonary
function because of disrupted immune responses).

Coherence

•	Biologically related findings within or across
studies, within an organ system or across
populations (e.g., sex), increase certainty
(generally only when observed in medium or
high confidence studies). Certainty is further
increased when a temporal or dose-dependent
progression of related effects is observed
within or across studies, or when related
findings of increasing severity are observed
with increasing exposure.

•	Coherence across findings within a unit of
analysis (e.g., consistent changes in disease
markers and biological precursors in exposed

• An observed lack of expected coherent changes (e.g., in well-

established biological relationships) within or across biologically related
units of analysis will typically decrease evidence certainty. This includes
mechanistic changes when included in the unit of analysis. However, as
described for decisions to increase certainty, confidence in the
understanding of the biological relationships between the endpoints
being compared, and the sensitivity and specificity of the measures
used, need to be carefully examined. The decision to decrease certainty
depends on the availability of evidence across multiple related
endpoints for which changes would be anticipated, and it considers
factors (e.g., dose and duration of exposure, strength of expected
relationship) across the studies of related changes.

This document is a draft for review purposes only and does not constitute Agency policy.

8-9	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Consideration

Increased evidence certainty
(of the human or animal evidence for hazard3)

Decreased evidence certainty
(of the human or animal evidence for hazard3)



humans) can increase certainty in the evidence
for an effect.

•	Coherence within or across biologically related
units of analysis can also increase certainty for
a given (or multiple) unit(s) of analysis. This
considers certainty in the biological
relationships between the endpoints being
compared, and the sensitivity and specificity of
the measures used.

•	Mechanistic support for, or biological
understanding of, the relatedness between
different endpoints within (or across different)
units of analysis, can inform an understanding
of coherence.



Other factors

• Unusual scenarios that cannot be addressed by
the considerations above, e.g., read-across
inferences supporting the adversity of observed
changes.

• Unusual scenarios that cannot be addressed by the considerations
above, e.g., strong evidence of publication bias.0

aAlthough the focus is on identifying potential adverse human health effects (hazards) of exposure, these factors can also be used to increase or decrease
certainty in the evidence supporting lack of an effect (e.g., leading to a judgment of compelling evidence of no effect). The latter application is not explicitly
outlined here.

bWithin this framework, evidence synthesis judgments reflect an interpretation of the evidence for a hazard; thus, consideration of the adversity of the findings
is an explicit aspect of the analyses. To better define how adversity is evaluated, the consideration of adversity is broken into the two, sometimes related,
considerations of the indirectness of the outcome measures and the interpreted biological significance of the effect magnitude.

Publication bias involves the influence of the direction, magnitude, or statistical significance of the results on the likelihood of a paper being published; it can
result from decisions made, consciously or unconsciously, by study authors, journal reviewers, and journal editors (Dickersin, 1990). This could make the
available evidence base unrepresentative. However, publication bias can be difficult to evaluate (NTP, 2019) and should not be used as a factor that decreases
certainty unless there is strong evidence.

This document is a draft for review purposes only and does not constitute Agency policy.

8-10	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

A structured framework approach is used to draw evidence synthesis judgments for human
and animal evidence. Tables 8-4 and 8-5 (for human and animal evidence, respectively) provide the
criteria that guide how to draw the strength of evidence judgments for each unit of analysis within
a health effect category and the terms used to summarize those judgments. These terms are applied
to human and animal evidence separately. The terms robust and moderate are characterizations for
judgments that the evidence (across studies) supports a conclusion that the effect(s) results from
the exposure being assessed. These two terms are differentiated by the quality and amount of
information available to rule out alternative explanations for the results. For example, repeated
observations of effects by independent studies or experiments examining various aspects of
exposure or response (e.g., different exposure settings, dose levels or patterns, populations or
species, biologically related endpoints) result in increased certainty in the evidence for hazard. The
term slight indicates situations in which there is some evidence supporting an association within
the evidence stream, but substantial uncertainties in the data exist to prevent judgments that the
effect(s) can be reliably attributed to the exposure being assessed. Indeterminate reflects judgments
for a wide variety of evidence scenarios, including when no studies are available or when the
evidence from studies of similar confidence has a high degree of unexplained inconsistency.
Compelling evidence of no effect represents a rare situation in which extensive evidence across a
range of populations and exposures has demonstrated that no effects are likely attributable to the
exposure being assessed. This category is applied at the health effect level (e.g., hepatic effects)
rather than more granular units of analysis level to avoid giving the impression of confidence in
lack of a health effect when aspects of potential toxicity have not been adequately examined.
Reaching this judgment is infrequent because it requires both a high degree of confidence in the
conduct of individual studies, including consideration of study sensitivity, as well as comprehensive
assessments of outcomes and lifestages of exposure that adequately address concern for the hazard
under evaluation.

This document is a draft for review purposes only and does not constitute Agency policy.

8-11	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 8-4. Framework for strength of evidence judgments from studies in
humans

Evidence
synthesis
judgment

Description

Robust (©©©)
...evidence in
human studies

(strong signal of
effect with very
little uncertainty)

A set of high or medium confidence independent studies (e.g., in different populations)
reporting an association between the exposure and the health outcome(s), with reasonable
confidence that alternative explanations, including chance, bias, and confounding, can be ruled
out across studies. The set of studies is primarily consistent, with reasonable explanations when
results differ; the findings are considered adverse (i.e., biologically significant and without
notable concern for indirectness); and an exposure-response gradient is demonstrated.
Additional supporting evidence, such as associations with biologically related endpoints in
human studies (coherence) or large estimates of risk or severity of the response, can increase
certainty but are not required. Supplemental evidence included in the unit of analysis (e.g.,
mechanistic studies in exposed humans or human cells) could raise the certainty in the evidence
to robust for a set of studies that otherwise would be described as moderate. Such evidence not
included in the unit of analysis can also inform evaluations of the coherence of the human
evidence, the directness of the outcome measures, and the biological significance of the
findings. Causality is inferred for a human evidence base of robust.

Moderate

(©©O)

...evidence in
human studies

(signal of effect
with some
uncertainty)

A set of evidence that does not reach the degree of certainty required for robust, but which
includes at least one high or medium confidence study reporting an association and additional
information increasing certainty in the evidence. For multiple studies, there is primarily
consistent evidence of an association with reasonable support for adversity, but there might be
some uncertainty due to potential chance, bias, or confounding or because of the indirectness of
some measures. When only a single study is available in the unit of analysis, there is a large
magnitude or severity of the effect, or a dose-response gradient, or other supporting evidence,
and there are no serious residual methodological uncertainties. Supplemental evidence included
in the unit of analysis might address the above factors and raise certainty in the evidence to
moderate for a set of studies that otherwise would be described as slight or, in exceptional
cases, could support raising to moderate evidence that would otherwise be described as
indeterminate. Mechanistic evidence not included in the unit of analysis can also inform
evaluations of the coherence of the human evidence, the directness of the outcome measures,
and the biological significance of the findings.

Slight

(©OO)

...evidence in
human studies

(signal of effect
with large amount
of uncertainty)

One or more studies reporting an association between exposure and the health outcome, but
considerable uncertainty exists and supporting coherent evidence is sparse. In general, the
evidence is limited to a set of consistent low confidence studies, or higher confidence studies
with significant unexplained heterogeneity or other serious residual uncertainties. It also applies
when one medium or high confidence study is available within the unit of analysis without
additional information strengthening the likelihood of a causal association (e.g., coherent
findings within the same study or from other studies). This category serves primarily to
encourage additional study when evidence does exist that might provide some support for an
association, but for which the evidence does not reach the degree of confidence required for
moderate.

This document is a draft for review purposes only and does not constitute Agency policy.

8-12	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Evidence
synthesis
judgment

Description

Indeterminate

(OOO)

...evidence in
human studies

(signal cannot be
determined for or
against an effect)

No studies available in humans or situations when the evidence is inconsistent and primarily of
low confidence. In addition, this might include situations in which higher confidence studies
exist, but there are major concerns with the evidence base such as unexplained inconsistency, a
lack of expected coherence from a stronger set of studies, very small effect magnitude (i.e.,
major concerns about biological significance), or uncertainties or methodological limitations that
result in an inability to discern effects from exposure. It also applies for a single low confidence
study in the absence of factors that increase certainty. A set of largely null studies could be
concluded to be indeterminate if the evidence does not reach the level required for compelling
evidence of no effect.

Compelling
evidence of no
effect
(...)

...in human
studies

(strong signal for
lack of an effect
with little
uncertainty)

A set of high confidence studies examining a reasonable spectrum of endpoints showing null
results (e.g., an odds ratio of 1.0), ruling out alternative explanations including chance, bias, and
confounding with reasonable confidence. Each of the studies should have used an optimal
outcome and exposure assessment and adequate sample size (specifically for higher exposure
groups and for susceptible populations). The set as a whole should include diverse sampling
(across sexes [if applicable] and different populations) and include the full range of levels of
exposures that human beings are known to encounter, an evaluation of an exposure-response
gradient, and an examination of at-risk populations and lifestages. Supplemental evidence can
help to address the above considerations or, when included in the unit of analysis, provide
additional support for this judgment.

Table 8-5. Framework for strength of evidence judgments from studies in
animals

Evidence
synthesis
judgment

Description

Robust (©©©)
...evidence in
animal studies

(strong signal of
effect with very
little

uncertainty)

The set of high or medium confidence, independent experiments (i.e., across laboratories,
exposure routes, experimental designs [for example, a subchronic study and a multigenerational
study], or species) reporting effects of exposure on the health outcome(s). The set of studies is
primarily consistent, with reasonable explanations when results differ (i.e., due to differences in
study design, exposure level, animal model, or study confidence), and the findings are considered
adverse (i.e., biologically significant and without notable concern for indirectness). At least two of
the following additional factors in the set of experiments increase certainty in the evidence:
coherent effects across multiple related endpoints (within or across biologically related units of
analysis); an unusual magnitude of effect, rarity, age at onset, or severity; a strong dose-response
relationship; or consistent observations across animal lifestages, sexes, or strains. Supplemental
evidence included in the unit of analysis (e.g., mechanistic studies in exposed animals or animal
cells) might raise the certainty of evidence to robust for a set of studies that otherwise would be
described as moderate. Such evidence not included in the unit of analysis can also inform
evaluations of the coherence of the animal evidence, the directness of the outcome measures,
and the biological significance of the findings.

This document is a draft for review purposes only and does not constitute Agency policy.

8-13	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Evidence
synthesis
judgment

Description

Moderate

(©©O)

...evidence in
animal studies

(signal of effect
with some
uncertainty)

A set of evidence that does not reach the degree of certainty required for robust, but which
includes at least one high or medium confidence study and additional information increasing
certainty in the evidence. For multiple studies or a single study, the evidence is primarily
consistent or coherent with reasonable support for adversity, but there are notable remaining
uncertainties (e.g., difficulty interpreting the findings due to concerns for indirectness of some
measures); however, these uncertainties are not sufficient to reduce or discount the level of
concern regarding the positive findings and any conflicting findings are from a set of experiments
of lower confidence. The set of experiments supporting the effect provide additional information
increasing certainty in the evidence, such as consistent effects across laboratories or species;
coherent effects across multiple related endpoints (can include mechanistic endpoints within the
unit of analysis); an unusual magnitude of effect, rarity, age at onset, or severity; a strong dose-
response relationship; or consistent observations across exposure scenarios (e.g., route, timing,
duration), sexes, or animal strains. Supplemental evidence included in the unit of analysis could
address the above factors and raise certainty in the evidence to moderate for a set of studies that
otherwise would be described as slight or, in exceptional cases, might support raising to moderate
evidence that would otherwise be described as indeterminate. Mechanistic evidence not included
in the unit of analysis can also inform evaluations of the coherence of the animal evidence, the
directness of the outcome measures, and the biological significance of the findings.

Slight

(©OO)

...evidence in
animal studies

(signal of effect
with large
amount of
uncertainty)

One or more studies reporting an effect on an exposure on the health outcome, but considerable
uncertainty exists and supporting coherent evidence is sparse. In general, the evidence is limited
to a set of consistent low confidence studies, or higher confidence studies with significant
unexplained heterogeneity or other serious uncertainties (e.g., concerns about adversity) across
studies. It also applies when one medium or high confidence experiment is available within the
unit of analysis without additional information increasing certainty in the evidence (e.g., coherent
findings within the same study or from other studies). Biological evidence from mechanistic
studies could also be independently interpreted as slight. This category serves primarily to
encourage additional study for which evidence does exist that might provide some support for an
association, but for which the evidence does not reach the degree of confidence required for
moderate.

Indeterminate

(OOO)

...evidence in
animal studies

(signal cannot
be determined
for or against an
effect)

No studies available in animals or situations when the evidence is inconsistent and primarily of
low confidence. In addition, this might include situations in which higher confidence studies exist,
but there are major concerns with the evidence base such as unexplained inconsistency, a lack of
expected coherence from a stronger set of studies, very small effect magnitude (i.e., major
concerns about biological significance), or uncertainties or methodological limitations that result
in an inability to discern effects from exposure. It also applies for a single low confidence study in
the absence of factors that increase certainty. A set of largely null studies could be concluded to
be indeterminate if the evidence does not reach the level required for compelling evidence of no
effect.

Compelling
evidence of no
effect
(...)

...in animal
studies
(strong signal
for lack of an
effect with little
uncertainty)

A set of high confidence experiments examining a reasonable spectrum of endpoints that
demonstrate a lack of biologically significant effects across multiple species, both sexes, and a
broad range of exposure levels. The data are compelling in that the experiments have examined
the range of scenarios across which health effects in animals could be observed, and an
alternative explanation (e.g., inadequately controlled features of the studies' experimental
designs; inadequate sample sizes) for the observed lack of effects is not available. Each of the
studies should have used an optimal endpoint and exposure assessment and adequate sample
size. The evidence base should represent both sexes and address potentially susceptible
populations and lifestages. Supplemental evidence can help to address the above considerations
or, when included in the unit of analysis, provide additional support for this judgment.

This document is a draft for review purposes only and does not constitute Agency policy.

8-14	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

8.2. EVIDENCE INTEGRATION

The phase of evidence integration combines animal and human evidence synthesis
judgments while also considering information on the human relevance of findings in animal
evidence, coherence across evidence streams ("cross-stream coherence"), information on
susceptible populations or lifestages, understanding of biological plausibility or MOA, and
potentially other critical inferences (e.g., read-across analyses) that generally draw on mechanistic
and other supplemental evidence (see Table 8-6). This analysis culminates in an evidence
integration judgment and narrative for each potential health effect category (i.e., each noncancer
health effect and specific type of cancer, or broader grouping of related outcomes as defined during
problem formulation). To the extent it can be characterized prior to conducting dose-response
analyses, exposure context is also provided.

With respect to susceptibility, the assessment describes the evidence (i.e., human, animal,
mechanistic) on populations and lifestages most likely to be susceptible to the hazards identified
and, to the extent possible, the factors that increase their risk for the hazards. In addition to
assessment-specific health effects evidence, background information about biological mechanisms
or ADME, as well as biochemical and physiological differences among lifestages and sexes, may be
used. At a minimum, particular consideration is given to infants and children, pregnant women, and
women of childbearing age. Many of the foundational analyses for summarizing susceptibility in the
evidence integration narrative are undertaken during evidence synthesis as patterns across studies
are evaluated with respect to consistency, coherence, and the magnitude and direction of effect
measures. Relevant factors for exploring patterns may include intrinsic factors (e.g., age, sex,
genetics, health status, behaviors) and certain extrinsic factors (e.g., socioeconomic status, access to
health care), although information on the latter is rarely available in human health studies of
environmental chemicals.

Table 8-6. Considerations that inform evidence integration judgments

Judgment

Description

Human relevance
of findings

Used to describe and justify the interpreted relevance of the data from experimental animals
(or other model systems) to humans. In the absence of chemical-specific evidence informing
human relevance, the evidence integration narrative will briefly describe the interpreted
underlying biological similarity across species. As noted in EPA guidelines (U.S. EPA, 2005a),
there needs to be evidence or a biological explanation to support an interpreted lack of
human relevance for findings in animals, and site concordance is neither expected nor
required. Thus, in the absence of specific evidence or cross-species understanding of the
underlying biology, it is appropriate to use a statement such as, "without evidence to the
contrary, [health effect] responses in animals are presumed relevant to humans."

Cross-stream
coherence

Used to address the concordance of biologically related findings across human, animal, and
mechanistic studies, considering features of the available evidence such as exposure timing
and cancer), it is not necessary or expected that effects manifest in humans are identical to
those observed in animals (e.g., tumors in animals can be predictive of carcinogenic potential

This document is a draft for review purposes only and does not constitute Agency policy.

8-15	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Judgment

Description



in humans, but not necessarily at the same site), although this typically provides stronger
evidence. Biological understanding of the manner in which the outcomes are manifest in
different species can inform cross-stream coherence. Evidence supporting a biologically
plausible mechanistic pathway across species adds coherence (see below).

Susceptible
populations and
lifestages

Used to summarize analyses relating to individual and social factors that may increase
susceptibility to exposure-related health effects in certain populations or lifestages, or to
highlight the lack of such information. These analyses are based on knowledge about the
health outcome or organ system affected and focus on the influence of intrinsic biological
factors but can also include consideration of mechanistic and ADME evidence.

Biological
plausibility and
MOA

considerations

Used to summarize the interpreted biological plausibility of an association between exposure
and the health effect, based primarily on the extent to which the available evidence comports
with the known development and characteristics of the health effect (and thus dependent on
sufficient information being available to draw such an interpretation). Importantly, because
this interpretation is dependent on canonical scientific knowledge about the health effect, the
lack of such understanding does not provide a rationale to decrease certainty in the evidence
for an effect (NTP, 2015; NRC, 2014). These analyses can be detailed (e.g., when
attempting to establish MOA understanding) and, if so, are typically conducted separately
(e.g., as part of the mechanistic evidence synthesis) and then referenced in the evidence
integration narrative.

Other critical

inferences

(optional)

Can be used to describe the consideration of other evidence or non-chemical-specific
information that informs evidence integration judgments (e.g., use of read-across analyses or
ADME understanding used to inform the other considerations described below; judgments on
other health effects expected to be linked to the health effect under evaluation).

ADME = absorption, distribution, metabolism, and excretion; MOA = mode of action.

Using a structured framework approach, one of five phrases is used to summarize the
evidence integration judgment based on the integration of the evidence synthesis judgments, taking
into account the additional considerations assessed across evidence streams: evidence
demonstrates, evidence indicates (likely), evidence suggests, evidence is inadequate, or strong evidence
supports no effect (see Table 8-7). The five evidence integration judgment levels reflect the
differences in the amount and quality of the data that inform the evaluation of whether exposure is
interpreted as capable of causing the health effect(s). As it is assumed that any identified health
hazards will only be manifest given exposures of a certain type and amount (e.g., a specific route; a
minimal duration, periodicity, and level), the evidence integration narrative and summary
judgment levels include the generic phrase, "given sufficient exposure conditions." This highlights
that, for those assessment-specific health effects identified as potential hazards, the exposure
conditions associated with those health effects will be defined (as will the uncertainties in the
ability to define those conditions) during dose-response analysis (see Section 8). More than one
evidence integration judgment level can be used when the evidence base is able to support that a
chemical's effects differ by exposure level or route fU.S. EPA. 2005al The analyses and judgments
are summarized in the evidence profile table (see Table 8-1).

This document is a draft for review purposes only and does not constitute Agency policy.

8-16	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

For evaluations of carcinogenicity, consistent with EPA's Cancer Guidelines (U.S. EPA.
2005a), one of EPA's standardized cancer descriptors is used to describe the overall potential for
carcinogenicity within the evidence integration narrative for carcinogenicity. These descriptors are:
(1) carcinogenic to humans, (2) likely to be carcinogenic to humans, (3) suggestive evidence of
carcinogenic potential, (4) inadequate information to assess carcinogenic potential, or (5) not
likely to be carcinogenic to humans. The standardized cancer descriptors will often align with the
evidence integration judgments (i.e., "evidence demonstrates" aligns with "carcinogenic to
humans") but not in all cases. For example, the evidence integration judgments are generally used
for individual tumor or cancer types and the standardized EPA descriptors are used to characterize
overall cancer hazard. For each type of cancer evaluated (e.g., lung cancer; renal cancer) or sets of
related cancer types, an evidence integration narrative and summary judgment level are provided
as described above for noncancer health effects. When considering evidence on carcinogenicity
across human and animal evidence, site concordance is not required (U.S. EPA. 2005a). If a
systematic review of more than one cancer type was conducted, then the strongest evidence
integration judgment(s) is used as the basis for selecting the standardized cancer descriptor in
accordance with the EPA cancer guidelines (U.S. EPA. 2005a). including application of the MOA
framework (incorporating an evaluation of evidence relevant to potential mutagenicity).

Similar to the description for summarizing noncancer judgments above, the cancer
descriptor and evidence integration narrative for carcinogenicity also consider the conditions of
carcinogenicity, including exposure (e.g., route; level) and susceptibility (e.g., genetics; lifestage), as
the data allow (Farland. 2005: U.S. EPA. 2005a. b). As with noncancer effects, the specific exposure
conditions necessary for carcinogenicity are further defined during dose-response analysis.

This document is a draft for review purposes only and does not constitute Agency policy.

8-17	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 8-7. Framework for summary evidence integration judgments in the evidence integration narrative

Summary evidence integration
judgment3 in narrative

Evidence
integration
judgment level

Explanation and example scenarios'3

The currently available evidence
demonstrates that [chemical] causes
[health effect] in humans0 given sufficient
exposure conditions. This conclusion is
based on studies of [humans or animals]
that assessed [exposure or dose] levels of
[range of concentrations or specific cutoff
level concentration01].

Evidence
demonstrates

A strong evidence base demonstrating that [chemical] exposure causes [health effect] in
humans.

•	This conclusion level js used if there is robust human evidence supporting an
effect.

•	This conclusion level could also be used with moderate human evidence and robust
animal evidence if there is strong mechanistic evidence that MOAs and key
precursors identified in animals are anticipated to occur and progress in humans.

The currently available evidence indicates
that [chemical] likely causes [health
effect] in humans given sufficient
exposure conditions. This conclusion is
based on studies of [humans or animals]
that assessed [exposure or dose] levels of
[range of concentrations or specific cutoff
level concentration].

Evidence indicates
(likely®)

An evidence base that indicates that [chemical] exposure likely causes [health effect] in
humans, although there may be outstanding questions or limitations that remain, and the
evidence is insufficient for the higher conclusion level.

•	This conclusion level is used if there is robust animal evidence supporting an effect
and slight-to-indeterminate human evidence, or with moderate human evidence
when strong mechanistic evidence is lacking.

•	This conclusion level could also be used with moderate human evidence
supporting an effect and moderate-to-indeterminate animal evidence, or with
moderate animal evidence supporting an effect and moderate-to-indeterminate
human evidence. In these scenarios, any uncertainties in the moderate evidence
are not sufficient to substantially reduce confidence in the reliability of the
evidence, or mechanistic evidence in the slight or indeterminate evidence base
(e.g., precursors) exists to increase confidence in the reliability of the moderate
evidence.

The currently available evidence
suggests that [chemical] may cause
[health effect] in humans given sufficient
exposure conditions. This conclusion is
based on studies of [humans or animals]
that assessed [exposure or dose] levels

Evidence suggests

An evidence base that suggests that [chemical] exposure may cause [health effect] in
humans, but there are very few studies that contributed to the evaluation, the evidence is
very weak or conflicting, and/or the methodological conduct of the studies is poor.

• This conclusion level is used if there is slight human evidence and indeterminate-
to-slight animal evidence.

This document is a draft for review purposes only and does not constitute Agency policy.

8-18	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Summary evidence integration
judgment3 in narrative

Evidence
integration
judgment level

Explanation and example scenarios'3

of [range of concentrations or specific
cutoff level concentration].



•	This conclusion level is also used with slight animal evidence and indeterminate-
to-slight human evidence.

•	This conclusion level could also be used with moderate human evidence and
slight or indeterminate animal evidence, or with moderate animal evidence and
slight or indeterminate human evidence. In these scenarios, there are
outstanding issues or uncertainties regarding the moderate evidence (i.e., the
synthesis judgment was borderline with slight), or mechanistic evidence in the
slight or indeterminate evidence base (e.g., null results in well-conducted
evaluations of precursors) exists to decrease confidence in the reliability of the
moderate evidence.

•	Exceptionally, when there is general scientific understanding of mechanistic
events that result in a health effect, this conclusion level could also be used if
there is strong mechanistic evidence that is sufficient to highlight potential
human toxicity'—in the absence of informative conventional studies in humans or
in animals (i.e., indeterminate evidence in both).

The currently available evidence is
inadequate to assess whether [chemical]
may cause [health effect] in humans.

Evidence
inadequate

This conveys either a lack of information or an inability to interpret the available evidence
for [health effect]. On an assessment-specific basis, a single use of this "inadequate"
conclusion level might be used to characterize the evidence for multiple health effect
categories (i.e., all health effects that were examined and did not support other conclusion
levels).8

•	This conclusion level is used if there is indeterminate human and animal evidence.

•	This conclusion level is_also used with slight animal evidence and compelling
evidence of no effect human evidence.

•	This conclusion level could also be used with sliaht-to-robust animal evidence and
indeterminate human evidence if strong mechanistic information indicated that
the animal evidence is unlikely to be relevant to humans.

A conclusion of inadequate is not a determination that the agent does not cause the
indicated health effect(s). It simply indicates that the available evidence is insufficient to
reach conclusions.

This document is a draft for review purposes only and does not constitute Agency policy.

8-19	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Summary evidence integration
judgment3 in narrative

Evidence
integration
judgment level

Explanation and example scenarios'3

Strong evidence supports no effect in

humans. This conclusion is based on
studies of [humans or animals] that
assessed [exposure or dose] levels of
[range of concentrations].

Strong evidence
supports no effect

This represents a situation in which extensive evidence across a range of populations and
exposure levels has identified no effects/associations. This scenario requires a high degree
of confidence in the conduct of individual studies, including consideration of study
sensitivity, and comprehensive assessments of the endpoints and lifestages of exposure
relevant to the heath effect of interest.

•	This conclusion level is used if there is compelling evidence of no effect in human
studies and compelling evidence of no effect-to-indeterminate in animals.

•	This conclusion level is also used if there is indeterminate human evidence and
compelling evidence of no effect animal evidence in models concluded to be
relevant to humans.

•	This conclusion level could also be used with compellina evidence of no effect in
human studies and moderate-to-robust animal evidence if strong mechanistic
information indicated that the animal evidence is unlikely to be relevant to
humans.

aEvidence integration judgments are typically developed at the level of the health effect when there are sufficient studies on the topic to evaluate the evidence
at that level; this should always be the case for "evidence demonstrates" and "strong evidence supports no effect," and typically for "evidence indicates
(likely)." However, some databases only allow for evaluations at the category of health effects examined; this will more frequently be the case for conclusion
levels of "evidence suggests" and "evidence inadequate." A judgment of "strong evidence supports no effect" is drawn at the health effect level,
terminology of "is" refers to the default option; terminology of "could also be" refers to situational options dependent on mechanistic understanding.
cln some assessments, these conclusions might be based on data specific to a particular lifestage of exposure, sex, or population (or another specific group). In
such cases, this would be specified in the narrative conclusion, with additional detail provided in the narrative text. This applies to all conclusion levels.
dlf concentrations cannot be estimated, an alternative expression of exposure level such as "occupational exposure levels," are provided. This applies to all
conclusion levels.

eFor some applications, such as benefit-cost analysis, to better differentiate the categories of "evidence demonstrates" and "evidence indicates," the latter
category should be interpreted as evidence that supports an exposure-effect linkage that is likely to be causal.

'Scientific understanding of adverse outcome pathways and of the human implications of new toxicity testing methods (e.g., from high-throughput screening,
from short-term in vivo testing of alternative species or from new in vitro testing) will continue to increase. This may make possible the development of
hazard conclusions when there are mechanistic or other relevant data that can be interpreted with a similar level of confidence to positive animal results in
the absence of conventional studies in humans or in animals.

Specific narratives for each of these health effects may also be deemed unnecessary.

This document is a draft for review purposes only and does not constitute Agency policy.

8-20	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

9.DOSE-RESPONSE ASSESSMENT: SELECTING
STUDIES AND QUANTITATIVE ANALYSIS

9.1. OVERVIEW

Selection of specific data sets for dose-response assessment and performance of the
dose-response assessment is conducted after hazard identification is complete and involves
database- and chemical-specific biological judgments. A number of EPA guidelines and support
documents detail data requirements and other considerations for dose-response modeling,
especially EPA's Benchmark Dose Technical Guidance (U.S. EPA. 20121. EPA's Review of the Reference
Dose and Reference Concentration Processes [(U.S. EPA. 2005a. 20021. Guidelines for Carcinogen Risk
Assessment (U.S. EPA. 2005a). and Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens (U.S. EPA. 2005b). This section of the Protocol provides an overview of
considerations for conducting the dose-response assessment, particularly statistical considerations
specific to dose-response analysis that support quantitative risk assessment Importantly, these
considerations do not supersede existing EPA guidelines.

For IRIS assessments, dose-response assessments are typically performed for both
noncancer and cancer hazards, and for both oral and inhalation routes of exposure following
chronic exposure5 to the chemical of interest, if supported by existing data. For noncancer hazards,
an oral reference dose (RfD) will be derived. (Inhalation toxicity values will not be derived in this
assessment of nitrate/nitrite.) An RfD is an estimate, with uncertainty spanning perhaps an order of
magnitude, of an exposure to the human population (including susceptible populations and
lifestages) that is likely to be without an appreciable risk of deleterious health effects over a lifetime
(U.S. EPA. 20021. In addition to an RfD, this assessment will attempt to derive organ- or system-
specific RfDs (osRfDs) when the data are sufficiently strong (i.e., noncancer conclusions of evidence
demonstrate or evidence indicates [likely]). An RfD may also be derived for cancer effects in cases in
which a nonlinear MOA is concluded that indicates a key precursor event necessary for
carcinogenicity does not occur below a specific exposure level ffU.S. EPA. 2005al. §3.3.4). In
addition to an RfD, when feasible and if the available data are appropriate for doing so, the
assessments will derive a less-than-lifetime toxicity value (a "subchronic" reference value) for
noncancer hazards. Both less-than-lifetime and hazard-specific values may be useful to EPA risk
assessors within specific decision contexts.

5Dose-response assessments may also be conducted for shorter durations, particularly if the evidence base
for a chemical indicates risks associated with shorter exposures to the chemical (U.S. EPA. 20021.

This document is a draft for review purposes only and does not constitute Agency policy.

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

23

24

25

26

27

28

29

30

31

32

33

34

35

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

When low-dose linear extrapolation for cancer effects is supported, particularly for
chemicals with direct mutagenic activity or those for which the data indicate a linear component
below the point of departure (POD), an OSF facilitates estimation of human cancer risks. Low-dose
linear extrapolation is also used as a default when the data are insufficient to establish the mode of
action fU.S. EPA. 2005al An OSF is a plausible upper-bound lifetime cancer risk from chronic
ingestion of a chemical (expressed as mg/kg-day). In contrast with reference doses (RfDs), an OSF
can be used in conjunction with exposure information to estimate cancer risk at a given dose.

The derivation of toxicity values also depends on the nature of the hazard conclusion.
Specifically, EPA generally conducts dose-response assessments and derives cancer values for
chemicals that are classified as carcinogenic or likely to be carcinogenic to humans. When there is
suggestive evidence of carcinogenic potential to humans, EPA generally would not conduct a
dose-response assessment and derive a cancer value. Similarly, for noncancer outcomes dose-
response is conducted based on having stronger evidence of a hazard (generally, "evidence
demonstrates" and "evidence indicates [likely]". When the noncancer outcome is considered
"evidence suggests" of potential hazard to humans, EPA generally would not conduct a
dose-response assessment and derive a RfC or RfD. Cases in which suggestive evidence might be
used to develop cancer risk estimates or noncancer toxicity value include when the evidence base
includes a well-conducted study (overall medium or high confidence for the outcome), quantitative
analyses may be useful for some purposes, (e.g., providing a sense of the magnitude and uncertainty
of potential risks, ranking potential hazards, or setting research priorities) (U.S. EPA. 2005a).

9.2. SELECTING STUDIES FOR DOSE-RESPONSE ASSESSMENT

9.2.1. Hazard and MOA Considerations for Dose Response

The assessment presents a summary of hazard identification conclusions to transition to
dose response considerations, highlighting (1) information used to inform the selection of
outcomes or broader health effect categories for which toxicity values will be derived, (2) whether
toxicity values can be derived to protect specific populations or life stages, (3) how dose response
modeling will be informed by pharmacokinetic information, and (4) the identification of
biologically based BMR levels. The pool of outcomes and study-specific endpoints is discussed to
identify which categories of effects and study designs are considered the strongest and most
appropriate for quantitative assessment of a given health effect, particularly among the studies that
exemplify the study attributes summarized in Table 9-1.

Also considered is whether there are opportunities for quantitative evidence integration.
Examples of quantitative integration, from simplest to more complex, include (1) combining results
for an outcome across sex (within a study); (2) characterizing overall toxicity, as in combining
effects that comprise a syndrome, or occur on a continuum (e.g., precursors and eventual overt
toxicity, benign tumors that progress to malignant tumors); and (3) conducting a meta-analysis or
meta-regression of all studies addressing a category of important health effects.

This document is a draft for review purposes only and does not constitute Agency policy.

9-2	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Some studies that are used qualitatively for hazard identification may or may not be useful
quantitatively for dose-response assessment due to such factors as the lack of quantitative
measures of exposure or lack of variability measures for response data. If the needed information
cannot be located, semiquantitative analysis may be feasible (e.g., via NOAEL/LOAEL). In the draft
and final assessments, specific endpoints considered for dose-response are summarized in a tabular
format that includes rationales for decisions to proceed (or not) for POD derivation.

In addition, mechanistic evidence that influences the dose-response analyses is highlighted,
for example, evidence related to susceptibility or potential shape of the dose-response curve
(i.e., linear, nonlinear, or threshold model). Mode(s) of action is summarized including any
interactions between them relevant to understanding overall risk. For cancer dose-response,
biological considerations relevant to dose-response for cancer are:

•	Is there evidence for direct mutagenicity?

•	Does tumor latency decrease with increasing exposure?

•	If there are multiple tumor types, which cancers have a longer latency period?

•	Is incidence data available (incidence data are preferred to mortality data)?

•	Were there different background incidences in different (geographic) populations?

•	While benign and malignant tumors of the same cell of origin are generally evaluated
together, was there an increase only in malignant tumors?

This document is a draft for review purposes only and does not constitute Agency policy.

9-3	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table 9-1. Attributes used to evaluate studies for derivation of toxicity values

Study attributes

Considerations

Human studies

Animal studies

Study confidence

High or medium confidence studies are highly preferred over low confidence studies. The selection of low confidence studies
should include an additional explanatory justification (e.g., only low confidence studies had adequate data for toxicity value
derivation). The available high and medium confidence studies are further differentiated on the basis of the study attributes
below, as well as a reconsideration of the specific limitations identified and their potential impact on dose-response analyses.

Rationale for choice of
species

Human data are preferred over animal data to eliminate
interspecies extrapolation uncertainties (e.g., in
pharmacodynamics, dose-response pattern in relevant
dose range, relevance of specific health outcomes to
humans).

Animal studies provide supporting evidence when adequate human
studies are available, and they are considered the studies of primary
interest when adequate human studies are not available. For some
hazards, studies of particular animal species known to respond
similarly to humans would be preferred over studies of other species.

Relevance of

exposure

paradigm

Exposure
route

Studies involving human environmental exposures (oral,
inhalation).

Studies by a route of administration relevant to human
environmental exposure are preferred. A validated pharmacokinetic
or PBPK model can also be used to extrapolate across exposure
routes.

Exposure
durations

When developing a chronic toxicity value, chronic or subchronic studies are preferred over studies of acute exposure durations.
Exceptions exist, such as when a susceptible population or life stage is more sensitive in a particular time window (e.g.,
developmental exposure).

Exposure
levels

Exposures near the range of typical environmental human exposures are preferred. Studies with a broad exposure range and
multiple exposure levels are preferred to the extent that they can provide information about the shape of the
exposure-response relationship (see the EPA Benchmark Dose Technical Guidance, §2.1.1) and facilitate extrapolation to more
relevant (generally lower) exposures.

Subject selection

Studies that provide risk estimates in the most susceptible groups are preferred.

Controls for possible
confounding3

Studies with a design (e.g., matching procedures, blocking) or analysis (e.g., covariates or other procedures for statistical
adjustment) that adequately address the relevant sources of potential critical confounding for a given outcome are preferred.

This document is a draft for review purposes only and does not constitute Agency policy.

9-4	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Study attributes

Considerations

Human studies

Animal studies

Measurement of exposure

Studies that can reliably distinguish between levels of
exposure in a time window considered most relevant
for development of a causal effect are preferred.
Exposure assessment methods that provide
measurements at the level of the individual and that
reduce measurement error are preferred.
Measurements of exposure should not be influenced by
knowledge of health outcome status.

Studies providing actual measurements of exposure (e.g., analytical
inhalation concentrations vs. target concentrations) are preferred.
Relevant internal dose measures may facilitate extrapolation to
humans, as would availability of a suitable animal PBPK model in
conjunction with an animal study reported in terms of administered
exposure.

Health outcome(s)

Studies that can reliably distinguish the presence or absence (or degree of severity) of the outcome are preferred. Outcome
ascertainment methods using generally accepted or standardized approaches are preferred.

Studies with individual data are preferred in general. For example, individual data allow you to characterize experimental
variability more realistically and to characterize overall incidence of individuals affected by related outcomes (e.g., phthalate
syndrome).

Among several relevant health outcomes, preference is generally given to those outcomes with less concern for indirectness or
with greater biological significance.

Study size and design

Preference is given to studies using designs reasonably expected to have power to detect responses of suitable magnitude.15
This does not mean that studies with substantial responses but low power would be ignored, but that they should be
interpreted in light of a confidence interval or variance for the response. Studies that address changes in the number at risk
(through decreased survival, loss to follow-up) are preferred.

aAn exposure or other variable that is associated with both exposure and outcome but is not an intermediary between the two.

bPower is an attribute of the design and population parameters, based on a concept of repeatedly sampling a population; it cannot be inferred post hoc using
data from one experiment (Hoenig and Heisev, 2001).

This document is a draft for review purposes only and does not constitute Agency policy.

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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

9.3. CONDUCTING DOSE-RESPONSE ASSESSMENTS

EPA uses a two-step approach for dose-response assessment that distinguishes analysis of
the dose-response data in the range of observation from any inferences about responses at lower,
generally more environmentally relevant, exposure levels that are generally needed to develop
toxicity values ((U.S. EPA. 2012. 2005a), see Section 3):

1)	Within the observed dose range, the preferred approach is to use dose-response modeling
to incorporate as much of the data set as possible into the analysis for the purpose of
deriving a POD, see Section 9.3.1 for more details.

2)	Derivation of cancer risk estimates or reference values nearly always involves extrapolation
to exposures lower than the POD and is described in more detail in Sections 9.3.2 and 9.3.3,
respectively.

When sufficient and appropriate human data and laboratory animal data are both available
for the same outcome, human data are generally preferred for the dose-response assessment
because their use eliminates the need to perform interspecies extrapolations.

For noncancer analyses, IRIS assessments typically derive a candidate value from each
suitable data set, whether for human or animal. Evaluating these candidate values grouped within a
particular organ/system yields a single organ/system-specific reference value for each
organ/system under consideration. Next, evaluation of these organ/system-specific reference
values results in the selection of a single overall reference value to cover all health outcomes across
all organs/systems. While this overall reference value is the focus of the assessment, the
organ/system-specific reference values can be useful for subsequent cumulative risk assessments
that consider the combined effect of multiple agents acting at a common organ/system.

For cancer analyses, if there are multiple tumor types in a study population (human or
animal), final cancer risk estimates will typically address overall cancer risk.

9.3.1. Dose-Response Analysis in the Range of Observation

For conducting a dose response assessment, pharmacodynamic ("biologically based")
modeling can be used when there are sufficient data to ascertain the mode of action and
quantitatively support model parameters that represent rates and other quantities associated with
the key precursor events of the modes of action. If there is not an applicable pharmacodynamic
model available to assess health effects associated with ingestion exposure to nitrate/nitrite,
empirical dose-response modeling is used to fit the data (on the apical outcomes or a key precursor
events) in the ranges of observation. For this purpose of empirical dose-response modeling, EPA
has developed a standard set of models f http: / /www.epa.gov/ncea /bmds] that can be applied to
typical dichotomous and continuous data sets, including those that are nonlinear. In situations
where there are alternative models with significant biological support, the users of the assessment
can be informed by the presentation of these alternatives along with the models' strengths and

This document is a draft for review purposes only and does not constitute Agency policy.

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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

uncertainties. The EPA has developed guidelines on modeling dose-response data, assessing model
fit, selecting suitable models, and reporting modeling results [see the EPA Benchmark Dose
Technical Guidance fU.S. EPA. 20121].

U.S. EPA Benchmark Dose Software (BMDS) is designed to model dose-response datasets in
accordance with EPA Benchmark Dose Technical Guidance fU.S. EPA. 20121. For noncancer (and
nonlinear cancer), a benchmark dose lower confidence limit (BMDL) is computed from a model
selected from the BMDS suite of models using statistical and graphical criteria. Linear analysis of
cancer datasets is generally based on the Multistage model, with degree selected following a U.S.
EPA Statistical Workgroup technical memo available on the BMDS website

fhttps://cfpub.epa.gov/ncea/bmds/recordisplay.cfm?deid=3083821. Modeling of cancer data may
in some cases involve additional, specialized methods, particularly for multiple tumors or early
removal from observation (due to death or morbidity). Additional judgments or alternative
analyses may be used if initial modeling procedures fail to yield results in reasonable agreement
with the data. For example, modeling may be restricted to the lower doses, especially if there is
competing toxicity at higher doses. Modeling may also need to accommodate cases of nonlinear
dose-response data.

For noncancer (and nonlinear cancer) datasets, EPA recommends (1) application of a
preferred set of models that use maximum likelihood estimation (MLE) methods (default models in
BMDS) and (2) selection of a POD from a single model based on criteria designed to limit model
selection subjectivity (auto implemented in BMDS version 3 and higher). For the linear analysis of
cancer datasets, EPA recommends (1) application of the Multistage MLE model; (2) selection of a
single Multistage degree; and (3) in cases for which tumors are observed in multiple organ systems,
use of a multi-tumor model (i.e., MS-Combo) that appropriately estimates combined tumor risk
(both (2) and (3) are available in BMDS).6

Version 3.2 and higher of BMDS also provides an alternative modeling approach that uses
Bayesian model averaging for dichotomous modeling average (DMA). EPA makes DMA available as
alternative approaches but has not yet finalized guidelines for their use. DMA may be applied to
nitrate/nitrite as a supplemental analysis; see the section on Supplemental Dose-Response
Analyses below for details.

For each modeled dataset for an outcome, a POD from the observed data should be
estimated to mark the beginning of extrapolation to lower doses. The POD is an estimated dose
(expressed in human equivalent terms) near the lower end of the observed range without
significant extrapolation to lower doses. For linear extrapolation of cancer risk, the POD is used to
calculate an OSF, and for nonlinear extrapolation, the POD is used in calculating an RfD.

6The Multistage degree selection process outlined in the memo is auto-implemented in the BMDS multitumor
model, which can be run on one or more tumor data sets, but only the noncancer model selection process is
auto-implemented for individual Multistage model runs in the current version, BMDS 3.2).

This document is a draft for review purposes only and does not constitute Agency policy.

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

34

35

36

37

38

39

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

The selection of the response level at which the POD is calculated is guided by the severity
of the endpoint If linear extrapolation is used, selection of a response level corresponding to the
POD is not highly influential, so standard values near the low end of the observable range are
generally used (for example, 10% extra risk for cancer bioassay data, 1% for epidemiologic data,
lower for rare cancers). Nonlinear approaches consider both statistical and biologic considerations.
For dichotomous data, a response level of 10% extra risk is generally used for minimally adverse
effects, 5% or lower for more severe effects or effects observed in studies with increased statistical
sensitivity. Lower BMRs are often supported for developmental toxicity studies. For continuous
data, a response level is ideally based on an established definition of biologic significance. In the
absence of such definition, one control standard deviation from the control mean is often used for
minimally adverse effects, one-half standard deviation for more severe effects. As with
dichotomous endpoints, lower BMRs may also be supported for endpoints observed in studies with
greater statistical sensitivity (e.g., developmental toxicity studies). The POD is the 95% lower bound
on the dose associated with the selected response level.

EPA has developed standard approaches for determining the relevant dose to be used in the
dose-response modeling in the absence of appropriate pharmacokinetic modeling. These standard
approaches also facilitate comparison across exposure patterns and species:

•	Intermittent study exposures are standardized to a daily average over the duration of
exposure. For chronic effects, daily exposures are averaged over the lifespan. Exposures
during a critical period, however, are not averaged over a longer duration ((U.S. EPA.
2005a), §3.1.1; fU.S. EPA. 1991al. §3.2). Note that this will typically be done after
modeling because the conversion is linear.

•	Doses are standardized to equivalent human terms to facilitate comparison of results
from different species. Oral doses are scaled allometrically using mg/kg3/4-day as the
equivalent dose metric across species. Allometric scaling pertains to equivalence across
species, not across life stages, and is not used to scale doses from adult humans or
mature animals to infants or children ffU.S. EPA. 2011. 2005a), §3.1.3).

•	It can be informative to convert doses across exposure routes. If this is done, the
assessment describes the underlying data, algorithms, and assumptions ffU.S. EPA.
2005a"). §3.1.4).

•	In the absence of study specific data on, for example, intake rates or body weight, the
EPA has developed recommended values for use in dose response analysis fU.S. EPA.
19881.

•	The preferred approach for dosimetry extrapolation from animals to humans is through
PBPK modeling. Elements of more than one published model can be combined if the
effort involved is minimal and no one model has all the features desired.

•	Briefly, PBPK model simulations are used to estimate internal dose metrics
corresponding to the applied doses for each experimental animal bioassay. By
simulating the exposure scenario for each toxicity study, the resulting internal metric

This document is a draft for review purposes only and does not constitute Agency policy.

9-8	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

effectively accounts for the difference between the pattern and a nominal daily
exposure. The set of internal dose metrics for each toxicity study and endpoint can then
be used in dose-response analysis to identify a BMDL or other POD for individual animal
toxicity studies. In this assessment, the internal dose metric is either the tissue-specific
rate of oxidative metabolism or a daily average blood concentration. The human version
of the PBPK model can then be used to estimate the exposure dose that would result in
internal dose at the POD. Any remaining uncertainty factors, including the factor of 10
for human inter-individual variability (UFH) will then be applied for derivation of the
HECs.

9.3.2.	Extrapolation: Slope Factors and Unit Risk

An OSF facilitates estimation of human cancer risks when low-dose linear extrapolation for
cancer effects is supported, particularly for chemicals with direct mutagenic activity or those for
which the data indicate a linear component below the POD. Low-dose linear extrapolation is also
used as a default when the data are insufficient to establish the mode of action fU.S. EPA. 2005a). If
data are sufficient to ascertain one or more modes of action consistent with low-dose nonlinearity,
or to support their biological plausibility, low-dose extrapolation may use the reference value
approach when suitable data are available fU.S. EPA. 2005al

An inhalation unit risk (IUR) was not included in the scope for this assessment

9.3.3.	Extrapolation: Reference Values

Reference value derivation is EPA's most frequently used type of nonlinear extrapolation
method. Although it is most commonly used for noncancer effects, this approach is also used for
cancer effects if there are sufficient data to ascertain the MOA and conclude that it is not linear at
low doses. For these cases, reference values for each relevant route of exposure are developed
following EPA's established practices ffU.S. EPA. 2005al see Section 3.3.4). In general, it has been
the IRIS program's preference to base cancer reference values on key precursor events in the MOA
that are necessary for tumor formation rather than on the incidence of tumors themselves. For
example, see the ethylene glycol monobutyl ether assessment in which the cancer RfD was based on
hemosiderin deposition in the liver vs. liver tumor incidence (HEROID: 4442193).

For each data set selected for reference value derivation, reference values are estimated by
applying relevant adjustments to the PODs to account for the conditions of the reference value
definition—for human variation, extrapolation from animals to humans, extrapolation to chronic
exposure duration, and extrapolation to a minimal level of risk (if not observed in the data set).
Increasingly, data-based adjustments fU.S. EPA. 2014) and Bayesian methods for characterizing
population variability fNRC. 2014) are feasible and may be distinguished from the uncertainty
factor (UF) considerations outlined below. The assessment will discuss the scientific bases for
estimating these data-based adjustments and UFs:

• Animal-to-human extrapolation: If animal results are used to make inferences about
humans, the reference value derivation incorporates the potential for cross-species

This document is a draft for review purposes only and does not constitute Agency policy.

9-9	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

differences, which may arise from differences in pharmacokinetics or
pharmacodynamics. If available, a biologically based model that adjusts fully for
pharmacokinetic and pharmacodynamic differences across species may be used.
Otherwise, the POD is standardized to equivalent human terms or is based on
pharmacokinetic or dosimetry modeling, which may range from detailed chemical-
specific to default approaches (U.S. EPA. 2014. 20111. and a factor of 101/2 (rounded to
3) is applied to account for the remaining uncertainty involving pharmacokinetic and
pharmacodynamic differences.

•	Human variation: The assessment accounts for variation in susceptibility across the
human population and the possibility that the available data may not represent
individuals who are most susceptible to the effect, by using a data-based adjustment or
UF or a combination of the two. Where appropriate data or models for the effect or for
characterizing the internal dose are available, the potential for data-based adjustments
for pharmacodynamics or pharmacokinetics is considered fU.S. EPA. 2014. 20021.7 8
When sufficient data are available, an intraspecies UF either less than or greater than
10-fold may be justified (U.S. EPA. 20021. This factor may be reduced if the POD is
derived from or adjusted specifically for susceptible individuals [not for a general
population that includes both susceptible and non-susceptible individuals; ffU.S. EPA.
20021. 54.4.5:(TJ.S. EPA. 19981. 54.2:(TJ.S. EPA. 19961.54:(TJ.S. EPA. 19941. §4.3.9.1;(US,
EPA. 1991al.§3.41]. When the use of such data or modeling is not supported, a UF with a
default value of 10 is considered.

•	LOAEL to NOAEL: If a POD is based on a LOAEL, the assessment includes an adjustment
to an exposure level where such effects are not expected. This can be a matter of great
uncertainty if there is no evidence available at lower exposures. A factor of 3 or 10 is
generally applied to extrapolate to a lower exposure expected to be without appreciable
effects. A factor other than 10 may be used depending on the magnitude and nature of
the response and the shape of the dose-response curve fU.S. EPA. 2002.1998.1996.
1994,1991a).

•	Subchronic-to-chronic exposure: When using subchronic studies to make inferences
about chronic/lifetime exposure, the assessment considers whether lifetime exposure
could have effects at lower levels of exposure. A factor of up to 10 may be applied to the
POD, depending on the duration of the studies and the nature of the response (U.S. EPA.
2002. 1998. 19941.

•	Database deficiencies: In addition to the adjustments above, if database deficiencies
raise concern that further studies might identify a more sensitive effect, organ system,
or life stage, the assessment may apply a database UF fU.S. EPA. 2002.1998.1996.1994.
1991a). The size of the factor depends on the nature of the database deficiency. For

7Examples of adjusting the pharmacokinetic portion of interhuman variability include the IRIS boron
assessment's use of nonchemical-specific kinetic data [e.g., glomerular filtration rate in pregnant humans as a
surrogate for boron clearance (U.S. EPA. 20041] and the IRIS trichloroethylene assessment's use of population
variability in trichloroethylene metabolism, via a PBPK model, to estimate the lower 1st percentile of the dose
metric distribution for each POD fMina etal.. 20211.

8Note that when a PBPK model is available for relating human internal dose to environmental exposure,
relevant portions of this UF may be more usefully applied prior to animal-to-human extrapolation, depending
on the correspondence of any nonlinearities (e.g., saturation levels) between species.

This document is a draft for review purposes only and does not constitute Agency policy.

9-10	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	example, the EPA typically follows the recommendation that a factor of 10 be applied if

2	both a prenatal toxicity study and a two-generation reproduction study are missing and

3	a factor of 101/2 (i.e., 3) if either one or the other is missing ((U.S. EPA. 20021. §4.4.5).

4	The POD for a reference value RfV) is divided by the product of these factors. U.S. EPA

5	(20021. section 4.4.5 recommends that any composite factor that exceeds 3,000 represents

6	excessive uncertainty and recommends against relying on the associated RfV.

This document is a draft for review purposes only and does not constitute Agency policy.

9-11	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

APPENDIX A. SYSTEMATIC EVIDENCE MAP FOR
HEALTH EFFECTS OF NITRATES AND NITRITES

A.l. INTRODUCTION

This systematic evidence map (SEM) was developed based on the IRIS Assessment Plan
(IAP) developed for nitrate and nitrite. Nitrate and nitrite are considered together, as both are
chemically related and metabolically linked, and their biological effects are determined by
conversion of nitrate to nitrite and vice versa. Review of the health effect literature for both
chemicals in a single health assessment also follows the approach taken by other health agencies
CCalEPA. 2018: ATSDR. 2017: WHO. 2016: Water and Air Quality Bureau. 2013: IARC. 2010: IPCS.
20051. More specifically, this SEM includes information for the six inorganic forms of nitrate and
nitrite listed in Table 4 of the Protocol, comprising: ammonium nitrate, sodium nitrate, sodium
nitrite, potassium nitrate, potassium nitrite, and calcium nitrate. These nitrate and nitrite salts are
the most common in the environment fATSDR. 20171. These salts are highly soluble in water and
dissociate under environmental conditions; in solution, they exist as ions fATSDR. 20171. Because
the cations are not expected to introduce significant differences in the toxicity of the different salts,
toxicity findings from all six compounds are considered relevant to an assessment of nitrate and
nitrite toxicity.

A.2. METHODS

The systematic review methods used to conduct the evidence map are described in the
Protocol document and follow the Office of Research and Development (ORD) Staff Standard
Operating Procedures for Developing Integrated Risk Information System (IRIS) Assessments
(Version 2.0, referred to as the "IRIS Handbook") (U.S. EPA. 20221.

A.2.1. Specific Aims

The specific aims for the SEM are presented below:

•	Identify epidemiological (i.e., human) and toxicological (i.e., experimental animal)
literature reporting health effects of exposure to nitrates and nitrites as outlined in the
problem formulation populations, exposures, comparators, and outcomes (PECO)
criteria (shown in Table 4-2 of the Protocol).

•	Identify supplemental material as outlined in Table 4-2 of the Protocol. Supplemental
material content includes mechanistic studies; non-PECO-relevant species/model
systems; toxicokinetic and absorption, distribution, metabolism, and excretion (ADME)

This document is a draft for review purposes only and does not constitute Agency policy.

A-l	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

studies; pharmacokinetic (PK) or physiologically based pharmacokinetic (PBPK) model
studies; exposure characteristics (no health outcome); human exposure biomarker
studies with health outcome; mixture studies; routes of exposure not pertinent to the
PECO; case studies; records with no original data; and conference abstracts.

•	Create a literature inventory of PECO-relevant studies. The literature inventory
summarizes basic features of study design, and health system(s) assessed.

•	Provide an overview of the evidence base, including the degree to which it supports
conducting a formal assessment for the effects of nitrates and nitrites on the specified
health effect categories.

A.2.2. Literature Search and Screening Strategies

Survey of Existing Regulatory Toxicity Values

Toxicity value is a broad term that encompasses reference values, probabilistic risk
estimates (i.e., slope factors and unit risk estimates), and assessment-based points of departure
(PODs). The term reference value applies to values designed to provide a "benchmark" or exposure
limit from which some level of protection to human life and health can be inferred. Reference values
are the most common final output from the dose-response assessment component of the risk
assessment paradigm set forth by the National Research Council fNRC. 2009] and are based on an
observed or estimated threshold for an effect, usually noncancer.

Health-based reference values for noncancer effects are presented either in units of
concentration (e.g., mg/L) or in terms of dose (e.g., milligrams per kilogram of body weight per day,
mg/kg-day). Reference values generally are derived by applying uncertainty and adjustment factors
to the exposure/dose level that elicits an effect observed in studies with human subjects or in
controlled animal experiments, the POD. The derivation methods and factors used in moving from a
POD to a final reference value vary according to the organization developing the values, often with
consideration of how the resulting values will be applied. Oral reference values often are used as
the basis for deriving standards for drinking water or acceptable levels in food.

Probabilistic risk estimates are most often developed for cancer effects when the default
assumption is that there is no level of exposure without some effect (i.e., non-threshold effects);
however, probabilistic approaches to estimate ranges for noncancer effect levels have also been
developed fBlessinger etal.. 20201. Probabilistic risk estimates are used to determine exposure
levels associated with an acceptable risk range (e.g., less than one-in-a-million probability for risks
above background for an adverse health effect). Assessment-based PODs are identified using the
same process as used in the derivation of reference values and are used in evaluations of risk when
specific conditions of use are part of a decision process to determine exposure or consumption
levels associated with acceptable level of risk.

A visual representation was developed to illustrate the available toxicity values for oral
exposure to nitrate/nitrite (see Figure 2-1 of the Protocol). The information displayed on this
graphical array of toxicity values was collected from searches of a number of authoritative sources;

This document is a draft for review purposes only and does not constitute Agency policy.

A-2	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

these sources, cited in Appendix B, were manually searched for health risk assessments for the oral
route of exposure. In addition to these sources, the ToxVal database on the EPA Chemicals
Dashboard fhttps: //comptox.epa.gov/dashboard/chemical lists/TOXVAL V51 was searched for
reference values, risk estimate values, and PODs as described in Appendix C.

A.2.3. Literature Inventory

The literature search and screening methods are described in Section 4 of the Protocol
document Human and animal studies that met problem formulation PECO criteria after full-text
review were briefly summarized using data extraction forms in the Health Assessment Workspace
Collaborative (HAWC; hawc.epa.gov). These study summaries are referred to as literature
inventories and are used to create interactive visualizations.

For animal studies, the following information was captured: chemical assessed, study type
(acute [<24 hours], short term [1-30 days], subchronic [30-90 days], chronic [>90 days,
multigenerational, peripubertal, developmental]), duration of treatment, route, species, strain, sex,
dose, or concentration levels tested, dose units, health system and specific endpoints assessed. For
epidemiological studies, the following information was summarized: chemical assessed, population
type (e.g., general population-adult, occupational, pregnant women, infants, and children), study
type (e.g., cross-sectional, cohort, case-control), sex, major route of exposure (if known), health
system and specific outcomes assessed. Summaries were extracted into HAWC by one team
member and the extracted data were quality checked by at least one other team member.

A.3. RESULTS

A.3.1. Available Health Values

The available health values are shown in Table A-l and Figure 2-1 of the Protocol. The IRIS
program currently does not include cancer risk values for nitrate or nitrite. The International
Agency for Research on Cancer (IARC) has determined that there is "inadequate" evidence of
carcinogenicity of nitrate in food or drinking water, "limited" evidence for the carcinogenicity of
nitrite in food, and "sufficient" evidence for the carcinogenicity of nitrite in combination with
amines or amides. IARC concludes that "ingested nitrate and nitrite under conditions that result in
endogenous nitrosation is probably carcinogenic to humans (Group 2A)" flARC. 20101.

The IRIS program lists reference dose (RfD) values of 1.6 mg/kg-day for nitrate and
0.1 mg/kd-g-day for nitrite, based on a critical effect of methemoglobinemia. ATSDR has
determined minimal risk levels (MRLs) of 4 mg/kg-day for nitrate and 0.1 mg/kg-day for nitrite
(applicable for acute, intermediate, and chronic durations of oral exposure) based upon the same
health endpoint (ATSDR. 2017). The Joint FAO/WHO Expert Committee on Food Additives (JECFA)
has also determined acceptable daily intake (ADI) values of 3.7 mg/kg-day for nitrate and
0.07 mg/kg-day for nitrite (based on heart and lung effects in rats) fWHO. 2003: TECFA. 19951.

This document is a draft for review purposes only and does not constitute Agency policy.

A-3	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	The EPA's maximum contaminant levels for nitrate and nitrite are 10 mg/L (or ppm) and

2	1 mg/L (or ppm), respectively. These are equivalent to ~44 mg nitrate/L as nitrate-nitrogen and

3	~3.3 mg nitrite/L as nitrite-nitrogen. California's Office of Environmental Health Hazard

4	Assessment lists public health goals (PHGs) of 45 mg/L and 3 mg/L for nitrate and nitrite,

5	respectively (the joint nitrate/nitrite PHG is 10 mg/L) fCalEPA. 20181. The FDA uses these same

6	values for allowable levels in bottled water fFDA. 20211. and these are also the same values that

7	Health Canada has determined for maximum allowable concentration values (Water and Air Quality

8	Bureau. 2013).

This document is a draft for review purposes only and does not constitute Agency policy.

A-4	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table A-l. Details on derivation of the available health effect reference values for oral exposure to nitrate and
nitrite

Reference
value
name

Chemical
form

Duration

Reference value

Health effect

Point of
departure

Qualifier

Source

Uncertainty/
modifying
factors

Notes on
derivation

Review
status

EPA RfD
(IRIS)3

Nitrate

Chronic

1.6 mg N/kg-d

Early clinical signs of
methemoglobinemia in
infants

10 mg nitrate-
nitrogen/L

NOAEL

Bosch et
al. (1950)
and

Walton
(1951)

Total UF = 1

Dose calculated15

Final
U.S. EPA
(1991b)

Nitrite

0.1 mg N/kg-d

10 mg N/L

NOEL

Walton
(1951)

Total UF = 1
MFC = 10

Dose calculated11

Final
U.S. EPA
(1987)

EPA p-RfD
(HEAST)

Nitrite

Subchronic

0.1 mg N/kg-d

Adopted IRIS RfD









Adopted chronic
IRIS RfD for
subchronic
duration

Provisional
U.S. EPA
(1997)

EPA RfD
(OW)

Nitrate

Chronic

1.6 mg N/kg-d

Methemoglobin
concentration in
infants >10%

1.6 mg
nitrate-
nitrogen/kg-d

NOAEL

Bosch et
al. (1950)
and

Walton
(1951)

Total UFe = 1

WOE approach

Final
U.S. EPA
(1990)

Nitrite

0.16 mg N/kg-d

Based on nitrate RfD









RfD adjusted'

ATSDR MRL

Nitrate

Acute (1-14 d)

4 mg NOs/kg-d

Methemoglobinemia in
infants due to nitrate-
contaminated water

44 mg/L

NOAEL

Walton
(1951)

Total UF = 1
UFh = 1

Dose calculated5

Final
ATSDR
(2017)

Intermediate
(15-365 d)

4 mg NOs/kg-d

This document is a draft for review purposes only and does not constitute Agency policy.

AS	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Reference
value
name

Chemical
form

Duration

Reference value

Health effect

Point of
departure

Qualifier

Source

Uncertainty/
modifying
factors

Notes on
derivation

Review
status





Chronic (>1 y)

4 mg NOs/kg-d











No duration
adjustment11



Nitrite

Acute (1-14 d)

0.1 mg N02/kg-d

0.2 mg/kg-d

NOAEL

Total UF = 1
UFh = 1
MF' = 2

Dose calculated'

Intermediate
(15-365 d)

0.1 mg N02/kg-d

Chronic (>1 y)

0.1 mg N02/kg-d

JECFAADI

Nitrate

Chronic

3.7 mg NOs/kg-d

No effects noted in
rats

370 mg/kg-d

NOEL

Speiiers et
al. (1989)

Total UF = 100

Derived values
not protective of
infants below
the age of 3 mo

Final
JECFA
(1995)
and WHO
(2003)

Nitrite

0.06 mg N02/kg-d

Hypertrophy of the
adrenal zona
glomerulosa in rats
exposed for 90 d

5.4 mg/kg-d

NOEL

Til et al.
(1988) and
Kuper F
(1995)

Methemoglobin
formation, dilated
bronchi and arteries,
lymphocyte
infiltration, and
alveolar

hyperinflation in rats

6.7 mg/kg-d

NOEL

Speiiers

et al.
(1989)

SCF ADI

Nitratek

Chronic

3.7 mg NOs/kg-d

No toxicity in rats

2,500 mg
NaNOs/kg-d

NOEL

Maekaw

Total UF = 500

MW adjustment1

Final

CEC (1992)
and SCF
(1997)

a et al.

(1982)

Nitrite

0.06 mg N02/kg-d

Hypertrophy of the
adrenal zona
glomerulosa in the rat

5.4 mg/kg-d

NOEL

Til et al.
(1988)

and

Kuper F
(1995)

Total UF = 100

NA

This document is a draft for review purposes only and does not constitute Agency policy.

A-6	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Reference
value
name

Chemical
form

Duration

Reference value

Health effect

Point of
departure

Qualifier

Source

Uncertainty/
modifying
factors

Notes on
derivation

Review
status









Histological changes in
the lung and heart of
rats

6.7 mg/kg-d

NOEL

Speiiers







et al.
(1989)

EFSA ADI

Nitrite

Chronic

0.07 mg N02/kg-d

Increased

methemoglobin levels

9.63 mg
NaN02/kg-d

BMDL

NTP
(2001a)

Total UF = 100

MW

adjustment"1

FinaK EFSA)



aThe IRIS RfDs have been adopted by NDEP, TCEQ, and MDEQ (TCEQ, 2023; NDEP.2020; Michigan DEQ. 2015).
bDose = NOAEL x water intake -f BW = 10 mg/L x 0.64 L/day -f 4 kg = 1.6 mg/kg-d.

CIRIS documentation states: "A modifying factor of 10 was applied because of the direct toxicity of nitrite."

dDose = NOEL x water intake 4 BW = 10 mg/L x 1 L/day 4 10 kg = 1.0 mg/kg-d.

eNo uncertainty factor is required since the POD is a NOAEL based on a sensitive subpopulation.

fN02 RfD = NO3 RfD x conversion factor = 1.6 mg nitrate-nitrogen/kg-d x 0.1 mg nitrite-nitrogen/ mg nitrate-nitrogen = 0.16 mg nitrite-nitrogen/kg-d.
gDose = NOAEL x water intake 4 BW = 44 mg/L x 0.525 L/day 4 5.33 kg = 4.33 mg/kg-d.

hThe toxicological profile states: "Repeated ingestion for intermediate- or chronic-duration time periods would be expected to result in changes in
methemoglobin levels similar to those elicited from a single exposure."

'A modifying factor is applied due to the increased susceptibility of infants to methemoglobinemia.

JN02dose = NOsdose x 0.05 = 4 mg/kg-d x 0.05 = 0.2 mg/kg-d. "The ingestion of 0.2 mg nitrite/kg/day by an adult would be expected to result in a nitrite blood
level similar to that achieved following ingestion of 4 mg nitrate/kg/day" (ATSDR, 2017).
kEFSA concurs with the nitrate ADI established by the Scientific Committee for Food (EFSA) (2017a).

'ADI = NOEL4 UF x NOs MW 4 NaNOs MW = 2,500 mg/kg-d 4 500 x 62 g/mol 4 85 g/mol = 3.7 mg/kg-d.
mADI = BMDL 4 UF x NO2 MW 4 NaNOz MW = 9.63 mg/kg-d 4 100 x 46 g/mol 4 69 g/mol = 0.07 mg/kg-d.

ADI = acceptable daily intake; ATSDR = Agency for Toxic Substances and Disease Registry; BMDL = benchmark dose level; BW = body weight; CEC = Commission
of the European Communities; EFSA = European Food Safety Authority; EPA = U.S. Environmental Protection Agency; HEAST = Health Effects Assessment
Summary Table; IRIS = Integrated Risk Information System; JECFA = Joint FAO/WHO Expert Committee on Food Additives; MDEQ = Michigan Department of
Environmental Quality; MF = modifying factor; MRL = minimal risk level; MW = molecular weight; NaN02 = sodium nitrate; NaNOs = sodium nitrate;

NDEP = Nevada Division of Environmental Protection; NO2 = nitrite; NO3 = nitrate; NOAEL = no-observed-adverse-effect level; NOEL = no-observed-effect
level; NTP = National Toxicology Program; OW = Office of Water; RfD = reference dose; SCF = Scientific Committee for Food; TCEQ = Texas Commission on
Environmental Quality; UF = uncertainty factor; UFh = inter-human variability; WHO = World Health Organization; WOE = weight of evidence.

This document is a draft for review purposes only and does not constitute Agency policy.

A-7	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

A.3.2. Literature Screening Results

The flow of studies for nitrate/nitrite during the screening process is summarized in Figure
A-l and available in an interactive format in a HAWC literature tree. The database searches yielded
73,395 unique records. Application of the SWIFT Review filters (human, animal/human health
models, and in vitro) reduced the number of studies for TIAB screening to 18,495. After TIAB
screening, 5,549 studies were excluded as not PECO relevant and another 1,080 were tagged as
supplemental material, leaving 557 studies that advanced to full-text screening. The remaining
11,374 studies were identified by the SWIFT-AS machine learning algorithm as not relevant. The
supplemental literature search yielded an additional 56 studies from other sources for a total of
613 studies that were considered for full-text screening.

The studies identified for full-text screening were processed in DistillerSR. Of these, 65 were
excluded as not meeting PECO criteria, text was unable to be obtained for 4, and 166 were tagged as
supplemental material. A total of 391 studies were considered PECO relevant, of which 244 were
human studies (178 human randomized controlled trials and 66 human observational studies) and
148 were animal studies (one study evaluated health endpoints in both animals and humans).

This document is a draft for review purposes only and does not constitute Agency policy.

A-8	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Literature Search (Jan 2016 - Aug 2022)

Figure A-l, Nitrate/ nitrite literature flow diagram.

A.3.3. Characterizing Animal and Epidemiological Studies

1	Human Studies

2	Literature Inventory

3	A survey of study designs and health systems assessed in the human studies that met PECO

4	criteria and tabular summary of study design and findings is provided in Figure A-2. Among the

5	244 human studies, there were 178 randomized controlled trials that administered controlled

6	quantities of oral nitrate or nitrite to identify potential health benefits; these studies were identified

This document is a draft for review purposes only and does not constitute Agency policy,

A-9	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

and inventoried but will be considered supplemental material as the focus of this work is on
potential adverse health effects due to exposure. The literature search also identified
66 observational epidemiology studies (n = 11 case-control, 2 nested case-control, 5 cross-
sectional, 8 ecological, and 40 cohort) in which nitrate/nitrite exposure was evaluated using
measurement in drinking water and/or food.

Cancer

10

12



5

1

28

Cardiovascular



13







13

Developmental



6







6

Endocrine





3

1



4

Gastrointestinal







1



1

Hematologic





3





3

Hepatic



1







1

Immune



2



1



3

Metabolic



3

1





4

Multi-System



3



1



4

Nervous



4







4

Ocular



1

1



1

3

Reproductive

1

5







6

Respiratory



1







1

Urinary



2







2

Whole Body



1

1





2

Grand Total

11

40

5

8

2

66



Case-control

Cohort

Cross-sectional

Ecological

Nested
case-control

Grand Total

Figure A-2. Survey of human studies that met PECO criteria summarized by
study design and health systems assessed.

This is a thumbnail image of the interactive dashboard. The numbers in the heat map inset indicate the number of
studies that investigated a health system within a study design. If a study evaluated multiple health outcomes, it
is shown here multiple times.

Animal Studies

Literature Inventory

A preliminary survey of study designs, species, form(s) of nitrate/nitrite evaluated, and
health effects evaluated in the animal studies that met PECO criteria is provided in Figure A-3. The
animal studies evaluated exposure to ammonium nitrate, potassium nitrate, sodium nitrate,
sodium nitrite, and mixed or unspecified forms of nitrate/nitrite. There were 148 animal studies
meeting PECO criteria, and many measured health endpoints in multiple categories. The number of
studies for each health effect category shown in the heatmap may be larger than reported in section
5.1 due to the inclusion of additional endpoints (e.g., mRNA expression) along with those used to
determine 'primary' health effect categories informed by each study. Most studies were conducted
in rats and mice, but data were also available from one study of rabbits. Among the 148 studies, 27
studies administered multiple doses; in general, these study designs are preferred for toxicity
value derivation over acute/short-term studies or studies that test a single dose level

A-10	DRAFT-DO NOT CITE OR QUOTE

This document is a draft for review purposes only and does not constitute Agency policy.


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

(U.S. EPA. 20021. although there may be circumstances for which other study designs are more
suitable.

Cardiovascular



7

2

1

2

1

24

27

60

Dermal













2



2

Developmental



1











3

4

Endocrine



3

2

1

1

1

21

16

45

Gastrointestinal













9

7

15

Hematologic

2

3



1



1

12

7

23

Hepatic

1

2



1





11

13

28

Immune



2









11

13

25

Metabolic

2

4

2

1

1

2

39

37

87

Multi-System













3

4

7

Musculoskeletal



1



1





9

1

12

Nervous

1

1





2



9

3

15

Ocular













1

1

1

Reproductive





1



2



8

6

17

Respiratory



3





1



2

3

9

Urinary

1

1









13

8

22

Whole Body

1

6

2

1

2



39

20

71

Grand Total

3

11

2

1

6

2

70

60

148



Ammonium
Nitrate

Nitrate

Nitrate/Nitrite

Nitrates (Mixed)

Nitrite

Potassium
Nitrate

Sodium Nitrate

Sodium Nitrite

Grand Total

Figure A-3. Survey of animal studies that met PECO criteria by form of
nitrate/nitrite administered and health systems.

This is a thumbnail image of the interactive dashboard that is filterable by health system, form of nitrate/nitrite
administered, and species. The numbers in the heat map inset indicate the number of studies that investigated a
health system within form of nitrate/nitrite administered. If a study evaluated multiple health outcomes or
presented several experiments, it is shown here multiple times.

Mechanistic Evidence

Results from Database Search

There were 86 mechanistic studies tagged as supplemental material. Among these, the
largest numbers of studies evaluated aspects of oxidative and nitrosative stress and hypoxia
(n = 37); modulation of enzyme activity (n = 25); or nitric oxide mediated cell signaling (n = 21).
Fewer (<20 studies) evaluated other mechanistic characteristics.

ToxCast and Tox21 High Throughput Screening Data

ToxCast and Tox21 high throughput screening data are available for each of the six forms of
nitrate/nitrite considered here:

Sodium nitrate: (link)

Sodium nitrite: (link)

Potassium nitrate: (link)

This document is a draft for review purposes only and does not constitute Agency policy.

A-ll	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Potassium nitrite: (link)

Ammonium nitrate: (link)

Calcium nitrate: (link)

Comparative Toxicogenomics Database

Nitrate and nitrite are included in the Comparative Toxicogenomics Database (CTDB).
Below is a summary of the top interacting genes based on analysis of 257 and 150 studies
presented in the CTDB, respectively (click here to see the entry for nitrates, and here to see the
entry for nitrites, in the CTDB). Note, these studies were reviewed to identify any that were not
otherwise retrieved from other sources (see Appendix D).

A.4. CONCLUSIONS

The SEM used systematic review methods to identify PECO-relevant studies published from
2016-2022 (no date restriction for calcium nitrate) for six specified forms of nitrate/nitrite. There
were 214 animal and human studies which evaluated effects of oral exposure to nitrate/nitrite,
comprising 148 animal studies and 66 observational human studies. The animal studies and
observational human studies, along with previously published studies as characterized in the
ATSDR Toxicological Profile fATSDR. 2017] and supporting information from the identified
supplemental material including mechanistic and ADME information, should be sufficient to
support hazard determination for the following health effect categories: cancer; cardiovascular;
developmental; endocrine; hematopoietic; hepatic; metabolic; nervous; reproductive; urinary.

This document is a draft for review purposes only and does not constitute Agency policy.

A-12	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

APPENDIX B. SURVEY OF EXISTING TOXICITY
VALUES

Table B-l lists websites which are searched for relevant human health reference values,
along with indications of the results of the search. In addition to these sources, the ToxVal database
on the Chemicals Dashboard fhttps://comptox.epa.gov/dashboard/chemical lists/TOXVAL V5] is
searched for both reference values and PODs as described in Appendix D.

Table B-l. Sources searched for existing human health reference values

Source3

Query and/or link

ATSDR

http://www. atsdr.cdc.gov/toxDrofiles/index.asD

httos://www. atsdr.cdc.gov/mrls/mrllist.asD

CalEPA

htto://www. oehha.ca.gov/tcdb/index.asD

httos://www. arb.ca.gov/toxics/healthval/healthval. htm

DWSHA

httDs://www.eDa.gov/sites/Droduction/files/2018-03/documents/dwtable2018.Ddf

Health Canada

httDs://www.canada.ca/en/services/health/Dublications/healthv-living.html

http://publications.gc.ca/site/archivee-

archived.html?url=httD://Dublications.gc.ca/collections/collection 2012/sc-hc/H 128-1-11-

638-eng.odf

http://publications.gc.ca/site/archivee-

archived.html?url=httD://Dublications.gc.ca/collections/Collection/H46-2-96-194E.Ddf

HEAST

httD://eDa-heast.ornl.gov/heast.DhD

httDs://neDis.eDa.gov/Exe/ZvPDF.cgi/200000GZ.PDF?Dockev=200000GZ.PDF

IRIS

htto://www. eoa.gov/iris/

Ml EGLE

https://www.michigan.gov/documents/dea/dea-rrd-chem-

CleanuoCriteriaTSD 527410 7.pdf

MDH

https://www.health.state.mn.us/communities/environment/risk/guidance/gw/table.html

NHMRC

https://www.nhmrc.gov.au/about-us/publications/australian-drinking-water-guidelines

NY DEC

https://www.dec.nv.gov/docs/remediation hudson pdf/techsuppdoc.pdf

OPP

https://iaspub.epa.gov/apex/pesticides/f?p=chemicalsearch:l

PPRTV

https://www.epa.gov/pprtv/provisional-peer-reviewed-toxicitv-values-pprtvs-assessments

RIVM

https://www.rivm.nl/bibliotheek/rapporten/711701092.pdf

https://www.rivm.nl/bibliotheek/rapporten/711701025.pdf

This document is a draft for review purposes only and does not constitute Agency policy.

B-l	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Source3

Query and/or link

TCEQ

https://www. tcea.texas.gov/remediation/trrD/trrDDcls. html

WHO

httD://www.who.int/iDcs/Dublications/ehc/en/

aATSDR = Agency for Toxic Substances and Disease Registry; CalEPA = California Environmental Protection Agency;
DWSHA = Drinking Water Standards and Health Advisories; HEAST = Health Effects Assessment Summary Tables;
IRIS = Integrated Risk Information System; MDH = Minnesota Department of Health; Ml EGLE = Michigan
Department of Environment, Great Lakes & Energy; NHMRC = National Health and Medical Research Council; NY
DEC = New York State Department of Environmental Conservation; OPP = Office of Pesticide Programs;

PPRTV = Provisional Peer-Reviewed Toxicity Values; RIVM = Rijksinstituut voor Volksgezondheid en Milieu, The
Netherlands Institute for Public Health and the Environment; TCEQ = Texas Commission on Environmental
Quality; WHO = World Health Organization.

This document is a draft for review purposes only and does not constitute Agency policy.

B-2	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

APPENDIX C. LITERATURE SEARCH STRATEGIES

Table C-2. Results of initial literature search

Database

Search terms

Number of
citations3

Web of
Science
(WoS)

Dates
covered:
1/1/2018-
8/17/2022

Search
date:

8/17/2022

TS=("14797-55-8" OR "14797-65-0" OR "13446-48-5" OR "7631-99-4" OR "7632-00-
0" OR "7758-09-0" OR "7757-79-1" OR "6484-52-2" OR "6484-52-2" OR "nitrate"
OR "nitrates" OR "nitrite" OR "nitrites" OR "sodium nitrate" OR "sodium nitrates"
OR "sodium nitrite" OR "sodium nitrites" OR "potassium nitrate" OR "potassium
nitrates" OR "potassium nitrite" OR "potassium nitrites" OR "ammonium nitrate" OR
"ammonium nitrates") AND PY=(2018-2022)

48,417

Web of
Science
(WoS)

Dates
covered:
1/1/2016-
12/31/2017

Search
date:

1/25/2023

TS=("14797-55-8" OR "14797-65-0" OR "13446-48-5" OR "7631-99-4" OR "7632-00-
0" OR "7758-09-0" OR "7757-79-1" OR "6484-52-2" OR "6484-52-2" OR "13477-34-
4" OR "10124-37-5" OR "nitrate" OR "nitrates" OR "nitrite" OR "nitrites" OR "sodium
nitrate" OR "sodium nitrates" OR "sodium nitrite" OR "sodium nitrites" OR
"potassium nitrate" OR "potassium nitrates" OR "potassium nitrite" OR "potassium
nitrites" OR "ammonium nitrate" OR "ammonium nitrates" OR "calcium nitrate")
AND PY=(2016-2017)

16,681

PubMed

Dates
covered:
1/1/2018-
8/17/2022

Search
date:

8/17/2022

(Updated

on

8/29/2023)

((" 14797-55-8"[tw] OR " 14797-65-0"[tw] OR " 13446-48-5"[tw] OR "7631-99-
4"[tw] OR "7632-00-0"[tw] OR "7758-09-0"[tw] OR "7757-79-l"[tw] OR "6484-52-
2"[tw] OR "6484-52-2"[tw] OR "nitrate"[tw] OR "nitrates"[tw] OR "nitrite"[tw] OR
"nitrites"[tw] OR "sodium nitrate"[tw] OR "sodium nitrates"[tw] OR "sodium
nitrite"[tw] OR "sodium nitrites"[tw] OR "potassium nitrate"[tw] OR "potassium
nitrates"[tw] OR "potassium nitrite"[tw] OR "potassium nitrites"[tw] OR
"ammonium nitrate"[tw] OR "ammonium nitrates"[tw]) AND ("2018"[Date -
Publication] : "3000"[Date - Publication]))

22,172

PubMed

((" 14797-55-8"[tw] OR " 14797-65-0"[tw] OR " 13446-48-5"[tw] OR "7631-99-4"[tw]
OR "7632-00-0"[tw] OR "7758-09-0"[tw] OR "7757-79-l"[tw] OR "6484-52-2"[tw] OR

8,393

This document is a draft for review purposes only and does not constitute Agency policy.

C-l	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Database

Search terms

Number of
citations3

Dates
covered:
1/1/2016-
12/31/2017

Search
date:

1/25/2023

"6484-52-2"[tw] OR " 13477-34-4"[tw] OR "10124-37-5"[tw] OR "nitrate"[tw] OR
"nitrates"[tw] OR "nitrite"[tw] OR "nitrites"[tw] OR "sodium nitrate"[tw] OR "sodium
nitrates"[tw] OR "sodium nitrite"[tw] OR "sodium nitrites"[tw] OR "potassium
nitrate"[tw] OR "potassium nitrates"[tw] OR "potassium nitrite"[tw] OR "potassium
nitrites"[tw] OR "ammonium nitrate"[tw] OR "ammonium nitrates"[tw] OR "calcium
nitrate") AND ("2016"[Date - Publication] : "2017"[Date - Publication]))



TOXNET

Dates
covered:
1/1/2016-
12/05/2017

@SYNO+@AND+@OR+(nitrate+nitrates+nitrite+nitrites+@TERM+@rn+14797-55-
8+@TERM+@rn+14797-65-0+@TERM+@rn+7631-99-4+@TERM+@rn+7757-79-
l+@TERM+@rn+6484-52-2+@TERM+@rn+7632-00-0+@TERM+@rn+7758-09-
0)+@RANGE+yr+2016+2017+@NOT+@org+"nih+reporter"

@SYNO+@AND+@OR+(nitrate+nitrates+nitrite+nitrites+@TERM+@rn+14797-55-
8+@TERM+@rn+14797-65-0+@TERM+@rn+7631-99-4+@TERM+@rn+7757-79-
l+@TERM+@rn+6484-52-2+@TERM+@rn+7632-00-0+@TERM+@rn+7758-09-
0)+@RANGE+yr+2017+2017+@NOT+@org+"nih+reporter"

1,992

TOTAL:

Merged reference sets (After removal of duplicates)

73,395

aThe numbers in this document are current as of October 6, 2023, but are subject to slight changes due to ongoing
deduplication efforts.

This document is a draft for review purposes only and does not constitute Agency policy.

C-2	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

APPENDIX D. PROCESS FOR SEARCHING AND
COLLECTING EVIDENCE FROM SELECTED OTHER
RESOURCES

D.l. REVIEW OF REFERENCE LISTS FROM EXISTING ASSESSMENTS
(FINAL OR PUBLICLY AVAILABLE DRAFT), JOURNAL REVIEW
ARTICLES, AND STUDIES CONSIDERED RELEVANT TO PECO BASED ON
FULL-TEXT SCREENING

Review of the citation reference lists is typically done manually because they are not
available in a file format (e.g., RIS) that permits uploading into screening software applications.
Manual review entails scanning the title, study summary, or study details as presented in the
resource for those that appear to meet the PECO criteria. Any records identified that were not
identified from the other sources are annotated with respect to source and screened as outlined in
Section 3.2.

D.2. EUROPEAN CHEMICALS AGENCY

A search of the ECHA registered substances database was conducted using the chemical
names. The registration dossier associated with the chemical name was retrieved by navigating to
and clicking the eye-shaped view icon displayed in the chemical summary panel. The general
information page and all subpages included under the Toxicological Information tab were reviewed
to identify any human or animal health effects information from 2016 onwards that would be
eligible for inclusion based on PECO criteria.

D.3. EPA CHEMVIEW

The EPA ChemView database fU.S. EPA. 2019al using the chemical CASRN was searched.
The prepopulated CASRN match and the "Information Submitted to EPA" output option filter were
selected before generating results. If results were available, the square-shaped icon under the "Data
Submitted to EPA" column was selected, and the following records were included:

•	High Production Volume Challenge Database (HPVIS)

•	Human Health studies (Substantial Risk Reports)

•	Monitoring (Includes environmental, occupational, and general entries)

•	TSCA Section 4 (Chemical testing results)

This document is a draft for review purposes only and does not constitute Agency policy.

D-l	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

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

•	TSCA Section 8(d) (Health and safety studies)

•	TSCA Section 8(e) (Substantial risk)

•	FYI (Voluntary documents)

All records for ecotoxicology and physical and chemical property entries were excluded.
When results were available, extractors navigated into each record until a substantial risk report
link was identified and saved as a PDF file. If the report could not be saved, due to file corruption or
broken links, the record was excluded during full-text review as "unable to obtain record." Most
substantial risk reports contained multiple document IDs, so citations were derived by
concatenating the unique report numbers (OTS; 8EHD Num; DCN; TSCATS RefID; and CIS)
associated with each document along with the typical author organization, year, and title. Once a
citation was generated, the study moved forward to DistillerSR with which it was screened
according to PECO and supplemental material criteria.

D.4. NTP CHEMICAL EFFECTS IN BIOLOGICAL SYSTEMS

This database is searched using the chemical CASRN
(https://manticore.niehs.nih.gov/cebssearch). All non-NTP data were excluded using the "NTP
Data Only" filter. Data tables for reports undergoing peer review are also searched for studies that
have not been finalized (https: //ntp.niehs.nih.gov/data/tables/index.html) based on a manual
review of chemical names.

D.5. ECOTOX DATABASE

EPA's ECOTOX Knowledgebase (https: //cfpub.epa.gov/ecotox/search.cfm) was searched
using the chemical names. Results were refined to terrestrial mammalian studies by selecting the
terrestrial tab at the top of the search page and sorting the results by species group. Results were
reviewed to verily that it was not already identified from the database search (or searches of "other
sources consulted") search prior to moving forward to screening.

D.6. EPA COMPTOX CHEMICAL DASHBOARD VERSION TO RETRIEVE A
SUMMARY OF ANY TOXCAST OR T0X21 HIGH THROUGHPUT
SCREENING INFORMATION

Version 3.0.9 of the CompTox Chemicals Dashboard (U.S. EPA. 2019b) was accessed for
high throughput screening (HTS) data by searching the Dashboard by CASRN. Next, the
"Bioactivity" section was selected and the availability of ToxCast/Tox21 HTS data for active and
inactive assays was examined in the "TOXCAST: Summary" tab. If active assays were reported, the
figure was copied for presentation in the SEM. This figure presents (i) scatterplot of scaled assay
responses vs. AC50 values for each active assay endpoint, and (ii) cytotoxicity limit as a vertical line.

This document is a draft for review purposes only and does not constitute Agency policy.

D-2	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

1	More detailed information on the results of ToxCast and Tox21 assays are available in the CompTox

2	Chemicals Dashboard section "ToxCast/Tox21," which includes chemical analysis data, dose-

3	response data and model fits, and "flags" assigned by an automated analysis, which might suggest

4	false positivity/negativity or indicate other anomalies in the data. This information is not

5	summarized further for the purposes of the SEM, which is focused on identifying the extent of

6	available evidence.

D.7. COMPARATIVE TOXICOGENOMICS DATABASE

7	This CTDB database fhttp: //ctdbase.org/] was searched using the chemical names in the

8	"keyword search" with pulldown menu set to "Chemicals." The reference list of studies reporting

9	gene/protein interactions with the query chemical were compared to existing references in HAWC.
10 Unique references screened according to PECO and supplemental material criteria.

This document is a draft for review purposes only and does not constitute Agency policy.

D-3	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Table D-l. Summary table for other sources search results

Source3

Source address

Search terms

Search date

Total unique number
of results not already
identified in
literature search

Records
found to be
PECO-
relevant

Review of reference
lists from existing
assessments (final or
publicly available draft)
or journal review
articles that focused on
human health

OEHAA 2018; EFSA 2017 (Sodium
nitrate); EFSA 2017 (Sodium nitrite);
various review articles

NA

NA

21

5

EPA CompTox
(Computational
Toxicology Program)
Chemicals Dashboard
(ToxVal)



Results from human health,
oral/ingestion route of exposure:
pod, toxicity value, lethality effect
level

5/25/2023







Nitrate:

https://comptox. epa.gov/dashboard/
chemical/hazard/DTXSID5024217

ATSDR MRL; IRIS NOAEL; RSL RfD;
IRIS RfD; OW RfD



0

0



Nitrite ion:

https://comptox.epa.gov/dashboard/
chemical/hazard/DTXSID5024219

ATSDR MRL; IRIS NOEL; HEAST
NOEL; RSL RfD; IRIS RfD; OW RfD;
DOD MEG; HEAST RfD



0

0



Sodium nitrate:

https://comptox.epa.gov/dashboard/
chemical/details/DTXSID6020937

COSMOS HNEL; COSMOS LEL; ECHA
IUCLID NOAEL; ChemlDplus LD50;
ECHA IUCLID LD50



0

0



Sodium nitrite:

https://comptox.epa.gov/dashboard/
chemical/hazard/DTXSID0020941

COSMOS LEL; COSMOS HNEL; HESS
NOEL; ECHA IUCLID NOEL; ECHA
IUCLID NOAEL; ECHA IUCLID LOAEL;
EFSA BMDL; ChemlDplus LD50;
ECHA IUCLID LD50



0

0

This document is a draft for review purposes only and does not constitute Agency policy.

D-4	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Source3

Source address

Search terms

Search date

Total unique number
of results not already
identified in
literature search

Records
found to be
PECO-
relevant





ToxRefDB LEL, NEL, LOAEL, NOAEL
(based on NTP 2001 report)









Potassium nitrate:

https://comptox. eDa.gov/dashboard/
chemical/details/DTXSID4029692

DOE Wildlife Benchmark; COSMOS
HNEL; COSMOS LEL; ECHA IUCLID
NOAEL; ChemlDplus LD50; ECHA
IUCLID LD50



0

0



Potassium nitrite:

https://comptox. epa.gov/dashboard/
chemical/hazard/DTXSID5042320

ECHA IUCLID NOEL; ECHA IUCLID
NOAEL; ECHA IUCLID LOAEL;
ChemlDplusLD50; ECHA IUCLID
LD50



0

0



Ammonium nitrate:

https://comptox.epa.gov/dashboard/

chemical/hazard/DTXSID2029668

ECHA IUCLID NOAEL; ChemlDplus
LD50; ECHA IUCLID LD50



0

0



Calcium nitrate:

https://comptox.epa.gov/dashboard/
chemical/hazard/DTXSID1039719

ECHA IUCLID NOAEL; ChemlDplus
LD50



0



ECHA, Chemical
Registration Dossiers





5/26/2023







Sodium nitrate:

https://echa.europa.eu/registration-

dossier/-/registered-

dossier/15423/1/1

EC number: 231-554-3



0

0



Sodium nitrite:

https://echa.europa.eu/registration-
dossier/-/registered-dossier/14890

EC number: 231-555-9



0

0

This document is a draft for review purposes only and does not constitute Agency policy.

D-5	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Source3

Source address

Search terms

Search date

Total unique number
of results not already
identified in
literature search

Records
found to be
PECO-
relevant



Potassium nitrate:

https://echa.europa.eu/registration-
dossier/-/registered-dossier/15481

EC number: 231-818-8



0

0



Potassium nitrite: No dossier
available

EC number: 231-832-4



0

0



Ammonium nitrate:

https://echa.europa.eu/registration-

dossier/-/registered-dossier/15999

EC number: 229-347-8



0

0



Calcium nitrate:

https://echa.europa.eu/registration-
dossier/-/registered-dossier/15487

EC number: 233-332-1



1

0

EPA ChemView

https://chemview.epa.gov/chemview

Nitrate; nitrite; nitrite ion;
potassium nitrate; potassium
nitrite; sodium nitrate; sodium
nitrite; ammonium nitrate; calcium
nitrate

5/25/2023

0

0

?tf=0&ch=14797-55-8 10124-37-

5 13477-34-4 14797-55-8 7631-99-

4 7757-79-1 14797-65-0 7632-00-

0 7758-09-

0&su=256737574985&as=31098&ac=

115166378999&ma=4-ll-1981377-

4 16848473-4 16848474-

4 49007566&gs=&tds=0&tdl=100&ta
sl=l&tas2=asc&tas3=undefined&tss

~

NTPCEBS

https://manticore.niehs.nih.gov/cebs
search/



5/24/2023







https://cebs.niehs.nih.gov/cebs/test
article/7631-99-4

Sodium nitrate: Only test article
purity



0

0



https://cebs.niehs.nih.gov/cebs/test
article/7632-00-0

Sodium nitrite: Link to NTP 2001
study



0

0

This document is a draft for review purposes only and does not constitute Agency policy.

D-6	DRAFT-DO NOT CITE OR QUOTE


-------
Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Source3

Source address

Search terms

Search date

Total unique number
of results not already
identified in
literature search

Records
found to be
PECO-
relevant



https://cebs. niehs.nih.gov/cebs/test
article/7757-79-1

Potassium nitrate: Only test article
purity



0

0

OECD Echem Portal

httpsi//hpvchemicals, oecd.org/UI/Se
arch.aspx

Potassium nitrate; potassium
nitrite; sodium nitrate; sodium
nitrite; ammonium nitrate; calcium
nitrate

5/24/2023

0

0

ECOTOX Database

httpsi//cfpub. epa.gov/ecotox/search.

cfm

Nitrate; nitrite (terrestrial,
mammalian studies only)

5/24/2023

0

0

Comparative
Toxicogenomics
Database (CTDB)

http://ctdbase.org/

Nitrate; nitrite; potassium nitrate;
potassium nitrite; sodium nitrate;
sodium nitrite; ammonium nitrate;
calcium nitrate

5/25/2023

39

0

TOTAL (after de-
duplication)







56

4

aPECO = populations, exposures, comparators, and outcomes; NA = not applicable; POD = point of departure; ECHA = European Chemicals Agency; NTP
CEBS = National Toxicology Program Chemical Effects in Biological Systems; OECD = Organisation for Economic Co-operation and Development.

This document is a draft for review purposes only and does not constitute Agency policy.

D-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

34

35

36

37

38

39

40

41

42

43

44

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

REFERENCES

ATSDR (Agency for Toxic Substances and Disease Registry). (2017). Toxicological profile for nitrate
and nitrite [ATSDRTox Profile], Atlanta, GA: U.S. Department of Health and Human Services,
Public Health Service, https://www.atsdr.cdc.gov/toxprofiles/tp204.pdf.

Bannach-Brown. A: Przybvla. P: Thomas. 1: Rice. ASC: Ananiadou. S: Liao. 1: Macleod. MR. (2018).
Machine learning algorithms for systematic review: reducing workload in a preclinical
review of animal studies and reducing human screening error (pp. 1-26). bioRxiv.
http://dx.doi.org/10.1101/255760.

Blessinger. T: Davis. A: Chiu. WA: Stanek. 1: Woodall. GM: Gift. 1: Thayer. KA: Bussard. D. (2020).
Application of a unified probabilistic framework to the dose-response assessment of
acrolein. Environ Int 143:105953. http://dx.doi.Org/10.1016/i.envint.2020.105953.

Bosch. HM: Rosenfield. AB: Huston. R: Shipman. HR: Woodward. FL. (1950). Methemoglobinemia
and Minnesota well supplies. J Am Water Works Assoc 42: 161-170.

Bryan. NS: Loscalzo. 1. (2011). Nitrite and nitrate in human health and disease. In NS Bryan; J

Loscalzo (Eds.). New York, NY: Humana Press. http://dx.doi.org/10.1007/978-l-6Q761-
616-0.

Bryan. NS: Petrosino. IF. (2017). Nitrate-reducing oral bacteria: Linking oral and systemic health.
Chapter 3. In NS Bryan; J Loscalzo (Eds.), Nitrate and nitrite in human health and disease
(2nd ed.). Boston, MA: Humana Press.

CalEPA (California Environmental Protection Agency). (2018). Public health goals for nitrate and
nitrite in drinking water. Sacramento, CA: Office of Environmental Health Hazard
Assessment

https://oehha.ca.gov/media/downloads/water/chemicals/phg/nitratephg051118.pdf.

CEC (Commission of the European Communities). (1992). Food science and techniques: Reports of
the scientific committee for food. (EUR13913). Luxembourg, Luxembourg.
http://ec.europa.eu/food/fs/sc/scf/reports/scf reports 26.pdf.

Coggan. AR: Baranauskas. MN: Hinrichs. RT: Liu. Z: Carter. ST. (2021). Effect of dietary nitrate on
human muscle power: a systematic review and individual participant data meta-analysis
[Review], JISSN 18: 66. http://dx.doi.org/10.1186/si 2970-021 -00463-z.

Coggan. ARR: Thies. (2020). Simultaneous Pharmacokinetic Analysis of Nitrate and its Reduced

Metabolite, Nitrite, Following Ingestion of Inorganic Nitrate in a Mixed Patient Population.
Pharm Res 37: 235-235. http:/ /dx.doi.org/10.1007/sl 1095-020-02959-w.

Cohen. AM: Hersh. WR: Peterson. K: Yen. PY. (2006). Reducing workload in systematic review

preparation using automated citation classification. J Am Med Inform Assoc 13: 206-219.
http: / /dx. doi. o r g /10.119 7 /i amia. M19 2 9.

Desimone. LA. (2009). Quality of water from domestic wells in principal aquifers of the United

States, 1991-2004. In Geological Survey Scientific Investigations (pp. 139). (SIR 2008-5227).
Reston, VA: U.S. Department of the Interior, U.S. Geological Survey.
http: / /pubs.usgs. gov/sir /2 008/5227/.

EFSA (European Food Safety Authority). (2017). Guidance on the use of the weight of evidence
approach in scientific assessments. EFSA J15: 1-69.
http://dx.doi.Org/10.2903/i.efsa.2017.4971.

(EFSA), EFSA. (2017a). Re-evaluation of sodium nitrate (E 251) and potassium nitrate (E 252) as
food additives. EFSA J15: e04787. http: //dx.doi.Org/10.2903 /i.efsa.2017.4787.

This document is a draft for review purposes only and does not constitute Agency policy.

R-l	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

40

41

42

43

44

45

46

47

48

49

50

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

(EFSA), EFSA. (2017b). Re-evaluation of potassium nitrite (E 249) and sodium nitrite (E 250) as
food additives. EFSA J15: e04786. http: //dx.doi.Org/10.2903 /i.efsa.2017.4786.

Farland. WH. (2005). [Memo to Science Policy council regarding implementation of the cancer
guidelines and accompanying supplemental guidance - Science Policy Council Cancer
Guidelines. Implementation Workgroup communication I: Application of the mode of action
framework in mutagenicity determinations for carcinogenicity]. Available online at
https://www.epa.gov/sites/production/files/2015-
01 /documents /cgiwgcommuniation i.pdf (accessed

FDA. Title 21: Food and drugs: Chapter I: Food and Drug Administration. Department of Health and
Human Services: Subchapter B: Food for human consumption: Part 165- Beverages. 21 CFR
§ 165 699-719 (2021). https://www.ecfr.gov/current/title-21/chapter-I/subchapter-
B/part-165

Fields. S. (2004). Global nitrogen: Cycling out of control. Environ Health Perspect 112: A556-A563.

Gonzalez. A: Hyde. E: Sangwan. N: Gilbert. TA: Viirre. E: Knight. R. (2016). Migraines are correlated
with higher levels of nitrate-, nitrite-, and nitric oxide-reducing oral microbes in the
American gut project cohort mSystems 1. http: //dx.doi.org/10.1128/mSystems.00105-16.

Higgins. TPT: Thomas. 1: Chandler. 1: Cumpston. M: Li. T: Page. Ml: Welch. VA. (2022). Cochrane
handbook for systematic reviews of interventions version 6.3. Higgins, JPT; Thomas, J;
Chandler, J; Cumpston, M; Li, T; Page, MJ; Welch, VA.
http: / /www, training. Cochrane. o r g /handbo ok.

Hill. AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-
300. http://dx.doi.Org/10.1177/003591576505800503.

Hoenig. TM: Heisev. DM. (2001). The abuse of power: The pervasive fallacy of power calculations for
data analysis. Am Stat 55: 19-24.

Howard. BE: Phillips. 1: Miller. K: Tandon. A: Mav. D: Shah. MR: Holmgren. S: Pelch. KE: Walker. V:
Roonev. AA: Macleod. M: Shah. RR: Thayer. K. (2016). SWIFT-Review: A text-mining
workbench for systematic review. Syst Rev 5: 87. http: //dx.doi.org/10.1186/si3643-016-
0263-z.

Howard. BE: Phillips. 1: Tandon. A: Maharana. A: Elmore. R: Mav. D: Sedvkh. A: Thayer. K: Merrick.
BA: Walker. V: Roonev. A: Shah. RR. (2020). SWIFT-Active Screener: Accelerated document
screening through active learning and integrated recall estimation. Environ Int 138:

105623. http://dx.doi.Org/10.1016/i.envint.2020.105623.

Hutson. SS... B..arber. N. L.... K..ennev. 1. F.... L..insev. K. S.... L..umia. D. S.... and Maupin. M. A. (2004).
Estimated use of water in the united states in 2000. USGS 1268.

IARC (International Agency for Research on Cancer). (2010). Ingested nitrate and nitrite, and
cyanobacterial peptide toxins [IARC Monograph] (pp. v-vii, 1-412). Lyon, France.
http://monographs.iarc.fr/ENG/Monographs/vol94/mono94.pdf.

IPCS (International Programme on Chemical Safety). (2005). SIDS Initial Assessment Report For
SIAM 20. Sodium Nitrite, http://www.inchem.org/documents/sids/sids/7632000.pdf.

TECFA (Joint FAO/WHO Expert Committee on Food Additives). (1995). Evaluation of certain food
additives and contaminants: Forty-fourth report of the Joint FA0/\NH0 Expert Committee
on Food Additives. (BIOSIS/96/25016). Geneva, Switzerland: World Health organization.

Kapil. V: Havdar. SM: Pearl. V: Lundberg. 10: Weitzberg. E: Ahluwalia. A. (2013). Physiological role
for nitrate-reducing oral bacteria in blood pressure control. Free Radic Biol Med 55: 93-100.
http://dx.doi.Org/10.1016/i.freeradbiomed.2012.ll.013.

Kulinkina. AY: Mohan. VR: Francis. MR: Kattula. D: Sarkar. R: Plummer. ID: Ward. H: Kang. G: Balrai.
V: Naumova. EN. (2016). Seasonality of water quality and diarrheal disease counts in urban
and rural settings in south India. Sci Rep 6: 20521. http://dx.doi.org/10.1038/srep20521.

Kuper F. T. il HP. (1995). Subchronic toxicity experiments with potassium nitrite in rats. In E
Council of (Ed.), Health aspects of nitrate and its metabolites (particularly nitrite):

This document is a draft for review purposes only and does not constitute Agency policy.

R-2	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

40

41

42

43

44

45

46

47

48

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Proceedings, international workshop, Bilthoven (Netherlands), 8-10 November 1994 (pp.
195-212). Strasbourg, France: Council of Europe Press.

Lin. YT: Cheng. CT: Chen. TW: Lin. Z. (2020). Incorporating exogenous and endogenous exposures into
dietary risk assessment of nitrates and nitrites in vegetables: A probabilistic integrated
toxicokinetic modeling approach. J Agric Food Chem 68: 1079-1090.
http://dx.doi.org/10.1021/acs.iafc.9b06720.

Maekawa. A: Ogiu. T: Onodera. H: Furuta. K: Matsuoka. C: Ohno. Y: Odashima. S. (1982).

Carcinogenicity studies of sodium nitrite and sodium nitrate in F-344 rats. Food Chem
Toxicol 20: 25-33. http://dx.doi.org/10.1016/s0278-6915C82180005-7.

Mattila. M: Niinisto. S: Takkinen. HM: Tapanainen. H: Reinivuo. H: Akerlund. M: Suomi. 1: Ahonen. S:
Ilonen. 1: Toppari. 1: Knip. M: Veiiola. R: Virtanen. SM. (2020). Maternal nitrate and nitrite
intakes during pregnancy and risk of islet autoimmunity and type 1 diabetes: The DIPP
cohort study. J Nutr 150: 2969-2976. http: //dx.doi.org/10.1093 /in/nxaa250.

Mensinga. TT: Speiiers. GT: Meulenbelt. 1. (2003). Health implications of exposure to environmental
nitrogenous compounds [Review], Toxicol Rev 22: 41-51.
http: / /dx.doi.org/10.2165 /00139709-200322010-00005.

Michigan DEO (Michigan Department of Environmental Quality). (2015). Chemical update
worksheet: Nitrite (DD). Lansing, MI. https://www.michigan.gov/-
/media/Proiect/Websites/egle/Documents/Programs/RRD/Remediation/Rules—
Criteria/Chemical-Update-Worksheets/N-"Z/Nitrite-
Datasheetpdf?rev=278b35ed9dca4a99b76e89b9fc22babl.

Mina. A: Stathopoulos. 1: Sinanian. T: Mcneice. L: Holmes. D: Fletcher. KL: Bottero. E: Banukumar. S:
Vazquez. S. (2021). Comparison of different methods used in drugs of abuse for sample
validity testing including pH methods, specific gravity methods, TECO™ Drug Adulteration
Test Strip and oxidant assay. 2: 550-566. http: //dx.doi.org/10.1515 /almed-2021-0026.

NASEM (National Academies of Sciences, Engineering, and Medicine). (2021). Review of U.S. EPA's
ORD staff handbook for developing IRIS assessments: 2020 version. Washington, DC:
National Academies Press, http: //dx.doi.org/10.17226/26289.

NDEP (Nevada Division of Environmental Protection). (2020). Nevada Division of Environmental
Protection basic comparison levels: 2020 update.
https://ndep.nv.gov/uploads/documents/ndep-bcls-august-202Q.pdf.

NRC (National Research Council). (2009). Science and decisions: Advancing risk assessment
Washington, DC: National Academies Press. http: //dx.doi.org/10.17226/122Q9.

NRC (National Research Council). (2014). Review of EPA's Integrated Risk Information System
(IRIS) process. Washington, DC: The National Academies Press.
http: / /dx. doi. o r g /10.172 2 6 /18 764.

NTP (National Toxicology Program). (2001a). Toxicology and carcinogenesis studies of sodium

nitrite (CAS No. 7632-00-0) in F344/N rats and B6C3F1 mice (drinking water studies) (pp.
280). (TR495). Research Triangle Park, NC: U.S. Department of Health and Human Services,
National Institutes of Health, National Toxicology Program.

https://ntrl.ntis.gov/NTRL/dashboard/searchResults/titleDetail/PB20011Q7676.xhtml.

NTP (National Toxicology Program). (2001b). Toxicology and carcinogenesis studies of sodium
nitrite (CASRN 7632-00-0) in F344/N rats and B6C3F1 mice (drinking water studies)
[NTP](pp. 7-273). (ISSN 0888-8051 NTP TR 495; NIH Publication No. 01-3954). Research
Triangle Park, NC. http: //ntp.niehs.nih.gov/ntp /htdocs/It rpts/tr495.pdf.

NTP (National Toxicology Program). (2015). Handbook for conducting a literature-based health

assessment using OHAT approach for systematic review and evidence integration. Research
Triangle Park, NC: U.S. Deptartment of Health and Human Services, National Toxicology

This document is a draft for review purposes only and does not constitute Agency policy.

R-3	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

40

41

42

43

44

45

46

47

48

49

50

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Program, Office of Health Assessment and Translation.
https://ntp.niehs.nih.gov/ntp/ohat/pubs/handbookian2015 508.pdf.

NTP (National Toxicology Program). (2021). Report on carcinogens, fifteenth edition. Research
Triangle Park, NC: U.S. Department of Health and Human Services, Public Health Service.
http://dx.doi.org/10.22427/NTP-QTHER-l 003.

Pokornv L. M. aturana I. B..ortle WH. (2006). Sodium Nitrate and Nitrite. In Kirk-Othmer
Encyclopedia of Chemical Technology (5th ed.). New York: John Wiley & Sons.
http://dx.doi.org/10.1002/0471238961.1915040916151115.a01.pub2.

SCF (Scientific Committee on Food). (1997). Reports of the Scientific Committee for Food.

Luxembourg: European Commission, https: //food.ec.europa.eu/system/files/2020-12 /sci-
com scf reports 38.pdf.

Schiinemann. H: Brozek. 1: Guvatt. G: Oxman. A. (2013). GRADE handbook. Available online at
https://gdt.gradepro.org/app/handbook/handbook.html (accessed April 22, 2022).

Sellars. S: Nunes. V. (2021). Synthetic nitrogen fertilizer in the U.S. farmdoc daily 11.

Speiiers. GTA: van Went. GF: van Apeldoorn. ME: Montizaan. GK: Tanus. TA: Canton. TH: van Gestel.
CAM: van der Heiiden. CA: Heiina-Merkus. E: Knapp. AGA: Luttik. R: de Zwart. D. (1989).
Integrated criteria document nitrate; effects. Appendix to RIVM Report No. 758473012.
Bilthoven, Netherlands: National Institute of Public Health and Environmental Protection.
https://www.rivm.nl/publicaties/integrated-criteria-document-nitrate.

Sterne. TAC: Hernan. MA: Reeves. BC: Savovic. I: Berkman. ND: Viswanathan. M: Henry. D: Altman.

DG: Ansari. MT: Boutron. I: Carpenter. TR: Chan. AW: Churchill. R: Peeks. IT: Hrobiartsson. A:
Kirkham. 1: Tiini. P: Loke. YK: Pigott. TP: Ramsay. CR: Regidor. D: Rothstein. HR: Sandhu. L:
Santaguida. PL: Schiinemann. HI: Shea. B: Shrier. I: Tugwell. P: Turner. L: Valentine. TC:
Waddington. H: Waters. E: Wells. GA: Whiting. PF: Higgins. TPT. (2016). ROBINS-I: A tool for
assessing risk of bias in non-randomised studies of interventions. BMJ 355: i4919.
http: / /dx. doi. o r g /10.113 6 /bmi. i4919.

Sterne. TAC: Savovic. 1: Page. MT: Elbers. RG: Blencowe. NS: Boutron. I: Cates. CI: Cheng. HY: Corbett.
MS: Eldridge. SM: Emberson. TR: Hernan. MA: Hopewell. S: Hrobiartsson. A: Tunqueira. PR:
Tuni. P: Kirkham. IT: Lasserson. T: Li. T: Mcaleenan. A: Reeves. BC: Shepperd. S: Shrier. I:
Stewart. LA: Tilling. K: White. IR: Whiting. PF: Higgins. TPT. (2019). RoB 2: a revised tool for
assessing risk of bias in randomised trials. BMJ 366:14898.
http://dx.doi.org/10.1136/bmi.14898.

TCEO (Texas Commission on Environmental Quality). (2023). TRRP protective concentration levels:
May 2023 PCL and supporting tables. Austin, TX: Texas Commission on Environmental
Quality :: TCEQ. Retrieved from

https://www.tceq.texas.gov/remediation/trrp/trrppcls.html

Til. HP: Falke. HE: Kuper. CF: Willems. MI. (1988). Evaluation of the oral toxicity of potassium nitrite
in a 13-week drinking-water study in rats. Food Chem Toxicol 26: 851-859.
http: //dx.doi.org/10.1016/0278-6915C88190026-9.

U.S. EPA (U.S. Environmental Protection Agency). (1987). Nitrite IRIS summary.
https://iris.epa.gov/ChemicalLanding/&substance nmbr=78.

U.S. EPA (U.S. Environmental Protection Agency). (1988). Recommendations for and documentation
of biological values for use in risk assessment [EPA Report], (EPA600687008). Cincinnati,
OH. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.

U.S. EPA (U.S. Environmental Protection Agency). (1990). Prinking Water Criteria Pocument for
Nitrate/Nitrite [EPA Report], (TR-1242-60B). Cincinnati, OH.

U.S. EPA (U.S. Environmental Protection Agency). (1991a). Guidelines for developmental toxicity
risk assessment Fed Reg 56: 63798-63826.

U.S. EPA (U.S. Environmental Protection Agency). (1991b). Integrated risk information system

(IRIS): Chemical assessment summary: Nitrate; CASRN 14797-55-8. Washington, PC: U.S.

This document is a draft for review purposes only and does not constitute Agency policy.

R-4	PRAFT-PO 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

40

41

42

43

44

45

46

47

48

49

50

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

Environmental Protection Agency. Integrated Risk Information System.

https: / /cfpub.epa.gov/ncea/iris /iris documents /documents /subst/0076 summarv.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (1994). Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry [EPA Report],
(EPA600890066F). Research Triangle Park, NC.

https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=71993&CFID=51174829&CFTOKE
N=25006317.

U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk
assessment [EPA Report], (EPA/630/R-96/009). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum.
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=30004Y0B.txt.

U.S. EPA (U.S. Environmental Protection Agency). (1997). Health effects assessment summary
tables: FY 1997 update [EPA Report], (EPA540R97036). Washington, DC: U.S.

Environmental Protection Agency, Office of Emergency and Remedial Response.
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=2 000Q0GZ.txt.

U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk

assessment [EPA Report], (EPA/630/R-95/001F). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http://www.epa.gov/risk/guidelines-
neurotoxicitv-risk-assessment

U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes. (EPA630P02002F). Washington, DC.
https://www.epa.gov/sites/production/files/2014-12/documents/rfd-final.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2004). Toxicological review of boron and

compounds. In support of summary information on the Integrated Risk Information System
(IRIS) [EPA Report], (EPA/635/04/052). Washington, DC: U.S. Environmental Protection
Agency, IRIS. http://nepis.epa.gov/exe/ZyPURL.cgi?Dockey=Pl 006CK9.txt.

U.S. EPA (U.S. Environmental Protection Agency). (2005a). Guidelines for carcinogen risk
assessment [EPA Report], (EPA630P03001F). Washington, DC.
https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for assessing

susceptibility from early-life exposure to carcinogens [EPA Report], (EPA/630/R-03/003F).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
https://www.epa.gov/risk/supplemental-guidance-assessing-susceptibilitv-earlv-life-
exposure-carcinogens.

U.S. EPA (U.S. Environmental Protection Agency). (2011). Recommended use ofbody weight 3/4 as
the default method in derivation of the oral reference dose. (EPA100R110001). Washington,
DC. https://www.epa.gOv/sites/production/files/2013-09/documents/recommended-use-
of-bw34.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance [EPA
Report], (EPA100R12001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, https: //www.epa.gov/risk/benchmark-dose-technical-guidance.

U.S. EPA (U.S. Environmental Protection Agency). (2014). Guidance for applying quantitative data to
develop data-derived extrapolation factors for interspecies and intraspecies extrapolation
[EPAReport], (EPA/100/R-14/002F). Washington, DC: Risk Assessment Forum, Office of
the Science Advisor, https://www.epa.gov/sites/production/files/2015-
01 /documents /ddef-final.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2015). Peer review handbook [EPA Report] (4th
ed.). (EPA/100/B-15/001). Washington, DC: U.S. Environmental Protection Agency, Science
Policy Council, https: //www.epa.gov/osa/peer-review-handbook-4th-edition-2015.

This document is a draft for review purposes only and does not constitute Agency policy.

R-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

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

U.S. EPA (U.S. Environmental Protection Agency). (2016). Chemical Contaminant Summaries for the
Third Six-Year Review of Existing National Primary Drinking Water Regulations [EPA
Report], Washington, DC: U.S. Environmental Protection Agency, Office of Water.
https://www.epa.gOv/sites/production/files/2016-12/documents/810rl6004.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2017). Guidance to assist interested persons in
developing and submitting draft risk evaluations under the Toxic Substances Control Act
(EPA/740/R17/001). Washington, DC: U.S Environmental Protection Agency, Office of
Chemical Safety and Pollution Prevention.
https://www.epa.gov/sites/production/files/2017-
06/documents/tsca ra guidance final.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2018). An umbrella Quality Assurance Project
Plan (QAPP) for PBPK models [EPA Report], (0RD QAPP ID No: B-0030740-QP-1-1).
Research Triangle Park, NC.

U.S. EPA (U.S. Environmental Protection Agency). (2019a). ChemView [Database], Retrieved from
https://chemview.epa.gov/chemview

U.S. EPA (U.S. Environmental Protection Agency). (2019b). CompTox Chemicals Dashboard
[Database], Research Triangle Park, NC. Retrieved from
https://comptox.epa.gov/dashboard

U.S. EPA (U.S. Environmental Protection Agency). (2020a). ORD staff handbook for developing IRIS
assessments (public comment draft) [EPAReport], (EPA/600/R-20/137). Washington, DC:
U.S. Environmental Protection Agency, Office of Research and Development, Center for
Public Health and Environmental Assessment

https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=350086.

U.S. EPA (U.S. Environmental Protection Agency). (2020b). Umbrella quality assurance project plan
(QAPP) for dosimetry and mechanism-based models. (EPA QAPP ID Number: L-CPAD-
0032188-QP-1-2). Research Triangle Park, NC.

U.S. EPA (U.S. Environmental Protection Agency). (2021). CompTox chemicals dashboard.
Washington, DC. Retrieved from https: //comptox.epa.gov/dashboard

U.S. EPA (U.S. Environmental Protection Agency). (2022). ORD staff handbook for developing IRIS
assessments [EPAReport], (EPA600/R-22/268). Washington, DC: U.S. Environmental
Protection Agency, Office of Research and Development, Center for Public Health and
Environmental Assessment

https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=356370.

Walton. G. (1951). Survey of literature relating to infant methemoglobinemia due to nitrate-
contaminated water. Am J Public Health Nations Health 41: 986-996.
http://dx.doi.Org/10.2105/ATPH.41.8 Pt 1.986.

Ward. MH. (2009). Too much of a good thing? Nitrate from nitrogen fertilizers and cancer [Review],
Rev Environ Health 24: 357-363. http://dx.doi.Org/10.1515 /reveh.2009.24.4.357.

Ward. MH: deKok. TM: Levallois. P: Brender. I: Gulis. G: Nolan. BT: VanDerslice. I. (2005).

Workgroup report: Drinking-water nitrate and health—Recent findings and research needs.
Environ Health Perspect 113: 1607-1614. http://dx.doi.org/10.1289/ehp.8043.

Water and Air Quality Bureau. H. ealthv Environments and Consumer Safety Branch. H..ealth
Canada. (2013). Guidelines for Canadian drinking water quality: Guideline technical
document: Nitrate and nitrite. In Guidelines for Canadian Drinking Water Quality. Ottawa,
Ontario: Minister of Health, http: //www.hc-sc.gc.ca/ewh-semt/pubs/water-
e au /nitrate nitrite /index- e ng.php.

WHO (World Health Organization). (2003). Nitrate and nitrite in drinking-water: Background
document for development of WHO Guidelines for drinking-water quality.
(WHO/SDE/WSH/04.03/56). Geneva, Switzerland.
http://apps.who.int/iris/handle/10665/75380.

This document is a draft for review purposes only and does not constitute Agency policy.

R-6	DRAFT-DO NOT CITE OR QUOTE


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Protocol for the Nitrate and Nitrite IRIS Assessment (Oral)

WHO (World Health Organization). (2016). Nitrate and nitrite in drinking-water: Background
document for development of WHO guidelines for drinking-water quality.
(WHO/FWC/WSH/16.52). Geneva, Switzerland.

http://www.who.int/water sanitation health/dwq/chem icals/nitrate-nitrite-backgrou nd-
ianl7.pdf.

Wolffe. TAM: Whalev. P: Halsall. C: Roonev. AA: Walker. VR. (2019). Systematic evidence maps as a
novel tool to support evidence-based decision-making in chemicals policy and risk
management Environ Int 130: 104871. http://dx.doi.Org/10.1016/i.envint.2019.05.065.

Zeilmaker. Ml: Bakker. MI: Schothorst. R: Slob. W. (2010). Risk assessment of N-

nitrosodimethylamine formed endogenously after fish-with-vegetable meals. Toxicol Sci
116: 323-335. http://dx.doi.org/10.1093/toxsci/kfq093.

Zeilmaker. Ml: Meulenbelt. 1: Kortbover. TM: Slob. W. (1996). Safety assessment of nitrate:
Mathematical modeling of nitrite formation in man and its application in the risk
assessment of nitrate. (Report No. 235802002). Bilthoven, Netherlands: National Institute
of Public Health and the Environment (RIVM).

This document is a draft for review purposes only and does not constitute Agency policy.

R-7	DRAFT-DO NOT CITE OR QUOTE


-------