Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment (Preliminary Assessment Materials) August 2023


A EPA

EPA/635/R-23/167
IRIS Assessment Protocol

www.epa.gov/iris

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

(Preliminary Assessment Materials)

August 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 Vanadium and Compounds (Inhalation) IRIS Assessment

DISCLAIMER

This document is a public comment draft. This information is distributed solely for review
purposes under applicable information quality guidelines. It has not been formally disseminated by
the Environmental Protection Agency. 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ii	DRAFT-DO NOT CITE OR QUOTE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 Vanadium and Compounds (Inhalation) IRIS Assessment

CONTENTS

AUTHORS | CONTRIBUTORS | REVIEWERS	viii

1.	INTRODUCTION	1-1

2.	SCOPING AND INITIAL PROBLEM FORMULATION	2-1

2.1.	BACKGROUND	2-1

2.1.1.	Physical and Chemical Properties	2-1

2.1.2.	Sources, Production, and Uses	2-2

2.1.3.	Environmental Fate and Transport	2-2

2.1.4.	Potential for Human Exposure and Populations with Potentially Greater Exposure	2-3

2.2.	SCOPING SUMMARY	2-5

2.3.	PROBLEM FORMULATION	2-5

2.4.	KEY SCIENCE ISSUES	2-6

3.	OVERALL OBJECTIVES AND SPECIFIC AIMS	3-1

3.1.	OBJECTIVES	3-1

3.2.	SPECIFIC AIMS	3-1

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

4.1.	POPULATIONS, EXPOSURES, COMPARATORS, AND OUTCOMES (PECO) CRITERIA FOR THE
SYSTEMATIC EVIDENCE MAP	4-1

4.2.	SUPPLEMENTAL CONTENT SCREENING CRITERIA	4-2

4.3.	USE OF EXISTING ASSESSMENTS	4-4

4.4.	LITERATURE SEARCH STRATEGIES	4-5

4.4.1.	Core Database Searches	4-5

4.4.2.	Searching Other Sources	4-6

4.4.3.	Non-Peer Reviewed Data	4-7

4.5.	LITERATURE SCREENING STRATEGY	4-8

4.5.1.	Title and abstract-level screening	4-9

4.5.2.	Full-text-level screening	4-9

4.5.3.	Multiple Citations with the Same Data	4-9

4.5.4.	Literature Flow Diagram	4-10

4.6.	LITERATURE INVENTORY	4-12

4.6.1. Studies That Meet Problem Formulation PECO Criteria	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 Vanadium and Compounds (Inhalation) IRIS Assessment

4.6.2. Organizational Approach for Supplemental Material	4-12

4.7. INITIAL LITERATURE INVENTORIES FOR VANADIUM (INHALATION)	4-12

5.	REFINE PROBLEM FORMULATION AND SPECIFY ASSESSMENT APPROACH	5-1

5.1.	REFINEMENTS TO PECO CRITERIA	5-1

5.1.1. Other Exclusions Based on Full-Text Content	5-3

5.2.	REFINEMENTS TO SUPPLEMENTAL CONTENT SCREENING CRITERIA	5-5

5.3.	UNITS OF ANALYSES FOR DEVELOPING EVIDENCE SYNTHESIS AND INTEGRATION
JUDGMENTS FOR HEALTH EFFECT CATEGORIES	5-10

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

6.2.1. Epidemiology Study Evaluation Considerations Specific to Vanadium	6-15

6.3.	EXPERIMENTAL ANIMAL STUDY EVALUATION	6-19

6.4.CONTROLLED	HUMAN EXPOSURE STUDY EVALUATION	6-29

6.5.	PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODEL DESCRIPTIVE SUMMARY

AND EVALUATION	6-29

6.6.	IN VITRO AND OTHER MECHANISTIC STUDY EVALUATION	6-29

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	9-8

9.3.1.	Dose-Response Analysis in the Range of Observation	9-8

9.3.2.	Extrapolation: Slope Factors and Unit Risk	9-11

9.3.3.	Extrapolation: Reference Values	9-11

10.	PROTOCOL HISTORY	10-1

REFERENCES	R-l

APPENDIX A. ELECTRONIC DATABASE SEARCH STRATEGIES	A-l

APPENDIX B. PROCESS FOR SEARCHING AND COLLECTING EVIDENCE FROM SELECTED

OTHER RESOURCES	B-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 Vanadium and Compounds (Inhalation) IRIS Assessment

TABLES

Table 2-1. Chemical identity of vanadium compounds with repeat dose inhalation toxicity data 2-4

Table 2-2. EPA program and regional office interest in the assessment of inhalation exposure to

vanadium and compounds 2-5

Table 4-1. Populations, exposures, comparators, and outcomes (PECO) criteria 4-1

Table 4-2. Categories of Potentially Relevant Supplemental Material 4-3

Table 5-1. Assessment PECO for the vanadium and compounds (inhalation) assessment 5-2

Table 5-2. Categories of potentially relevant supplemental material 5-6

Table 5-3. Human and animal endpoint grouping categories 5-10

Table 6-1. Information relevant to evaluation domains for epidemiology studies 6-6

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

epidemiology studies 6-7

Table 6-3 Criteria for evaluating exposure measurement in epidemiology studies of vanadium 6-16

Table 6-4. Domains, questions, and general considerations to guide the evaluation of animal

toxicology studies 6-21

Table 6-5. Domains, questions, and general considerations to guide the evaluation of in vitro

studies 6-31

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 judgments of the certainty of the evidence for hazard for

each unit of analysis 8-7

Table 8-4. Framework for evidence synthesis judgments from studies in humans 8-11

Table 8-5. Framework for evidence synthesis 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 (in addition to the

health effect category-specific evidence integration judgment) 9-4

Table 9-2. Example table used in assessment to show endpoint consideration judgments for POD

derivation 9-6

Table 9-3. Specific example of presenting endpoints considered for dose-response modeling and

derivation of points of departure 9-6

Table A-l. Database search strategies for vanadium and compounds A-l

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

v DRAFT-DO NOT CITE OR QUOTE

-------
FIGURES

Figure 1-1. IRIS systematic review problem formulation and method documents	1-2

Figure 4-1. Literature search flow diagram for vanadium and compounds	4-11

Figure 4-2. Inventory heatmap of PECO-relevant vanadium and compounds	4-14

Figure 4-3. Inventory heatmap of PECO-relevant vanadium and compounds (inhalation

exposure) animal studies by study design and health system	4-15

Figure 6-1. Overview of IRIS study evaluation process	6-2

-------
ABBREVIATIONS

ADME absorption, distribution, metabolism,

and excretion
AOP adverse outcome pathway
ATSDR Agency for Toxic Substances and

Disease Registry
BMD benchmark dose

BMDL benchmark dose lower confidence limit
BMDS Benchmark Dose Software

CAS Chemical Abstracts Service
CASRN Chemical Abstracts Service registry

number
CO I	conflict of interest

CPAD Chemical and Pollutant Assessment
Division

CPHEA Center for Public Health and
Environmental Assessment
DNA deoxyribonucleic acid
EPA Environmental Protection Agency
HAWC Health Assessment Workspace

Collaborative
HEC human equivalent concentration
HERO Health and Environmental Research
Online

IAP	IRIS Assessment Plan

IARC International Agency for Research on
Cancer

IRIS Integrated Risk Information System
IUR inhalation unit risk

LOAEL	lowest-observed-adverse-effect level

LOEL	lowest-observed-effect level

MeSH	Medical Subject Headings

MOA	mode of action

NOAEL no-observed-adverse-effect level
NOEL no-observed-effect level
NTP National Toxicology Program
OAR Office of Air and Radiation
OECD Organisation for Economic

Co-operation and Development
ORD Office of Research and Development
OSF oral slope factor

PBPK physiologically based pharmacokinetic
PECO populations, exposures, comparators,

and outcomes
PK	pharmacokinetic

PM	particulate matter

POD point of departure
RfC	inhalation reference concentration

RfD oral reference dose
ROBINS I Risk of Bias in Nonrandomized Studies

of Interventions
UF	uncertainty factor

UFh human variation uncertainty factor
WOS Web of Science

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

AUTHORS | CONTRIBUTORS | REVIEWERS

Assessment Managers

Survanaravana Vulimiri. Ph.D.	EPA/ORD/CPHEA

Kathleen Newhouse. M.S.

Assessment Team

Francesca Branch. DrPH	EPA/ORD/CPHEA

David Farrar. Ph.D.

Elizabeth Radke. Ph.D.

Erin Yost. Ph.D.

John Stanek, Ph.D.

Andre Weaver. Ph.D.

Michelle Angrish. Ph.D.

Bevin Blake. Ph.D.

Michele Tavlor. Ph.D.

Martha Powers. Ph.D.

Dustin Kapraun. Ph.D.

Bidya Prasad, Ph.D. (formerly EPA)

Executive Direction

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

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

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

Kristina Thayer, Ph.D. (CPAD Director)

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

Paul White, Ph.D. (CPAD Senior Science Advisor)

Ravi Subramaniam, Ph.D. (CPAD Senior Science

Advisor)

Janice Lee, Ph.D. (Branch Chief)

Elizabeth Radke-Farabaugh, Ph.D. (Branch Chief)

Viktor Morozov, Ph.D. (Branch Chief)

Glenn Rice, Ph.D. (Branch Chief)

Deborah Segal, M.S. (Acting Branch Chief)

Vicki Soto, B.S. (Branch Chief)

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Production Team

Maureen Johnson
Ryan Jones
Dahnish Shams
Avanti Shirke
Jessica Soto-Hernandez
Samuel Thacker
Garland Waleko

Brittany Schulz

EPA/ORD/CPHEA

Student Services Contractor, Oak Ridge Associated
Universities (0 RAU]

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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

1. INTRODUCTION

The Integrated Risk Information System (IRIS) Program is undertaking an assessment of the
health effects from inhalation exposure to vanadium and compounds.1 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.

A draft IRIS Assessment Plan (IAP) for Vanadium and Compounds (Inhalation Exposure)
was presented at a public science meeting on July 14, 2021 (see http s://www, ep a. gov/ iris/ir is-
public-science-meeting-jul-2021) to seek input on the problem formulation components of the
assessment plan (U.S. EPA. 2021a). The 2021 IAP specified why vanadium and compounds were
selected for evaluation, described the objectives and specific aims of the assessment, provided draft
populations, exposures, comparators, and outcomes (PECO) criteria, and identified key areas of
scientific complexity. This assessment is being developed at the request of EPA's Office of Air and
Radiation (OAR). It may also be used to support actions in other EPA Program and Regional Offices
and can inform efforts to address inhalation exposure to vanadium by tribes, states, and
international health agencies.

This protocol document includes the IAP content, revised based on 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 describes what the assessment covers,
this protocol describes how the assessment is conducted (see Figure 1-1). The methods described
in this protocol are based on the Office of Research and Development (ORD) Staff Handbook for
Developing Integrated Risk Information System (IRIS) Assessments (referred to as the "IRIS
Handbook") fU.S. EPA. 20221.

The IRIS Program posts assessment protocols on its website. Public input received is
considered during preparation of the draft assessment.

*An assessment of oral exposure to vanadium and compounds was initiated prior to the inhalation
assessment and is being performed separately

fhttps://cfpub.epa.gov/ncea/iris drafts /recordisplav.cfm?deid=3487921

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Systematic Review: A structured and documented process for transparent
literature review using explicit, pre-specified scientific methods to identify, select,
assess, and summarize the findings of similar but separate studies.

Assessment
Initiated /

Scoping/Initial
Problem
Formulation

Specify Assessment
Approach

Assessment
Developed

Assessment Plans:

What the assessment will
cover

How the assessment will be conducted

Figure 1-1. 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

2.SCOPING AND INITIAL PROBLEM FORMULATION

2.1. BACKGROUND

The IRIS Program finalized an assessment of vanadium and compounds in 1987 that
included a reference dose (RfD) but did not derive inhalation toxicity values due to lack of
inhalation data (U.S. EPA. 1987). Since then, several relevant studies on vanadium inhalation
toxicity have been completed, including a two-year inhalation study conducted by the National
Toxicology Program fNTP. 20021. The focus of this document is on inhalation exposure to vanadium
and compounds and its potential impacts on human health. In this assessment "vanadium" refers to
the element vanadium as part of environmentally occurring compounds. Vanadium as a pure metal
is not found in the environment since it is unstable. Vanadium alloys (such as ferrovanadium) are
not considered within scope because they are mixtures (see Table 5.1). Oral exposure to vanadium
and compounds is currently under evaluation in a separate assessment

fhttps: //cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=3487921. Section 2.1 provides a
brief overview of aspects of the physiochemical properties, human exposure, and environmental
fate characteristics of vanadium. This overview provides a summaiy of background information for
contextual purposes only as it falls outside the scope of a human health toxicity assessment. This
overview is not intended to provide a comprehensive description of the available information on
these topics and is not recommended for use in decision-making. The reader is encouraged to refer
to the source materials cited below, more recent publications on these topics, and authoritative
reviews or assessments focused on these topics.

2.1.1. Physical and Chemical Properties

Vanadium has a complex chemistry, existing in the environment with three common
oxidation states (+3, +4, +5) (Gustafsson. 2019). Pure elemental vanadium does not exist naturally
(Rehder. 2015: ATSDR. 2012). Burning of fossil fuels containing vanadium results in the production
of vanadium as oxides, including VO, V203, V02, V205, which are emitted as fly ash into the
atmosphere (Sturini etal.. 2010: Crans etal.. 1998: Mamane and Pirrone. 1998). In general, specific
vanadium compounds relevant to environmental inhalation have not been well characterized in the
literature, and few methods are available to speciate vanadium in particulate matter (PM) (Shafer
etal.. 2012: Sturini etal.. 2010). Vanadium pentoxide (V205), a +5 vanadium species, is the most
common compound of vanadium used for industrial applications such as metal alloy production
and catalytic processes. In crude oils, vanadium is present as an organometallic complex, and upon
burning in boilers or furnaces, vanadium is left behind as vanadium oxides in residual oil fly ash
flPCS. 20011. Residual oil fly ash is a mixture of different vanadium compounds and other metals
and components of PM fHauser etal.. 19951. Most studies of human inhalation of vanadium, such as

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

2-1 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

those which evaluate vanadium as a component of PM, do not identify the specific vanadium
compound, but instead report total vanadium in air or biological matrix. Evidence from one study
suggests that pentavalent vanadium, in particular vanadium pentoxide, can be among the vanadium
compounds present in PM emitted from diesel engines and in PM in urban atmospheric aerosols
fShafer etal.. 20121.

When dissolved in water, vanadium speciation is a complex function of factors such as pH,
redox potential, and vanadium concentration. Vanadate species (+5) predominate under oxic
conditions and high pH, while vanadyl (+4) occurs under suboxic conditions and low pH and
trivalent vanadium (+3) occurs under anoxic conditions fGustafsson. 2019: Huang et al.. 20151.
Vanadium pentoxide undergoes hydrolytic reactions in water generating vanadate solutions
(Cohen. 2007: Crans etal.. 19981.

Table 2-1 lists the chemical identity of elemental vanadium (for reference only) and
inorganic vanadium compounds that have been used in inhalation toxicology studies.

2.1.2. Sources, Production, and Uses

Vanadium is a transition metal that occurs naturally in Earth's crust and is a component of
various minerals and most ores, tars, coal, and petroleum crude oils with heavy oils and bitumen
from tar sands being especially rich in vanadium (WHO. 1988). Natural sources of vanadium in the
air include continental dust, marine aerosol, and volcanic emissions (ATSDR. 2012). Vanadium has
been reported to have natural background concentrations in the air ranging from tenths of a
nanogram to a few nanograms (WHO. 2000). Vanadium is produced worldwide through mining or
recycling residues and waste materials. In the US, the main method of production of vanadium is
through reclamation (U.S. Department of Commerce. 2021).

The use of vanadium in industrial applications (e.g., steel production, vanadium redox-flow
batteries, and catalytic converters) could contribute to the release of vanadium into the
environment fSchlesinger etal.. 2017: ATSDR. 20121. However, fossil fuel combustion is thought to
be the major anthropogenic source of vanadium to the atmosphere, with vanadium found adsorbed
onto PM as a result fSchlesinger et al.. 2017: ATSDR. 20121.

2.1.3. Environmental Fate and Transport

As noted above, industrial processes, primarily burning of vanadium rich fuel, are reported
to be the major source of vanadium in the atmosphere. Vanadium oxides generated during such
combustion combine into particulate fly ash, also called residual oil fly ash (ROFA). ROFA is a
mixture of different vanadium compounds and other metals and components of particulate matter
fHauser et al.. 19951. At high temperatures encountered in combustion stacks (100-500°C), lower
oxides of vanadium will ultimately oxidize to varius vanadium oxides including V2O5 (U.S. EPA.
1985). Deposition is likely the only sink for atmospheric vanadium (Tullar and Suffet. 1975).

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

2-2 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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

The ambient air concentrations of vanadium in the United States varies widely depending
on factors including urban or rural location, seasonality, and geography (as reviewed by fATSDR.
201211. Generally, populations in cities located in the northeastern states have higher vanadium
concentrations occuring during winter months when more fuel oil is burnt for heating and
electricity flARC. 20061. In addition, populations near port cities have greater exposure to
vanadium due to higher concentrations of vanadium in marine vessel fuel and emissions (Spada et
al.. 2018): (Agrawal etal.. 2009): (Peltier and Lippmann. 2010). Relatively recent publications,
using air monitoring between 2007 and 2009, have reported average vanadium concentrations in
New York City of approximately 5 ng/m3 and ranging from approximately 2-15 ng/m3 which were
closely associated with ship traffic (Ito etal.. 2016): (Peltier and Lippmann. 2010). Air monitoring
in 2011 near the Seattle and Tacoma ports measured median vanadium concentrations of
approximately 6-8 ng/m3 (Spada etal.. 2018). However, recent regulations limiting sulfur content
of marine fuel oil have resulted in decreased vanadium emissions near ports, due to increased fuel
refinement (Kodros etal.. 2022): (Tao et al.. 2013): (Spada etal.. 2018).

Occupational exposure to vanadium in humans can occur through the inhalation of dust
generated during vanadium processing and through the inhalation of residual oil fly ash (ROFA)
during cleaning of oil-burning boilers and furnaces. Other occupational exposure occurs through
cleaning of oil boilers, vanadium pentoxide production, and metallurgical processes (e.g., ferroalloy
and V2O5 production facilities) associated with production of vanadium-containing vapors that
would condense forming respirable aerosols (Kucera and Sabbioni. 1998).

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

2-3 DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Table 2-1. Chemical identity of vanadium compounds with repeat dose inhalation toxicity data (short term,
subchronic, and chronic)

Name

Elemental
vanadium3

Vanadium
pentoxide

Bismuth
orthovanadate

Sodium
orthovanadate

Sodium
metavanadate

Ammonium
metavanadate

Vanadium
dioxide

CASRN

7440-62-2

1314-62-1

14059-33-7

13721-39-6

13718-26-8

7803-55-6

12036-21-4

DTXSIDb

2040282

2023806

20893971

2037269

3044336

1052533

5065194

Empirical formula

V

V2O5

Bi04V

Na3V04

NaV03

NH4VO3

VO2

Molecular mass (g/mol)

50.942

181.878

323.918

183.907

121.928

116.978

82.94

Oxidation state

0

+5

+5

+5

+5

+5

+4

Selected synonym(s)

Vanadium

Vanadium oxide; mu-
oxido[tetrakis(oxido)]
divanadium;
divanadium
pentoxide; vanadic
anhydride;
vanadin(V) oxide;
vanadium(V) oxide

Bismuth
vanadate(V)

(BiVCU);
bismuth(3+)
tetraoxidovana-
date(3"); bismuth
vanadium oxide;
vanadic acid;
bismuth vanadate
(BiVCU); bismuth
vanadate yellow

Trisodium

tetraoxidovana-

date(3");

sodium

vanadium

oxide;

trisodium

vanadate;

sodium

vanadate(V);

vanadic acid,

trisodium salt

Sodium vanadate;
sodium

trioxidovanadate(1_);
sodium vanadium
oxide; sodium
vanadium trioxide;
vanadic acid,
monosodium salt;
sodium
vanadate(V)

Ammonium
trioxovanadate(1_);
ammonium
trisoxidovanadate(1_);
ammonium
monovanadate;
ammonium
vanadate(V); vanadic
acid, ammonium salt;
ammonium vanadium
oxide; ammonium
vanadium trioxide

Bisoxidovanadium;
dioxido de
vanadio; dioxyde
de vanadium;
divanadium
tetraoxide;
divanadium
tetroxide;
vanadium (IV)
oxide; vanadium
dioxide;
vanadium (IV)
oxide

Solubility (g/100 ml)

Insoluble

0.8 (20°C)

-

-

-

-

-

Melting point (°C)

1.9 x 103 c

690c

-

858c

630c

200d

-

Boiling point (°C)

3.0 x 103 c

1.75 x 103c

-

-

-

-

-

Elemental vanadium included for reference only.

bDTXSIDs are unique substance identifiers used for curation by the EPA's Distributed Structure Searchable Toxicity (DSSTox) project.

Experimental average values for physicochemical properties are provided in EPA's CompTox Chemicals Dashboard at https://comptox.epa.gov/dashboard/. If
no experimental values are available on EPA's CompTox Chemicals Dashboard, "-"is shown.
dATSDR (2012)

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

2-4	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

2.2. SCOPING SUMMARY

During scoping, the IRIS Program met with EPA program and regional offices that had
interest in an IRIS assessment for inhalation exposure to vanadium and compounds to discuss
specific assessment needs. Table 2-2 summarizes input from this outreach. EPA's Office of
Transportation and Air Quality within the Office of Air and Radiation (OAR) nominated vanadium
compounds (including vanadium pentoxide) for an inhalation health assessment (both cancer and
noncancer) under the IRIS Program. Vanadium pentoxide has been used as a catalyst to control
emissions from diesel engines employed in mobile sources such as on-highway heavy-duty trucks,
nonroad equipment, and marine vessels.2 Under certain conditions, the use of vanadium in diesel
engine emission control devices can result in the potential for exposures to vanadium compounds,
such as vanadium pentoxide. A vanadium (inhalation) assessment could therefore help inform
decisions about potential health risks from increased vanadium in the atmosphere.

Table 2-2. EPA program and regional office interest in the assessment of
inhalation exposure to vanadium and compounds

EPA program
or regional
office

Oral3

Inhalation

Statute/Regulation

Anticipated uses/Interest

Office of Air
and Radiation

Clean Air Act

Vanadium and compounds (including vanadium
pentoxide) are mobile source air toxics.
Toxicological information developed for this
assessment may be used to inform risk
management decisions.

2.3. PROBLEM FORMULATION

The IRIS Program published an assessment of vanadium and compounds in 1987 that
included a reference dose (RfD) for vanadium pentoxide, but no inhalation toxicity values (U.S. EPA.
1987). A draft IRIS assessment addressing inhaled vanadium pentoxide was released for public
comment and external peer review in 2011 (U.S. EPA. 2011b). but was not finalized due to
recognition of a cross-Agency need for an assessment with broader consideration of vanadium
compounds (potentially aiding in the evaluation of toxic effects and helping to inform data gaps)
(see: https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=125). EPA's

20AR has issued an Advance Notice regarding plans for a new rulemaking that would establish new emission
standards for oxides of nitrogen (NOx) and other pollutants for highway heavy-duty engines
fhttps://www.epa.gov/regulations-emissions-vehicles-and-engines/advance-notice-proposed-rule-control-
air-pollution-newl.

3The IRIS Program is conducting a separate assessment of oral exposure to vanadium and compounds (U.S.
EPA. 2021b).

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

2-5 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Provisional Peer-Reviewed Toxicity Values (PPRTV) program developed a 2008 assessment on
vanadium pentoxide which identified respiratory inflammation in female rats as the most sensitive
endpoint for inhaled vanadium pentoxide in a two-year exposure study conducted by the National
Toxicology Program fNTP. 20021. Later, in 2009, the PPRTV program also finalized an assessment
that included soluble inorganic vanadium compounds other than vanadium pentoxide fU.S. EPA.
20091. However, this assessment found the evidence base was inadequate to support the derivation
of chronic or subchronic inhalation toxicity values for soluble inorganic vanadium compounds
other than vanadium pentoxide.

The IAP for Vanadium and Compounds (Inhalation) was released in May 2021 and
presented at a public science meeting on July 14, 2021 (see https://www.epa.gov/iris/iris-public-
science-meeting-iul-20211 to seek input on the problem formulation components of the assessment
plan and key science issues fU.S. EPA. 2021al. This protocol considers input received on the 2021
IAP.

2.4. KEY SCIENCE ISSUES

The 2021 IAP for inhalation exposure to vanadium and compounds (inhalation) identified
several key science issues based on the preliminary literature survey results and review of past
assessments on inhalation exposure to vanadium (see Section 2.3).

• Key Science Issue #1: Consideration of vanadium speciation and oxidation state.

Considering oxidation status could be important as preliminary results from oral exposure
studies in rodents indicates increased toxicity of vanadium in the +5 oxidation state
compared to vanadium +4 (Roberts et al.. 2016). As noted in Section 2, vanadium in solution
can convert between oxidation states and will form different species as a function of factors
including pH, concentration, and redox potential. Study evaluations for the available
inhalation studies, to the extent possible, will consider factors that could affect vanadium
oxidation state and speciation [e.g., study methods that involved aerosolizing vanadium
pentoxide (or other vanadium compounds) from solution, e.g., Gonzalez-Villalva et al.
f20111. rather than exposure to vanadium as a dust, e.g., NTP f20021]. In addition, data to
inform potential conversion between vanadium species and oxidation states in the body
also will be evaluated and discussed in the assessment

• Key Science Issue #2: Interpretation of data on noncancer respiratory responses to
vanadium pentoxide.

The two year NTP f20021 study reports increasing incidences of nonneoplastic lesions in
the upper and lower respiratory tract of rats and mice (both sexes) with increasing
vanadium pentoxide concentrations. Responses in all vanadium pentoxide exposure groups
were highly elevated compared to controls. Information on the biology underlying these
findings will aid interpretation of their use for hazard identification. Depending on the
hazard identification decisions, methods for low-dose extrapolation and the associated
uncertainties with any such approaches also would need to be explored and justified.

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

2-6 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

• Key Science Issue #3: Interpretation of data on rodent tumor responses.

The NTP (2002) study also reports that tumor responses (alveolar/bronchiolar neoplasms)
in male and female mice were highly elevated at all concentrations of vanadium pentoxide
exposure: 70-80% increased incidence at the lowest tested vanadium concentration;
control incidence in male mice was high (44%), but background incidence in females was
very low (2%). Tumor incidence in male rats was elevated slightly but not statistically
significant compared to controls. Previous reviews analyzed this tumor incidence against
concurrent controls as well as historical controls, which will be useful in interpreting these
data as they are considered in the assessment. In summary, aspects of the rodent tumor
data noted above, and the uncertainties will be considered in the assessment

• Key Science Issue #4. Cancer MOA for alveolar/bronchiolar neoplasms.

There is some support for both a mutagenic MOA and an MOA dependent on cytotoxicity
and reparative regeneration (and potentially other undetermined mechanisms) as
suggested in the EPA PPRTV assessment (U.S. EPA. 2008). A similar lack of clearly
delineated MOA(s) for alveolar/bronchiolar lung tumors with vanadium pentoxide
exposure was proposed in the unfinalized draft IRIS Assessment of Vanadium Pentoxide
fU.S. EPA. 2011bl. As reported in these reviews, mutagenicity tests for vanadium pentoxide
appear generally negative, but there is evidence of DNA strand breaks, aneuploidy,
cytotoxicity, and cell proliferation. A focused evaluation of the available evidence regarding
cancer MOA(s) for alveolar/bronchiolar neoplasms, including judgments regarding human
relevance, is expected to be a key component of the vanadium (inhalation) IRIS assessment.

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

15

16

17

18

19

20

21

22

23

24

25

26

27

28

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

3.OVERALL OBJECTIVES AND SPECIFIC AIMS

3.1.	OBJECTIVES

The overall objective of this assessment is to identify adverse health effects and
characterize exposure-response relationships for these effects to support development of
inhalation toxicity values. This assessment will use systematic review methods to evaluate the
epidemiological and toxicological literature for vanadium compounds, including consideration of
relevant mechanistic evidence. The evaluation conducted in this assessment will be consistent with
relevant EPA guidelines.4

3.2.	SPECIFIC AIMS

•	Develop a systematic evidence map (SEM) to identify epidemiological (i.e., human),
toxicological (i.e., experimental animal), and supplemental literature pertinent to
characterizing the health effects of inhalation exposure to vanadium. The PECO criteria used
to develop the SEM (referred to as "problem formulation PECO") 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 fThaver etal.. 2022: NASEM. 2021: Wolffe et al.. 20191.

° Supplemental material content includes: mechanistic studies, including in vivo, in vitro,
ex vivo, or in silico models; non-mammalian model systems; pharmacokinetic and
absorption, distribution, metabolism, and excretion (ADME) studies; PM studies (as
measured via air pollution monitoring stations); exposure characteristics (with no
health outcome); data pertinent to identify susceptible populations; mixture studies;
non-PECO routes of exposure; case studies (of 1-3 individuals); records with no original
data; conference abstracts, and errata.

•	Use the results of the SEM to (1) develop 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 available. If a PBPK or PK model is selected for use, the most reliable dose

4EPA guidelines: 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

metric will be applied based on analyses of the available dose metrics and the outcomes to
which they are being applied.

• 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 certainty 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, biological plausibility, and
other critical inferences supported by mechanistic, ADME, or PK/PBPK analyses.

• For each health effect category, summarize evidence synthesis (certainty 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 inhalation
reference concentrations (RfCs) and organ- or system-specific RfCs. Apply pharmacokinetic
and dosimetry modeling (possibly including PBPK modeling) to account for interspecies
differences, as appropriate. Derive an inhalation unit risk (IUR) 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

4.LITERATI)RE SEARCH, SCREENING, AND
INVENTORY

The literature search and screening processes described in this section were used to
develop a SEM (here) 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 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 are conducted using the processes described in this
chapter. The literature inventory is continually updated with new studies as the assessment
progresses.

4.1. POPULATIONS, EXPOSURES, COMPARATORS, AND OUTCOMES
(PECO) 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 presented in the IAP are presented in Table 4-1. The SEM
PECO criteria were intentionally broad to identify all the available evidence in humans and animal
models. As part of problem formulation, the SEM PECO is refined to develop the assessment PECO
and these refinements are presented in Section 5.1.

Table 4-1. Populations, exposures, comparators, and outcomes (PECO) criteria

PECO element

Evidence

Populations

Human: Any population and life stage (occupational or general population, including children
and other potentially sensitive populations).

Animal: Nonhuman mammalian animal species (whole organism) at any life stage (including
preconception, in utero, lactation, peripubertal, and adult stages). Studies of transgenic
animals will be tracked as mechanistic studies under "potentially relevant supplemental
material."

Exposu res

Relevant forms: Any forms of vanadium.

Human: Any exposure to vanadium compound(s) via the inhalation route, either explicitly
stated or considered plausible based on exposure assessment. Exposure can be based on
administered concentration, biomonitoring data (e.g., urine, blood, or other specimens),
environmental or occupational measurements (e.g., air concentration), or job title or
residence. Studies will be included if biomarkers of vanadium exposure are evaluated but the

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 Vanadium and Compounds (Inhalation) IRIS Assessment

PECO element

Evidence



exposure route is unclear. Other exposure routes including oral will be tagged as "potentially
relevant supplemental information."

Animal: Any exposure to vanadium compound(s) via the inhalation route. Studies involving
exposures to mixtures will be included only if they include an arm with exposure to a singular
vanadium compound alone, otherwise, they will be tagged as "potentially relevant
supplemental information." Other exposure routes, including intratracheal instillation,
intranasal or oropharyngeal administration, oral, dermal, or injection, will be tagged as
"potentially relevant supplemental information."

Comparators

Human: A comparison or referent population exposed to lower levels (or no
exposure/exposure below detection limits) to vanadium compounds, or exposure for shorter
periods of time, or cases versus controls. However, worker surveillance studies are considered
to meet PECO criteria even if no 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 in nonoccupational or occupational settings will be tagged as "potentially relevant
supplemental information."

Animal: A concurrent control group exposed to vehicle-only treatment, untreated control, or
other treatment group with a different exposure duration.

Outcomes

All health outcomes (both cancer and noncancer). In general, endpoints related to clinical
diagnostic criteria, disease outcomes, histopathological examination, or other
apical/phenotypic outcomes are considered to meet PECO criteria and are prioritized for
evidence synthesis over outcomes such as biochemical measures.

PKorPBPK
models

Studies describing pharmacokinetic (PK) or physiologically based pharmacokinetic (PBPK)
models for any form of vanadium will be included.

Classical Pharmacokinetic (PK) 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, where movement of a chemical into, between, and out of the compartments is
quantified empirically by fitting model parameters to ADME (absorption, distribution,
metabolism, and excretion) data. This category is for papers that provide detailed descriptions
of PK models, that are not a PBPK model.

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

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

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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

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

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

3 assessment (e.g., individual studies that contribute to a well-established scientific conclusion).

4 Because it is often difficult to assess the impact of individual studies tagged as supplemental

5 material on assessment conclusions at the screening stage, the tagging structure, allows for easy

6 retrieval later in the assessment process. Table 4-2 presents the supplemental tagging structure

7 presented in the July 2021 IAP. This structure was slightly refined to align with the IRIS Handbook

8 methodsfU.S. EPA. 20221 and the updated supplemental material tagging structure used in the

9 draft assessment is presented in Section 5.2.

Table 4-2. Categories of Potentially Relevant Supplemental Material

Category

Evidence

Toxicokinetic (ADME)

Toxicokinetic (ADME) studies are primarily controlled experiments,
where 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 (D), metabolized (M), or excreted/eliminated (E) through
urine, breath, feces.

The most informative studies are by the inhalation route and 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
timepoint, however, also inform distribution.

ADME data also can be collected from human subjects who have had
environmental or workplace exposures that are not quantified or fully
defined. To be useful, however, such data must involve either repeated
measurements over a 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, separately tracking
the in vitro ADME studies may be appropriate.

*Studies describing environmental fate and transport or metabolism in
bacteria are not tagged as ADME.

Exposure characteristics (no health
outcome assessment)

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).

Mixture studies

Mixture studies that are not considered to meet PECO criteria 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.

Case reports

Case reports of fewer than three subjects that describe health outcomes
after exposure.

Records with no original data

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

Conference abstracts/abstract only

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

4.3. USE OF EXISTING ASSESSMENTS

2 The ATSDR Toxicological Profile for Vanadium (ATSDR. 2012) was selected as the starting

3 point for the literature search because it is the most recent review of health effects of vanadium and

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

4-4 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

compounds published by a U.S. federal government agency that has undergone public comment and
external peer review. In addition, reference lists from existing assessments (final or publicly
available drafts) were manually screened. References were identified from: PPRTV assessment of
vanadium pentoxide fU.S. EPA. 20081. PPRTV assessment of vanadium and its soluble compounds
other than vanadium pentoxide fU.S. EPA. 20091. IRIS External Review Draft assessment of
vanadium pentoxide fU.S. EPA. 2011bl. International Agency for Research on Cancer (IARC)
document on vanadium pentoxide flARC. 20061 as well as references pertinent to vanadium from
the most recent Integrated Science Assessment for Particulate Matter fU.S. EPA. 2019cl. All
references from the 2012 ATSDR Toxicological Profile for Vanadium, literature searches, and other
relevant assessments were extracted by an EPA information specialist and stored in the Health and
Environmental Research Online (HERO) database.5

4.4. LITERATURE SEARCH STRATEGIES

4.4.1. Core Database Searches

Database searches were conducted to identify records that had been published since
development of the 2012 ATSDR Toxicological Profile for Vanadium. The sources listed below were
searched for records published between 2010 and 2021. The start date of 2010 was selected to
ensure records published near the time of release of the ATSDR document were captured.

• PubMed (National Library of Medicine)

• Web of Science (Thomson Reuters)

• Toxline (National Library of Medicine)6

The database searches focused only on the chemical name (and synonyms or trade names)
with no additional limits. The search terms were based on previous vanadium review efforts by
IRIS and were reviewed carefully to ensure that a wide array of vanadium compounds were
encompassed. Because each database has its own search architecture, the resulting search strategy
was tailored to account for each database's unique search functionality. The detailed search
strategies are presented in Appendix A. Literature searches are conducted using EPA's Health and
Environmental Research Online (HERO) database.

The database searches will be updated throughout the assessment's development and
review process to identify newly published literature. The last full literature search update is
conducted several months prior to the planned release of the IRIS draft assessment for public

5Health and Environmental Research Online: https://hero.epa.gov/hero/.

6The Toxline database was migrated to PubMed prior to the March 2020 literature search update, so the
Toxline search was conducted only in March 2019.

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

4-5 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

comment. The results returned (i.e., the number of references from each electronic database or
other literature source described in 4.4.2 and 4.4.3), including the results of any literature search
updates, are documented in the literature flow diagrams (see Section 4.4.2), which also reflect the
literature screening decisions. The IRIS Program takes extra steps to ensure identification of
pertinent studies by encouraging the scientific community and the public to identify additional
studies and ongoing research and by considering late breaking studies that would impact the
credibility of the conclusions, even during the review process. Studies identified after peer review
begins are considered for inclusion only if they meet the PECO criteria and could fundamentally
alter the assessment's primary conclusions or address key uncertainties fU.S. EPA. 20221.

4.4.2. Searching Other Sources

The literature search strategies described above are designed to be broad, but like in any
search strategy, studies can be missed [e.g., cases where 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 may have been missed in the
database search. Records that appear to meet the problem formulation PECO criteria are uploaded
into the screening software, annotated with respect to source of the record, and screened using the
methods described in Appendix B. The list of other sources consulted includes:

•	Manual review (at the title level) of reference list in studies screened as meeting problem
formulation PECO after full-text review.

•	Manual review (at the title level) of the reference list from publicly available final or draft
assessments from EPA (e.g., IRIS and PPRTV) and other non-EPA Agencies (e.g., IARC
[International Agency for Research on Cancer]) or published journal review specifically
focused on human health.

•	References from EPA's Toxicity Values database (ToxValDB), accessed via EPA's CompTox
Chemicals Dashboard fhttps: //comptox.epa.gov/dashboard/]. to identify studies or
assessments that present point of departure (POD) information. ToxValDB collates publicly
available toxicity dose-effect related summary values typically used in risk assessments,
many of which are from "gray literature" and are not available in databases such as Pub
Med or Web of Science. These include POD data collected from data sources within EPA's
ACToR (Aggregated Computational Toxicology Resource) and ToxRefDB (Toxicity
Reference Database), and no-observed and lowest-observed (adverse) effect levels (NOEL,
NOAEL, LOEL, LOAEL) data extracted from repeated dose toxicity studies submitted under
European Union (EU) REACH regulation (Registration, Evaluation, Authorisation and
Restriction of Chemicals). Also included are RfDs from EPA's IRIS and dose descriptors from
EPA's PPRTV documents. Acute toxicity information is extracted from several different
sources, including OECD eChemPortal, ECHA (European Chemicals Agency), NLM (National
Library of Medicine) HSDB (Hazardous Substances Data Bank), ChemlDplus via EPA TEST
(Toxicity Estimation Software Tool), and the EU JRC (Joint Research Centre) AcutoxBase.
Data from the EU COSMOS project (Integrated in Silico Models for the Prediction of Human

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

4-6	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Repeated Dose Toxicity of COSMetics to Optimise Safety) have also been included in
ToxValDB. Although many of the resources included in the "Other Sources Consulted" list
are represented in ToxValDB, they are also manually searched because most of the
ToxValDB entries have not undergone quality control to ensure accuracy or completeness
and might not include recent studies.

• European Chemicals Agency (ECHA) registration dossiers to identify data submitted by
registrants http://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 (TSCA) 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 For Your Information (FYI; voluntary documents).
Other databases accessible via ChemView include EPA's High Production Volume (HPV)
Challenge database

(https: //sor.epa.gov/sor internet/registry/substreg/list/details.do?listId=74) and the
Toxic Release Inventory database.

• National Toxicology Program (NTP) Chemical Effects in Biological Systems (CEBS) database
of study results and research projects.

• The Organisation for Economic Co-operation and Development (OECD) Screening
Information DataSet (SIDS) High Production Volume Chemicals
http://www.inchem.org/pages/sids.html.

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

• The EPA CompTox (Computational Toxicology Program) Chemical Dashboard (U.S. EPA.
2019b) to retrieve a summary of any ToxCast or Tox21 high-throughput screening
information. This data can be used to generate mechanistic insight, predict outcomes using
appropriate models, and potentially inform dose-response modeling. The data's importance
for outcome prediction and dose-response modeling depends on the context, size, and
quality of retrieved results and the lack of availability of other data typically used for these
purposes.

• References identified by the nominating program office, during public comment periods, by
technical consultants, and during peer review.

4.4.3. Non-Peer Reviewed Data

IRIS assessments rely mainly on publicly accessible, peer-reviewed studies. However, it is
possible that unpublished data directly relevant to the PECO may be identified during assessment
development In these instances, the EPA will try to get permission to make the data publicly
available (e.g., in HERO); data that cannot be made publicly available are not used in IRIS
assessments. In addition, on rare occasions where unpublished data would be used to support key
assessment decisions (e.g., deriving a toxicity value). EPA may obtain external peer review if the
owners of the data are willing to have the study details and results made publicly accessible, or if an

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

4-7 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

unpublished report is publicly accessible (or submitted to EPA in a nonconfidential mannerlfU.S.
EPA. 20151.This independent, contractor-driven peer review includes an evaluation of the study
similar to that for peer review of a journal publication. The contractor would identify and typically
select two to three scientists knowledgeable in scientific disciplines relevant to the topic as
potential peer reviewers. Persons invited to serve as peer reviewers would be screened for conflict
of interest. In most instances, the peer review would be conducted by letter review. The study
authors are informed of the outcome of the peer review and given an opportunity to clarify issues
or provide missing details. The study and its related information, if used in the IRIS assessment,
would become publicly available. In the 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 as long as 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).

4.5. LITERATURE SCREENING STRATEGY

Records identified from the literature searches are housed in HERO. 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. Briefly, 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
(https://www.sciome.com/swift-review/searchstrategies/). 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 title
and abstract (TIAB) and full-text screening in DistillerSR (Evidence Partners;
https://distillercer.com/products/distillersr-systematic-review-software/), as described below.
The impact of application of the SWIFT evidence stream filters on the number of studies for TIAB
screening is presented in Figure 4-1.

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

4-8 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

4.5.1. Title and abstract-level screening

The studies prioritized by SWIFT Review are imported into DistillerSR software for TIAB
screening by two independent reviewers. Reviewers complete a structured form asking whether a
study meets PECO criteria or contains potentially relevant supplemental material. Studies
considered relevant or "unclear" based on meeting all PECO criteria at the TIAB level are
considered for inclusion and advanced to full-text screening.

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.

4.5.2. Full-text-level screening

Full-text references are sought through EPA's HERO database for studies screened as
meeting problem formulation PECO criteria, or "unclear" based on TIAB screening. Full-text
screening occurs in DistillerSR. Full-text copies of these citations 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 citations are documented, e.g., study did not meet
problem formulation 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. Conflicts between screeners in applying the supplemental material tags are
resolved similarly, erring on the side of over tagging. Note that more granular sub-tagging of
supplemental material occurs during preparation of the literature inventory as described in Table
4.2.

4.5.3. Multiple Citations with the Same Data

When there are multiple citations using the same or overlapping data, all citations are
included, with one selected for use as the primary citation; the others are considered as secondary
publications with annotation in HAWC and HERO indicating their relationship to the primary
citation during data extraction. For epidemiology studies, the primary citation 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 citation is typically the one with the longest duration of exposure, the
largest sample size, or with the outcome(s) most informative to the PECO. For both epidemiology
and animal studies, the assessment includes relevant data from all citations of the study; although,
if the same outcome is reported in more than one citation, the data are extracted only once. For

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

4-9 DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

1	corrections, retractions, and other companion documents to the included citations, a similar

2	approach to annotation is taken and the most recently published data are incorporated into the

3	assessments.

4.5.4. Literature Flow Diagram

4	The results of the screening process are posted on the project page for the assessment in

5	the HERO database fhttps: //heronet.epa.gov/heronet/index.cfm/proiect/page/project id/29521

6	and studies have been "tagged" with appropriate category descriptors (e.g., included, excluded,

7	potentially relevant supplemental material). Results for SEM screening against the problem

8	formulation PECO are also summarized in a literature flow diagram (see Figure 4-1).

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

4-10	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

2012 ATSDR
Toxicological Profile
for Vanadium,
n = 363

Database Searches (2010 - 2022*),
n = 29,976 after duplicate removal

PubMed
(n = 5,664

WOS
= 29,092

ToxLine3
(n = 15)

I

n = 30,332 records after duplicate removal

SWIFT Review software

Identification of potentially relevant

studies based on application of
SWIFT-Review evidence stream and
health outcome tags
n = 3,778

T

TITLE AND ABSTRACT SCREENING

Title & Abstract Screening

(n = 3,733)	

w

ISA PM studies (n =

"=27) |

FULL TEXT SCREENING

Full-Text Screening

(n= 396)

1

Studies meeting PECO criteria (n = 126)

•	Human health effects studies (n = 95)

•	Animal health effect studies (n = 31)

•	PBPK models (n = 0)

Studies meeting PECO criteria that also
reported potentially relevant
supplemental (n=40)

*Studies from the most recent search update
have not yet been included in the inventory.

Other Resources

200S & 2009 PPRTV
(n = 75)

2011 IRIS
(n = 49)

2006 IARC
(n = 241)

OAR
(n = 10)

n = 4 additional records identified through
curation of references cited in a review
article (Bishayee et al. 2010)

Excluded (n= 2,269)

Tagged as Supplemental (n= 1,095)

Excluded (n= 77)

Tagged as Supplemental (n= 193)

Sum of excluded and supplemental (n= 270)

Tagged as Supplemental {n= l,32Sb)
mechanistic (nongenotoxic) (n =607),
mechanistic (genotoxic) (n = 95),
non-mammalian model (n = 19),
non-inhalation route (n = 392),
ADME/TK (n = 87),
exposure characteristics (n = 116),
mixture studies (n = 46),
case studies or case series (n = 9),
no original data (n = 170),
conference abstract (n = 30)
non-English studies (n = 2)

PM studies (n = 44 )
acute animal studies (n = )

Sum of excluded and supplemental (n= 3,364)



Figure 4-1. Literature search flow diagram for vanadium and compounds.

aThe Toxline database was migrated to PubMed after the 2019 literature search update, thus it was not included in
subsequent literature search updates.

bThese numbers represent the total number of unique citations that were identified; because some citations are
given multiple tags, the sum of the individual material tags is greater than the total number of citations.

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

28

29

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

4.6.	LITERATURE INVENTORY

During title/abstract or full-text-level screening, citations that meet problem formulation
PECO criteria are categorized by evidence type (human or animal) or category of supplemental
information (e.g., mechanistic, PBPK, ADME). Next, study design details for citations that meet
problem formulation PECO criteria are summarized as described in Section 4.6.1. A more granular
tagging of supplemental material may be conducted as described in Section 4.6.2. The results of this
categorization and tagging are referred to as the literature inventory and is the key analysis output
of the SEM.

4.6.1.	Studies That Meet Problem Formulation PECO Criteria

Human and animal studies that met problem formulation PECO criteria after full-text
reviews are briefly summarized using DistillerSR Hierarchical Data Extraction (HDE) forms to
create literature inventories which were used to display the extent and nature of the available
evidence. Data extraction details for the literature inventory are presented in Section 7. These study
summaries are exported from DistillerSR in Excel format and imported into Tableau software
(https://www.tableau.com/) to create interactive literature inventory visualizations. The literature
inventories are used to inform the assessment PECO criteria and evaluation plan. More detail on the
process of summarizing studies is presented in Section 7 (Data Extraction of Study Methods and
Results).

4.6.2.	Organizational Approach for Supplemental Material

Inventories may also be created for other categories of studies that were tagged as
"potentially relevant supplemental material" during screening, including mechanistic studies (e.g.,
in vitro or in silico models), ADME studies, and other studies that do not meet the specific PECO
criteria but that may still be relevant to the research question(s). Here, the objective is to create an
inventory of studies that can be tracked and further summarized as needed—for example, by model
system, key characteristic [e.g., of carcinogens, fSmith etal.. 20161] mechanistic endpoint, or key
event—to support analyses of critical questions that arise at various stages of the systematic review.
The analysis of biological processes underlying vanadium-induced respiratory lesions, including
lung tumor formation, was identified as a key science issue during problem formulation (see
Section 2.4). Studies tagged as containing mechanistic information are inventoried to identify and
organize data that can be used to evaluate the MOA(s) for vanadium induced respiratory effects.

4.7.	INITIAL LITERATURE INVENTORIES FOR VANADIUM (INHALATION)

Literature inventories for PECO-relevant citations were created to develop summary-level,
sortable lists that include some basic study design information (e.g., study population, exposure

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

4-12	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

1	information such as doses administered or biomarkers analyzed, age/life stage7 of exposure,

2	endpoints examined). These literature inventories facilitate subsequent review of individual studies

3	or sets of studies by topic-specific experts. The literature inventory of studies meeting the problem

4	formulation PECO criteria are presented in Figures 4-2 and 4-3 for human and animal studies,

5	respectively. An interactive version of these figures, including additional study design details and a

6	high-level summary of the results is available here.

7Age/life stage of chemical exposure are considered according to EPA's Guidance on Selecting Age Groups for
Monitoring and Assessing Childhood Exposures to Environmental Contaminants and EPA's A Framework for
Assessing Health Risk of Environmental Exposures to Children.

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

4-13	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Epidemiological Studies Examining Inhalation Exposure to Vanadium by Population, Study Design, and Health System

Heat Map	References

Health System

children <18 yrs

pregnant women

Adults - general
population

Adults - occupational

Grand Total

Akkas etal., 2020
Alqhazo and Rashaid, 2018

Cancer





9



9

Alves etal., 2020

Cardiovascular

1

2

9

5

17

Bai fit al., 2019

Dermal







3

3

Barthetal., 2002

Developmental

7

1





8

Bell etal., 2010

Gastrointestinal





1



1

Blaurock-Busch et al., 2012

Hematologic





2

2

5

Boice, Mumma, and Blot, 20..

Hepatic







3

3

Brier etal., 2020

Immune

3



2

3

8

Cal ogero et al., 2021

Metabolic

1

3

8

1

13

Chehbant et al., 2020

Nervous

7



7

2

16

*-« —

Ocular
Renal





1

4
4

5
5

Study Design

Reproductive



3

6



9

case-control

Respiratory

1



4

15

20

cohort

Syst e m i c/V.' ho1 e Body





1

2

3

cross-sectional

Other

2



2

5

8

ecologic

nested case-control

Grand Total

19

9

46

22

95

O
©
Q
O

o

Q
Q
O
O
O
Q

Note: Co! jmi totals, row totals arc grand totals i-idicate total numbersof disti-

Study Details

Exposure Measurement

% I

§" | j|

Health System	Chemical Name	Population	Study Design	Sex	Reference	';§ o	o

cross-sectional	both	Zammitetal., 2020	g

vanadium pentoxide

Aduits - general population

Other

male

Zenz and Berg, 1967

¦





Adults - occupational

cross-sectional

male

Kiviluoto et al., 1981

¦

¦





Other

male

Irsigleretal., 1999

¦

¦

Metabol ic not specif ied

Adults - general population

case-control

both

Li etal., 2017



¦









Lv et al., 2020



¦









Pedro etal., 2019



¦









Tinkov etal., 2020



¦









Wang et al., 2014



¦





cohort

female

Wang et al., 2020a



¦





cross-sectional

both

Huang et al., 2021



¦







male

Alves etal., 2020



¦



children <18 yrs

cohort

both

Karakis et al., 2021



¦



pregnant women

case-control

female

Rezaeieta!., 2021



¦





cohort

female

Wang et al., 2013



¦









Zhang et al., 2021a



¦

vanadium pentoxide

Adults - occupational

cross-sectional

male

K luotoetal., 1981

¦

¦

' Jejl 3 not specified

Adults - general population

case-control

both

Krhira et al., 2015



¦









Pagiia et al., 2016



¦







female

Nayloretal., 1S84



a







not reported

Rochfit al, 2013



¦





cross-sectional

both

Gu etal., 2021



¦









Lv etal., 2021



¦









Palfi or =1 ?n?f>



	M

Legend: Exposure Measurement

| ai r	| occupational | biomonitoring | other

Figure 4-2. Inventory heatmap of PECO-relevant vanadium and compounds
(inhalation exposure) human studies by study design and health system. An
interactive version, which includes a list of citations with additional study
details and summary of the results, is available here.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Toxicological Studies Examining Inhalation Exposure to Vanadium by Study Design, Species, and Health System

Heat Map Hover over Study Design column headers and click the small [•] to collapse Species column headers.



acute



short-term



subchronic



chronic







non-

non-











non-





Grand

Health System

human rat rabbit

human

mouse

rat

rabbit

mouse

rat

human

mouse rat

rabbit

Total



irimate

srimate











jrimate







Cancer















1 1



1

Cardiovascular





1

1



2

1



2 1



2

Dermal

















1 1



1

Endocrine





1











1 1



2

Gastrointestinal









1







1 1



2

Hematologic











D

1



1 1



5

Hepatic





1

1

1

2

1



1 1

1

3

Immune

1



1

2





1



1 1



5

Metabolic











1









1

Nervous





WM





2





1 1



7

Renal





i

1





1



1 1

1

3

Reproductive





mm

1



1

1



1 1

5

Respiratory

1 1

1

2

ma

1

2

1

1

1 1

1

11

Systemity'Whole Body

1



mm

m



1

1



1 1



7

Other







i











1

Grand Total

12 1

1

9

6

1

11

1

1

4 1

1

31

Note: Column totals., row totals, and grand totals indicate total numbers of distinct references.

Study Details

References

Avila-Costa et al.r 2004
Avila-Costa et al., 2005
Avila-Costa eta I., 2006
Bae et al., 2020
Cano-Gutierrez et al., 2012
Cohen etal., 1996

Multidose Study

yes
no

not reported
Grand Total

Chemical(s) Evaluated

Ammonium metavanadate
Bismuth orthovanadate
Sodium metavanadate
Sodium vanadate
Vanadium dioxide
Vanadium pentoxide
Grand Total

Health System

Cancer

Cardiovascular

Route of	Study

Chemical Name Species Sex Exposure	Design Dosing Duration

Vanadium pentoxide mouse both inhalation	chronic 2 y (6 h/d x 5 d/wk)

rat both

All dose levels Dose units Reference

0,1,2,4	mg/mA3 NTP, 2002

inhalation chronic 2 y (6 h/d x 5 d/wk)
Vanadium pentoxide mouse both inhalation chronic 2 y (6 h/d x 5 d/wk)

short-term 16 d (6 h/d x 5 d/wk)

0,0.5,1,2	mg/mA3	NTP, 2002

0,1,2,4	mg/mA3	NTP, 2002

0,2,4, 8,16,32	mg/mA3	NTP, 2002

subchronic 3 mon (6 h/d x 5 d/wk) 0,1,2,4.8,16	mg/mA3	NTP, 2002

90 d (6 h/d x 5 d/wk) 0,16	mg/mA3	Moyer etal., 2002

male inhalation chronic 24 mon (6 h/d x 5 d/wk) 0,4	mg/mA3	Moyer et al., 2002

rat both inhalation chronic 2 y (6 h/d x 5 d/wk) 0,0.5,1,2	mg/mA3	NTP, 2002

short-term 16 d (6 h/d x 5 d/wk) 0,2,4.8,16,32	mg/mA3	NTP. 2002

subchronic 3 mon (6 h/d x 5 d/wk) 0,1,2,4,8,16	mg/mA3	NTP, 2002

0,4,8,16	mg/mA3	NTP, 2002

Vanadium pentoxide mouse both inhalation chronic 2 y (6 h/d x 5 d/wk) 0,1,2,4	mg/mA3	NTP, 2002

rat both inhalation chronic 2 y (6 h/d x 5 d/wk)

„	..	^	.u_	both inhalation chronic 2 y (6 h/d x 5 d/wk)

0, 0.5,1,2
0,1, 2, 4

mg/mA3 NTP, 2002
mg/mA3 NTP. 2002

Note: NOELs/LOELs are based on author-reported statistical significance.

©
©
Q
Q
©

Legend: Results



¦ LOEL(s):

| no effect(s) reported

Figure 4-3. Inventory heatmap of PECO-relevant vanadium and compounds
(inhalation exposure) animal studies by study design and health system. An
interactive version, which includes a list of citations with additional study
details and summary of the results, is available here.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

5.REFINE PROBLEM FORMULATION AND SPECIFY
ASSESSMENT APPROACH

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 inventoiy. 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).

5.1. REFINEMENTS TO PECO CRITERIA

Refinements to the problem formulation PECO criteria were made based on the creation of
initial literature inventories, which are presented in here. The assessment PECO criteria (see Table
5-1) 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 vanadium and compounds (inhalation) IRIS assessment will focus on the health
outcome categories identified in the literature inventory that appear to have sufficient information
available to support hazard identification, i.e., respiratory, immune, reproductive, developmental,
hepatic, renal, cardiometabolic, hematologic, nervous, and cancer. It is clear that in the absence of
additional inhalation studies there will not be sufficient evidence to draw conclusions about
gastrointestinal, dermal, ocular, or endocrine effects. Thus, unless more evidence becomes
available, studies on these health outcomes will not undergo study evaluation or evidence synthesis
to inform hazard characterization, and the information relating to those effects will be briefly
summarized at the literature inventory level. Animal toxicological studies reporting effects tagged
as "Systemic/Whole Body" (body weight, food/water consumption, mortality) that do not evaluate
any other health systems also will not undergo study evaluation or evidence synthesis but can be
considered to help interpret findings for other outcomes and will be summarized at the literature
inventoiy level.

Many observational epidemiological studies evaluated health outcomes in relation to
internal biomarkers of vanadium exposure (e.g., total vanadium in blood, urine, nails). However, the
primary route of vanadium exposure is unclear. This body of studies will be evaluated 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

synthesized as part of both the vanadium and compounds oral and vanadium and compounds
inhalation assessments and may contribute to hazard identification. Studies where total vanadium
was measured only as a component of PM, as measured via stationary monitoring sites were tagged
as "potentially relevant supplemental material." Across these studies, there is considerable
potential for exposure misclassification due to spatial and temporal heterogeneity of vanadium
exposures [PM mass has less heterogeneity than the individual components fU.S. EPA. 2019cl]:
thus, exposure estimates from stationary monitoring stations may not adequately distinguish
between individuals. In addition, there are concerns for confounding by other PM components,
which are difficult to disentangle given their high correlations. Even when multi-pollutant modeling
is performed, there is potential for amplification bias from highly correlated co-exposures
fWeisskopf et al.. 20181. Because of these limitations, these studies are expected to be of lower
confidence overall and are unlikely to provide sufficient evidence of an association with a health
effect on their own. These studies will only undergo study evaluation and inclusion in evidence
synthesis if they can inform the evaluation of an outcome that has evidence of adversity based on
other epidemiology or animal toxicology data.

Table 5-1. Assessment PECO for the vanadium and compounds (inhalation)
assessment

PECO element

Evidence

Populations

Human: Any population and life stage (occupational or general population, including children
and other potentially sensitive populations).

Exposu res

Relevant forms: All forms of vanadium except alloys. Vanadium alloys will be tracked as
supplemental.

Human: Exposure to vanadium compound(s) via the inhalation route, either explicitly stated or
considered plausible based on exposure assessment. Exposure can be based on administered
concentration, biomonitoring data (e.g., urine, blood, or other specimens), environmental or
occupational measurements (e.g., air concentration), or job title or residence. Studies will be
included if biomarkers of vanadium exposure are evaluated but the exposure route is unclear.
Other exposure routes including oral will be tagged as "potentially relevant supplemental
information." Human inhalation exposure to vanadium (as a component of PM) measured via
population stationary air monitoring sites will be tagged as "potentially relevant supplemental

material" due to high potential for exposure misclassification (see Section 6.2.1).

Animal: Any exposure to vanadium compound(s) via the inhalation route. Studies involving
exposures to mixtures will be included only if they include an arm with exposure to a singular
vanadium compound alone, otherwise, they will be tagged as "potentially relevant
supplemental information." Other exposure routes, including intratracheal instillation,
intranasal or oropharyngeal administration, oral, dermal, or injection, will be tagged as
"potentially relevant supplemental information." Acute studies (<24 hours) will be included in
the literature inventory as they can be helpful to interpret findings from studies more directly

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 Vanadium and Compounds (Inhalation) IRIS Assessment

PECO element

Evidence



informative for developing a chronic toxicity value; however, these studies will be tagged as

potentially relevant supplemental material and will not undergo study evaluation or full data

extraction.

Comparators

Human: A comparison or referent population exposed to lower levels (or no
exposure/exposure below detection limits) to vanadium, or exposure for shorter periods of
time, or cases versus controls. However, worker surveillance studies are considered to meet
PECO criteria even if no 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 in
nonoccupational or occupational settings will be tagged as "potentially relevant supplemental
information."

Animal: A concurrent control group exposed to vehicle only treatment, untreated control, or
other treatment group with a different exposure duration.

Outcomes

Health outcomes: respiratory, immune, reproductive, developmental, hepatic, renal.

cardiometabolic, hematologic, nervous, and cancer. In general, endpoints related to clinical
diagnostic criteria, disease outcomes, histopathological examination, or other
apical/phenotypic outcomes are considered to meet PECO criteria and are prioritized for
evidence synthesis over outcomes such as biochemical measures.

PKorPBPK
models

Studies describing pharmacokinetic (PK) or physiologically based pharmacokinetic (PBPK)
models for any form of vanadium will be included.

Classical Pharmacokinetic (PK) 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, where movement of a chemical into, between, and out of the compartments is
quantified empirically by fitting model parameters to ADME (absorption, distribution,
metabolism, and excretion) data. This category is for papers that provide detailed descriptions
of PK models, that are not a PBPK model.

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

Underlined text show modifications in the assessment PECO criteria compared to the problem formulation 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

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

5-3	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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

documented at the level of full-text exclusions in literature flow diagrams with a rationale of
"critical reporting limitation."

A similar approach is taken for in vitro studies that are prioritized for focused analysis
during assessment development (i.e., the critical reporting deficiency may preclude them from
consideration). For each piece of information, if the information can be inferred (when not directly
stated) for an exposure/endpoint combination, the study should be included.

Critical reporting information for different study types are summarized below:

Epidemiology studies

•	Sample size

•	Exposure characterization and/or measurement method

•	Outcome ascertainment method

•	Study design
Animal studies

•	Species

•	Test article name

•	Levels and duration of exposure

•	Route of exposure

•	Quantitative or qualitative (e.g., photomicrographs; author-reported lack of an effect on the
outcome) results for at least one endpoint of interest

In vitro studies prioritized for focused analysis

•	Cell/tissue type(s) or test system

•	Test article name

•	Concentration and duration of treatment

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

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

5-4	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

5.2. REFINEMENTS TO SUPPLEMENTAL CONTENT SCREENING CRITERIA

As noted in the refinements to PECO criteria (see Section 5.1), studies evaluating exposure
to vanadium as a component of PM and animal studies with acute (<24 hour) exposure durations
will be considered as potentially relevant supplemental material in the assessment. A revised list of
supplemental content screening criteria that includes these two categories is presented in Table 5-
2.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Table 5-2. Categories of potentially relevant supplemental material

Category(Tag)

Description

Typical Assessment Use

Pharmacokinetics Data Potentially Informative to Assessment Analyses

Pharmacokinetic (ADME)

Pharmacokinetic (ADME) studies are primarily controlled experiments,
where 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 (D), metabolized (M), or excreted/eliminated (E) through
urine, breath, feces.

•	The most informative studies are by the inhalation route and
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 timepoint, however, also inform distribution.

•	ADME data also can be collected from human subjects who have
had environmental or workplace exposures that are not
quantified or fully defined. To be useful, however, such data
must involve either repeated measurements over a 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, separately tracking the in vitro ADME
studies may be appropriate.

*Studies describing environmental fate and transport or metabolism in
bacteria are not tagged as ADME.

ADME studies are inventoried and prioritized for
possible inclusion in an ADME synthesis section on
the chemical's PK properties and for conducting
quantitative adjustments or extrapolations (e.g.,
animal-to-human). Specialized expertise in PK is
necessary for inventory and prioritization.

Standard operating procedures for PBPK/PK model
evaluation and the identification, organization, and
evaluation of ADME studies are outlined in yAn
Umbrella Quality Assurance Project Plan (QAPP)for
PBPKmodels (U.S. EPA, 2018b).

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Category(Tag)

Description

Typical Assessment Use

Supplemental Evidence Potentially Informative to Assessment Analyses

Mechanistic studies

Studies reporting measurements related to a health outcome that inform
the biological or chemical events associated with phenotypic effects, in
either mammalian and nonmammalian model systems, including in vitro,
in vivo (by any routes of exposure, includes transgenic models), ex vivo,
and in silico studies. Genotoxicity tests are considered "mechanistic."
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).

Prioritized studies of mechanistic endpoints are
described in the mechanistic synthesis sections;
subsets of the most informative studies may
become part of the units of analysis. Mechanistic
evidence can provide support for the relevance of
animal effects to humans and biological plausibility
for evidence integration judgments (including MOA
analyses, e.g., using the MOA framework in the US
EPA Cancer Guidelines (U.S. EPA, 2005a)).

Nonmammalian model
systems

Studies in nonmammalian model systems, e.g., zebrafish, birds, C.
elegans.

Studies of non-PECO animal models can be
summarized to inform evaluations of consistency
(e.g., across species), coherence, or adversity;
subsets of the most informative studies may be
included in the unit of analysis. These studies may
also be used to inform evidence integration
judgments of biological plausibility and/or MOA
analyses and thus may be summarized as part of
the mechanistic evidence synthesis.

Non-inhalation route of
administration

Studies in which humans or animals (whole organism) were exposed via a
non-inhalation route (e.g., oral, injection, or dermal) and intratracheal,
intranasal, or oropharyngeal routes of exposure. This categorization
generally does not apply to epidemiological studies in which the
exposure route may be unclear; such studies are considered to meet
PECO criteria when inhalation exposure is plausible (further review of
these studies will include consideration of whether route attribution can
be inferred). Studies evaluating oral exposure to vanadium and
compounds are also under evaluation in a separate IRIS assessment
(https://cfpub.epa.gov/ncea/iris drafts/recordisplav.cfm?deid=348792).

Routes of exposure which fall outside of the PECO
mav be summarized to inform evidence synthesis
and integration judgments, and/or MOA analyses.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Category(Tag)

Description

Typical Assessment Use

Susceptible population

Studies that help to identify potentially susceptible subgroups, including
studies on the influence of intrinsic factors such as sex, lifestage
(including pregnancy and developmental period), or genotype, as well as
some other factors (e.g., health status). These are often co-tagged with
other supplemental material categories, such as mechanistic or ADME.
Studies meeting PECO criteria that also address susceptibility should be
co-tagged as supplemental.

*Susceptibility based on most extrinsic factors, such as increased risk for
exposure due to residential proximity to exposure sources, is not
considered an indicator of susceptible populations for the purposes of IRIS
assessments.

Provides information on factors that might
predispose sensitive populations or lifestages to a
higher risk of adverse health effects following
exposure to the chemical. This information is
summarized during evidence integration for each
health effect and is considered during dose-
response, where it can directly impact modeling
decisions.

PM studies

Human studies evaluating health outcomes associated with the vanadium
component of particulate matter (as measured via air pollution

Studies which provide more detailed exposure
monitoring (e.g., personal air sampling, or

monitoring stations). No study evaluation will be done on these studies.

occupational exposure measurements) will be
considered PECO relevant and will undergo studv
evaluation.

Acute studies

Animal studies with acute exposure durations (defined as less than 24
hours) that otherwise meet PECO criteria.

Acute animal studies are retained in the literature
inventory since they can be helpful in interpreting
studies used in developing chronic toxicity values.

Non-English language
studies

Records in foreign language with the abstracts in English.

For non-English language 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.

Background Information Potentially Useful to Problem Formulation and Protocol Development
(These studies fall outside the scope of IRIS assessment analyses)

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Category(Tag)

Description

Typical Assessment Use

Exposure characteristics (no
health outcome assessment)

This information may be useful for developing
exposure criteria for study evaluation or refining
problem formulation decisions.

Notably, providing an assessment of typical human
exposures (e.g., sources, levels) falls outside the
scope of an IRIS assessment.

Mixture studies

Animal studies which included co-exposure to multiple chemicals are not
considered to meet PECO criteria because they do not contain an
exposure or treatment group assessing only the chemical of interest.

Mixture studies are tracked to help inform
cumulative risk analyses, which may provide useful
context for risk assessment but fall outside the
scope of an IRIS assessment.

Case studies or case series

Case reports of 1-3 subjects that describe health outcomes after
exposure.

Tracking case studies can facilitate awareness of
potential human health issues missed by other
types of studies during problem formulation.

Reference Materials

Records with no original
data

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

Studies that are tracked for potential use in
identifying missing studies, background
information, or current scientific opinions (e.g.,
hypothesized MOAs).

Posters or conference
abstracts

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

Underlined text show modifications in the categories of potentially relevant supplemental material since release of the IAP.

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

5-9 DRAFT-DO NOT CITE OR QUOTE

-------
1

2

3

4

5

6

7

8

9

10

11

12

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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

The planned units of analysis based on outcomes identified in the assessment PECO are
summarized in Table 5-2. General considerations for defining the units of analysis are presented in
the IRIS Handbook (U.S. EPA. 20221. Each unit of analysis is initially synthesized and judged
separately within an evidence stream (see Section 8.1)

Evidence integration judgments focus on the stronger within evidence stream synthesis
conclusions when multiple units of analysis are synthesized. The evidence synthesis judgments are
used alongside other key considerations (i.e., human relevance of findings in animal evidence,
coherence across evidence streams, information on susceptible populations or lifestages, and other
critical inferences that draw on mechanistic evidence) to draw an overall evidence integration
judgment for each health effect category or more granular health outcome grouping (see Section
8.2). As new evidence to inform potential vanadium-associated health hazards become available,
updates to the units of analysis will be considered as appropriate.

Table 5-3. Human and animal endpoint grouping categories.

Health Effect Categories
for Evidence Integration

Units of Analysis for Evidence Synthesis Integration
(Each bullet represents a unit of analysis)



Human

Animals

Cancer

• Tumors and precancerous
lesions

• Tumors and precancerous lesions

Cardiometabolic

•	Cardiovascular outcomes (e.g.,
CVD mortality, hospital
admissions)

•	Clinical effects (e.g., blood
pressure, pulse)

•	Serum lipids, glucose, A1C

•	Histopathology

•	Heart weight

•	Serum lipids, glucose

Developmental

•	Fetal viability/pregnancy
outcomes

•	Fetal structural alterations

•	Offspring mortality/survival

•	Offspring growth

•	Developmental milestones (e.g., eye
opening, incisor eruption, etc)

•	Structural alterations (e.g., external,
skeletal, and soft tissue findings)

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Health Effect Categories
for Evidence Integration

Units of Analysis for Evidence Synthesis Integration
(Each bullet represents a unit of analysis)



Human

Animals

Hematologic

• Red blood cell count and size,
hematocrit or hemoglobin,
platelets

• Red blood cell count and size,

hematocrit or hemoglobin, platelets

Hepatic

• Clinical chemistry (e.g., ALT,
AST, ALP, GGT)

•	Histopathology

•	Liver weight

•	Clinical Chemistry (e.g., ALT, AST,
ALP, GGT)

Immune

•	Immune cell counts

•	Functional immune assays

•	Autoimmune response

•	Immune organ weight histopathology

•	Immune cell counts

•	Functional immune assays

Nervous

•	Neurodevelopmental/
neurobehavioral disorders

•	Neurological or sensory
symptoms

•	Histopathology

•	Brain weight

•	Functional observational battery,
including motor activity and reflex
responses

•	Learning and memory

Renal/Urinary

•	Clinical Chemistry (e.g., BUN,
CREA, KIM1, NGAL)

•	Urinalysis (protein, glucose)

•	Renal function (e.g., GFR)

•	Histopathology

•	Organ weight

•	Clinical Chemistry (e.g., BUN, CREA,
KIM1, NGAL)

•	Urinalysis (e.g., protein, glucose)

•	Renal function (e.g., GFR)

Reproductive

•	Fertility

•	Pregnancy outcomes

•	Menstrual disorders

•	Fertility and pregnancy outcomes

•	Histopathology

•	Reproductive organ weight

•	Reproductive hormones

•	Dam body weight/body weight gain

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

5-11	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Health Effect Categories
for Evidence Integration

Units of Analysis for Evidence Synthesis Integration
(Each bullet represents a unit of analysis)



Human

Animals

Respiratory

•	Pulmonary function (e.g., FEV,
FVC, MEF)

•	Asthma incidence/severity

•	Respiratory symptoms (e.g.,
wheezing, irritation, shortness
of breath)

•	Lung weight

•	Histopathology

•	Pulmonary function

ALT = alanine aminotransferase; AST = aspartate aminotransferase; A1C = glycated hemoglobin; BUN = blood urea
nitrogen; CREA = creatinine; CK = creatine kinase; FEV = forced expiratory volume; FVC = forced vital capacity;
MEF = maximal expiratory flow

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

5-12	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

6.STUDY EVALUATION (RISK OF BIAS AND
SENSITIVITY)

The general approach for evaluating the primary health effect studies meeting PECO criteria
for all study types is described in Section 6.1. Instructional and informational materials for study
evaluations are available at https: //hawcprd.epa.gov/assessment/100000039/. The approach is
conceptually the same for epidemiology, animal toxicology, and controlled human exposure, 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(U.S. EPA. 2018b) (see
Section 6.5).

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 Vanadium and Compounds (Inhalation) IRIS Assessment

(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 judgments and overall study rating

Domain judgments

Judgment

Interpretation

0

Good

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



Adequate

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



Deficient

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

•

Critically
Deficient

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 orconcerns identified; potential forbias
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 andthe 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 IRIS study evaluation process, (a) An overview of the
evaluation process, (b) The evaluation domains and definitions for ratings
(i.e., domain and overall judgments, performed on an outcome-specific basis).

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

6-2	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

To calibrate the assessment specific considerations, the study evaluation process includes a
pilot phase to assess and refine the evaluation process. Following this pilot, at least two reviewers
independently evaluate studies to identify characteristics that bear on the informativeness (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: //hawcprd.epa.gov/assessment/l00000039/1 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.

1991).

Authors might be queried to obtain critical information, particularly that involving missing
key study design or results information that or additional analyses that could address potential
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:

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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

•	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.

•	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 pre-defined
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

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

6-4	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 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 useable 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. Critically deficient and Uninformative ratings are
uncommon; these ratings are reserved for critical flaws where the study findings are truly
uninterpretable due to identified biases. The most frequent situation where they are used for
epidemiology studies is when potential confounding has not been considered using any method
(e.g., adjustment, stratification, restriction), including unadjusted correlation coefficients or means
in cases/controls in a heterogeneous population where confounding is likely.

6.2. EPIDEMIOLOGY STUDY EVALUATION

Evaluation of epidemiology studies of health effects to assess risk of bias and study
sensitivity are 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, whereas study sensitivity is typically
concerned with identifying the latter.

The principles and framework used for evaluating epidemiology studies are adapted from
the principles in the Cochrane Risk of Bias in Nonrandomized Studies of Interventions [ROBINS-I;
Sterne etal. (2016)] but modified to address environmental and occupational exposures. The types
of information that may be the focus of those criteria are listed in Table 6-1. Core and prompting
questions, presented in Table 6-2, 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. Exposure- and outcome- specific criteria to use
during study evaluation are developed using the core and prompting questions and refined during a
pilot phase with engagement from topic specific experts. The protocol may also be adjusted in the
early phases of the study evaluation process if corrections are identified based on initial literature

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 Vanadium and Compounds (Inhalation) IRIS Assessment

1 reviews. Exposure and confounding domain considerations specific to vanadium are presented in

2 Sections 6.2.1.

Table 6-1. Information relevant to evaluation domains for epidemiology
studies

Domain

Types of information that may need to be collected or are important for evaluating

the domain

Exposure
measurement

Source(s) of exposure (e.g., consumer products, occupational, an industrial accident) and
source(s) of exposure data, blinding to outcome, level of detail for job history data, when
measurements were taken, type of biomarker(s), assay information, reliability data from
repeated-measures studies, validation studies.

Outcome
ascertainment

Source of outcome (effect) measure, blinding to exposure status or level, how
measured/classified, incident vs. prevalent disease, evidence from validation studies, prevalence
(or distribution summary statistics for continuous measures).

Participant
selection

Study design, where and when was the study conducted, and who was included? Recruitment
process, exclusion and inclusion criteria, type of controls, total eligible, comparison between
participants and nonparticipants (or followed and not followed), and final analysis group. Does
the study include potential susceptible populations or life stages (see discussion in Table 8.6)?

Confounding

Background research on key confounders for specific populations or settings; participant
characteristic data, by group; strategy/approach for consideration of potential confounding;
strength of associations between exposure and potential confounders and between potential
confounders and outcome; and degree of exposure to the confounder in the population.

Analysis

Extent (and if applicable, treatment) of missing data for exposure, outcome, and confounders;
approach to modeling; classification of exposure and outcome variables (continuous vs.
categorical); testing of assumptions; sample size for specific analyses; and relevant sensitivity
analyses.

Sensitivity

What are the ages of participants (e.g., not too young in studies of pubertal development)? What
is the length of follow-up (for outcomes with long latency periods)? Choice of referent group, the
exposure range, and the level of exposure contrast between groups (i.e., the extent to which the
"unexposed group" is truly unexposed, and the prevalence of exposure in the group designated
as "exposed").

Selective
reporting

Are results presented with adequate detail for all the endpoints and exposure measures
reported in the methods section, and are they relevant to the PECO? Are results presented for
the full sample as well as for specified subgroups? Were stratified analyses (effect modification)
motivated by a specific hypothesis?

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Table 6-2. 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

Exposu re
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?

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?

Is the degree of
exposure

misclassification 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 etiological ly 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 etiological ly 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.

Exposure measurement was not independent of outcome
status.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

Follow-up questions

Criteria that apply to most exposures and outcomes

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:

•	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?

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.

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.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

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?

•	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?

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 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/continuation is
likely?

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
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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

Follow-up questions

Criteria that apply to most exposures and outcomes







exposure, recruitment materials stated outcome of interest
and potential participants are aware of or are concerned
about specific exposures).

Confounding

Is confounding of the
effect of the
exposure likely?

Is confounding adequately addressed by
considerations in:

•	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)?

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

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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

Follow-up questions

Criteria that apply to most exposures and outcomes







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
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

• 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

Good

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

Follow-up questions

Criteria that apply to most exposures and outcomes

the necessary
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)?

the effect estimate (if
there is enough
information)?

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

•	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 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).

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

Follow-up questions

Criteria that apply to most exposures and outcomes







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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

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

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

6.2.1. Epidemiology Study Evaluation Considerations Specific to Vanadium

6.2.1.1 Exposure measurement

Exposure to vanadium by the inhalation route may be measured based on occupational
exposure (e.g., job duties), air sampling, or biomonitoring data, or a combination of these. Criteria
for evaluating each of these information types are summarized in Table 6-3, with some additional
considerations described below.

Biomarker measurements of total vanadium may represent exposure via any route. Where
possible, evaluations will indicate the likely predominant route; studies where exposure is likely to
be primarily via inhalation will be given more weight. Measurements from urine, blood, hair, or
toenails will be considered to be relevant to either acute or long-term continuous exposure. Metal
concentrations in hair or toenails may reflect exposures during the previous several months based
on their rate of growth, although the precise exposure window has not been investigated for
vanadium (Gutierrez-Gonzalez et al.. 2019). Toenail vanadium was strongly correlated with
vanadium in hair (r = 0.61) in a study of 26 adults (primarily workers) (Rairiska et al.. 2005).
Validated reference values are available for hair, blood, plasma, and urine using Inductively
Coupled Plasma Mass Spectrometry (ICP-MS) (Goulle etal.. 2005). Quality control procedures
include the use of certified reference material generated by individual laboratories, recovery
analysis, procedural blanks, duplicate samples, or spike samples. Sample mass has been associated
with concentrations measured in toenails; therefore, correction methods are necessary.

Well-established and sensitive methods for measurement of total vanadium concentrations
include measurement using graphite furnace atomic absorption (GF-AAS) (with a preconcentration
procedure), isotope dilution mass spectrometry (ID-MS), inductively coupled plasma mass
spectrometry (ICP-MS), and neutron activation analysis (NAA) with radiochemical separation.
Detection limits of these methods have been summarized previously fATSDR. 20121. Because toxic
properties of vanadium species differ, measurements that report vanadium species are preferred to
measurements of total vanadium. If only total vanadium were measured in the sample media used
for internal biomarker measurement, and there were no other serious limitations in the
measurement of exposure (e.g., invalid measure, inappropriate timing of measurement, inadequate
detail of analysis including quality control), the exposure measurement domain would be rated
adequate rather than good to reflect the reduced sensitivity resulting from combining the effects of
vanadium species.

Occupational exposure to vanadium compounds can occur in a variety of occupational
settings, including mining and/or processing of vanadium ore, maintenance of oil-burning boilers,
and some steel production processes. In most occupational studies of vanadium inhalation
exposures, the vanadium compound is identified in the indicated job category. It is preferred if
these categories are validated by a quantitative measure such as personal sampling of air or
biomonitoring matrices in at least some participants.

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

6-15 DRAFT-DO NOT CITE OR QUOTE

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Vanadium in air is generally measured via regular monitoring for particulate matter (PM)
and its components. However, the spatial distribution for the monitoring of PM in air is limited; in
the U.S., only a few hundred monitoring stations capture data that allow for analysis of individual
PM components (from the Chemical Speciation Network (CSN) and the Interagency Monitoring of
Protection Visual Environments (IMPROVE) systems). Given that vanadium levels are more
spatially and temporally variable than PM mass fU.S. EPA. 2019c! there is considerable uncertainty
as to whether monitoring measurements can provide accurate estimates of individual exposure. In
addition, some studies used land use regression (LUR) models supplemented by a short period of
air monitoring that predicted levels of vanadium and other PM components using variables such as
land density, population density, altitude, traffic intensity, and road network; at this time, none of
these models have been validated for prediction of vanadium and thus their ability to distinguish
exposures is uncertain. A small number of studies measured exposure using personal or home air
samples which may provide more specific estimates of individual exposure when samples are
collected over enough time to capture variability.

Table 6-3 Criteria for evaluating exposure measurement in epidemiology
studies of vanadium

Rating

Criteria

Good

Biomarker measures:

•	Evidence that exposure was consistently assessed using well-established analytical
methods. Well-established and sensitive methods include measurement of total
vanadium using GF-AAS (with a preconcentration procedure), ID-MS, ICP-MS, and NAA
with radiochemical separation.

Occupational measures:

•	For a specific job site(s): Evidence that measurement of current/recent exposure is based
on personal samples (air or biomonitoring). Ideally, this would cover all workers or
randomly selected workers within specific areas/jobs/tasks for at least one full shift,
allowing for examination of variation in exposure among workers at a particular worksite,
but this is not required (i.e., categorization by job duties with validation using personal
samples in a sample of workers is acceptable). OR for long-term exposure, monitoring
data covering a substantial portion of the time period of interest specific to work
locations, job titles, and tasks with information provided on changes in exposure
conditions over time; job histories are available for a substantial period of employment in
exposed jobs.

Air measures:

•	Personal/Home samples: Integrated personal measurements using passive monitors over
multiple 24-hour periods, or time-weighted summary concentrations incorporating
concentrations in residence and school/workplace. OR Area measurements in home using
passive or active monitors, with an average of measurements in one or more rooms
(average over longer periods is preferred, with multiple seasons if estimating annual

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Rating

Criteria

average). For either type of measure, sampling details are provided including type and
placement of samplers, sampling periods, and chemical analysis methods.

And all of the following (where relevant):

• Measurement of vanadium included species or exposure was to a specific species.

• Exposure was assessed in a relevant time-window (i.e., temporality is established, and
sufficient latency occurred prior to disease onset) for development of the outcome.

• There is evidence that a sufficient number of the exposure data measurements are above
the limit of quantification for the assay.

• Details on quality control provided include measures to avoid contamination in sampling,
sample handling and storage of blood and urine samples, sample mass (minimum 10 mg
with adjustment for mass (Gutierrez-Gonzalez et al., 2019)) for toenails. QA statistics on
precision and accuracy reported.

• There is sufficient specificity/sensitivity and range or variation in exposure measurements
that would minimize potential for exposure measurement error and misclassification by
allowing exposure classifications to be differentiated (i.e., can reliably categorize
participants into groups such as high vs. low exposure).

Adequate

Biomarker measures

• Evidence that exposure was consistently assessed using methods described in Good, but
there were some concerns about quality control measures or other potential for non-
differential misclassification.

Occupational measures

• For a specific job site(s): With known exposure to vanadium at the site, evidence that
current/recent exposure based on job duties alone (without personal samples) are used
with a comparison group where exposure levels are known to be low (i.e., similar to
background levels in the general population) or monitoring data is less comprehensive,
raising the possibility of nondifferential misclassification. OR for long-term exposure,
monitoring data is less comprehensive with regard to time, work site, job title or tasks, or
job history data are less complete than described in Good.

• For population-based occupational studies: Job exposure matrix that incorporates
industry, time period, tasks, and material used, and has validation data confirming its
ability to differentiate between exposure levels.

Air measures

• Personal/Home samples: sampling occurs over a shorter period than described in Good, or
some details on sampling and analysis are not provided but appear appropriate.

And all of the following (where relevant):

• Exposure was assessed in a relevant time-window for development of the outcome.

• There is evidence that a sufficient number of the exposure data measurements are above
the limit of quantification for the assay.

• The laboratory analysis included some data on standard quality control measures with
demonstrated precision and accuracy.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Rating

Criteria

Measurement of total vanadium will reduce the rating from Good to Adequate unless the exposure
is known to be a specific vanadium species.

Deficient

Any of the following:

• There is a lack of detail on the sampling or analytical methods that reduces the ability to
assess exposure misclassification.

• There is some concern, but no direct evidence, that the exposure was assessed using
methods that have not been validated or empirically shown to be consistent with
methods that directly measure exposure.

• Exposure was assessed in a relevant time-window(s) for development of the outcome, but
there could be some concern about the potential for bias due to reverse causation
between exposure and outcome, but there is no direct evidence that it is present.

• There is some concern over insufficient specificity/sensitivity and range or variation in
exposure measurements that may result in considerable exposure measurement error
and misclassification when exposure classifications are compared (i.e., data do not lend
themselves to reliably categorize participants into groups such as high vs. low exposure,
and/or there is considerable uncertainty in exposure values which do not allow for
confidence in the examination of small per unit changes in continuous exposures)

Critically
deficient

Any of the following:

• Exposure was assessed in a time-window that is unknown or not relevant for
development of the outcome. This could be due to clear evidence of bias due to reverse
causation between exposure and outcome, or other concerns such as the lack of temporal
ordering of exposure and disease onset, insufficient latency, or having exposure
measurements that are not reliable measures of exposure during the etiologic window(s).

• Direct evidence that bias was likely since the exposure was assessed using methods with
poor validity.

• Evidence of differential exposure misclassification (e.g., differential recall of self-reported
exposure).

• There is evidence that an insufficient number of the exposure data measurements were
above the limit of quantification for the assay.

1 6.2.1.2 Confounding by co-exposures

2 Exposure to vanadium via the inhalation route is typically co-occurring with other chemical

3 exposures. In the general population, overall PM mass and other individual PM components are

4 highly correlated with vanadium (U.S. EPA. 2019c). while in occupational studies, co-exposures

5 depend on the specific job duties. These co-exposures represent potentially important confounders

6 when estimating the effect of an individual component from a larger mixture. The likelihood of

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

6-18 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

confounding by co-exposures will be considered during study evaluation. In order for confounding
to occur, the co-pollutant would need to be associated with both vanadium and the outcome of
interest, and not act as an intermediate in the causal pathway. Thus, where correlations across
exposures are reported, they will be reviewed to assess the likelihood that confounding could
explain the observed results. In addition, many studies, particularly those published recently, may
also have performed multipollutant modeling to adjust for co-exposures. These analyses can
provide additional context, but even when they are available, it is often not possible to fully
disentangle the associations due to high correlations. This stems from the potential for
amplification bias that can occur following adjustment of highly correlated exposures fWeisskopf et
al.. 20181. Thus, in most studies, there may be some residual uncertainty about the risk of
confounding by co-exposures. A Good rating for the confounding domain will be reserved for
situations where there is minimal concern for substantial confounding across co-exposures as well
as other sources of confounding. This could occur in studies where there are robust results
following multipollutant modeling (i.e., minimal change between single- and multi-pollutant
models), which would also indicate minimal concern for amplification bias. Potential confounding
by co-exposures may result in a Deficient rating if there is considerable concern that the observed
effect could be explained by correlated co-exposures.

Because of the challenge in evaluating individual studies for confounding by co-exposures,
this issue will also be assessed across studies during the evidence synthesis phase, primarily when
there is support for an association with adverse health effects in the epidemiology evidence (i.e.,
moderate, or robust evidence in humans, as described below). Analyses may include comparison of
results across studies in populations with different exposure mixture profiles (e.g., general
population vs. occupational) and considering results of multi-pollutant models across studies when
available. In situations where there is considerable uncertainty regarding the impact of residual
confounding by co-exposures, this will be captured as a factor that decreases the overall strength of
evidence (see Section 10.1).

6.3. EXPERIMENTAL ANIMAL STUDY EVALUATION

Using the principles described in Section 6.1, the animal studies of health effects to assess
risk of bias and sensitivity are evaluated for the following domains: allocation, observational
bias/blinding, confounding, selective reporting, attrition, chemical administration and
characterization, endpoint measurement and validity, results presentation and comparisons, and
sensitivity (see Table 6-4).

The rationale for judgments is documented at the outcome level. The evaluation
documentation in HAWC includes the identified limitations and their expected impact on the overall
confidence level. To the extent possible, the rationale will reflect an interpretation of the potential
influence on the outcome-specific results, including the direction or magnitude of influence (or
both).

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 Vanadium and Compounds (Inhalation) IRIS Assessment

1	Vanadium speciation chemistry in animal toxicological studies will be considered in the

2	Exposure methods sensitivity domain. The highest confidence will be placed in studies that report

3	the vanadium compound that was used and have analytical chemistry data indicating the vanadium

4	species present. Considering oxidation status could be important as results from some oral

5	exposure studies in rodents suggest increased toxicity of vanadium in the +5 oxidation state

6	compared to vanadium +4 fRoberts etal.. 20161: fNational Toxicology Program fNTPll. Study

7	evaluations for the available inhalation studies, to the extent possible, will consider factors that

8	could affect vanadium oxidation state and speciation (e.g., study methods that involved aerosolizing

9	vanadium pentoxide from solution, rather than exposure to vanadium pentoxide as a dust).

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Table 6-4. 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 allocation3)?

•	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 where 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-lesionsb).

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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core question

Prompting questions

General considerations

• 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).

• 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?

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.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core question

Prompting questions

General considerations



•	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?

•	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 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

For each study:

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

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.

•	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

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

6-23	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core question

Prompting questions

General considerations

is not a risk of bias
consideration.



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

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?

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

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. 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:

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 Vanadium and Compounds (Inhalation) IRIS Assessment

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 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. *

•	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.

•	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). **

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

These limitations typically also raise a concern for insensitivity

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

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.*

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.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core question

Prompting questions

General considerations

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

*Failure 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.

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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core question

Prompting questions

General considerations





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
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.

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

•	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)

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, 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 towards 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.

6-27	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 towards 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 6.1.1).

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

6-28	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

6.4. CONTROLLED HUMAN EXPOSURE STUDY EVALUATION

This study design involves human volunteers to test specific hypotheses about short-term
exposures and biological responses that inform potential mechanisms and understanding of
exposure-response patterns. The exposures are generated in the laboratory to achieve
predetermined concentrations for periods of minutes to hours. For study evaluation, a process
incorporating aspects of the approaches used for epidemiology studies and experimental animal
studies, as well as the ROBINS-I tool discussed in Section 6.2 fSterne etal.. 20161. are used to
evaluate controlled exposure studies in humans. Controlled human exposure studies are 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), 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. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODEL
DESCRIPTIVE SUMMARY AND 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.

No PBPK models for vanadium and compounds were identified in the survey of the
literature. If the comprehensive literature search or updates to that initial search identify any PBPK
models, they will be evaluated in accordance with the Quality Assurance Project Plan for PBPK
models fIJ.S. FPA. 2020b. 2018b).

6.6. IN VITRO AND OTHER MECHANISTIC STUDY EVALUATION

As described in Section 4.4, the initial literature screening identifies sets of other potentially
informative studies, including mechanistic studies, as "potentially relevant supplemental
information." Mechanistic information includes any experimental measurement related to a health
outcome that informs the biological or chemical events associated with phenotypic effects. These
measurements can improve understanding of the mechanisms involved in the biological effects
following exposure to a chemical but are not generally considered by themselves adverse outcomes.
Mechanistic data are reported in a diverse array of observational and experimental studies across

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

6-29 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 Vanadium and Compounds (Inhalation) IRIS Assessment

species, model systems, and exposure paradigms, including in vitro, in vivo (by various routes of
exposure), ex vivo, and in silico studies.

Individual study-level evaluations of mechanistic endpoints are pursued only in some select
cases. For some chemical assessments, it may be necessary to identify assay-specific considerations
for study endpoint evaluations, on a case-by-case basis, to provide a more detailed summary and
evaluation for the most relevant individual studies. This may be done, for example, when the
scientific understanding of a critical mechanistic event or MOA is less established or lacks scientific
consensus, when the reported findings on a mechanistic endpoint are conflicting, when the
available mechanistic evidence addresses a complex and influential aspect of the assessment, or
when in vitro or in silico data make up the bulk of the evidence base and there is little or no
evidence from epidemiological studies or animal bioassays.

If a subset of individual mechanistic studies is identified for evaluation, the study evaluation
considerations will differ depending on the type of endpoints, study designs, and model systems or
populations evaluated. Note that because the evaluation process is outcome specific, overall
confidence classifications for human or animal studies that have already been determined will not
automatically apply to mechanistic endpoints if reported in the same study; instead, a separate
evaluation of the mechanistic endpoints should be performed because the utility of a study may
vary for the different outcomes reported. Developing specific considerations requires a familiarity
with the studies to be evaluated and cannot be conducted in the absence of knowledge of the
relevant study designs, measurements, and analytic issues. Knowledge of issues related to the
hazards and the outcomes identified in the revised evaluation plan is also important for developing
specific evaluation considerations. One challenge is that novel methodologies for studying
mechanistic evidence are continually being developed and implemented and often no "standard
practices" exist.

The evaluation of mechanistic studies applies similar principles as those described above
for the evaluation of experimental animal studies. Table 6-5 provides the standard domains and
core questions for evaluating studies conducted in in vitro test systems, along with some basic
considerations for guiding the evaluation. The evaluation process focuses on assessing aspects of
the study design and conduct through three broad types of evaluations: reporting quality, risk of
bias, and study sensitivity. Some domain considerations are tailored to the chemical, as well as the
assay(s) and/or endpoint(s) being evaluated. Assessment teams work with subject-matter experts
to develop specific considerations. These specific considerations are determined before performing
the study evaluation, although they may be refined as the study evaluation proceeds (e.g., during
pilot testing). Assessment-specific and/or assay-specific considerations are documented and made
publicly available in the assessment.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Table 6-5. 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

For each study:

• Are there any known or presumed
differences across treatment groups
(e.g., co-exposures, culture conditions,

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.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations

controlled for and consistent
across experimental groups?

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 feature inherent
to the physicochemical properties of
the test substance(s) that have the
potential to bias the 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
characterization.)

•	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?

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 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

For each study:

• Are results presented for all

endpoints/outcomes described in the
methods?

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.

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

6-32	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations

endpointsand replicates
described in the methods?
Note: The appropriateness
of the analysis or results
presentation is considered
under results presentation.

•	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?

•	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.

•	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

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.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations



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?

•	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).

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.

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 feature inherent

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

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

6-34	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations

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

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?

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:

•	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.*

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

6-35	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations





• 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)**

Controls are not included or considered inappropriate.

*These limitations typically also raise a concern for insensitivity

**Sample size alone is not a reason to conclude an individual study is critically

deficient.

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.*

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.

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

6-36	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations





•	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 data* (e.g., presentation of mean
without variance data; concurrent control data are not presented;
failure to report or address overt cytotoxicity).

*Failure to describe any findings for assessed outcomes (i.e., report lacks any
qualitative or quantitative description of the results in tables, figures, 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.

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

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?

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

6-37	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations

of detection), frequency, and duration
of exposure sensitive for the
assay/model system of interest,
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?

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
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 are 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

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

6-38	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Domain and core
question

Prompting questions

General considerations





•	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).

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).

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

6-39	DRAFT-DO NOT CITE OR QUOTE

-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 problem formulation PECO criteria after full-text review are briefly summarized in
DistillerSR HDE forms. These study summaries are exported from DistillerSR in Excel format and
imported into Tableau software (https://www.tableau.com/) to create interactive literature
inventoiy visualizations used to display the extent and nature of the available evidence. (See below
for studies decisions related to studies meeting the assessment PECO).

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 days]8 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. Animal studies that meet
the assessment PECO undergo a subsequent phase of full data extraction in HAWC that includes
detailed presentation of results (described below). For studies that meet problem formulation
PECO criteria (but not the assessment PECO) the SEM (initial) literature inventoiy summary
includes the no-observed-effect level/low-observed-effect level (NOEL/LOEL) based on author-
reported statistical significance. Expert judgment may be used to identify NOEL/LOELs in cases
where only qualitative results are reported (e.g., "no effects on liver weight were observed at any
dose level") or when the findings indicate an apparent clear and strong effect of exposure (e.g.,
large magnitude of change) but the authors did not present a statistical comparison. When findings
are not analyzed by the authors and are not readily interpretable, then NOEL/LOELs are not
identified, and the extraction field entry indicates "not reported."

For human studies, the following information is summarized in DistillerSR HDE 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, specific
endpoints assessed and a quantitative summary of findings at the endpoint level (or narrative only
if the finding was qualitatively presented). In contrast to the animal studies, epidemiological studies

8EPA 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

that met assessment PECO did not undergo additional more detailed data extraction in HAWC
because that module in HAWC was under development at the time of preparation of this protocol.

For animal studies that met the assessment PECO criteria, HAWC is used for full extraction
of study methods and results. For animal studies, compared with the literature inventory forms
used to describe studies that meet problem formulation PECO criteria, full data extraction in HAWC
includes summarizing more details of study design (e.g., diet, chemical purity) and gathering effect
size information. Instructions on how to conduct data extraction in HAWC are available at
fhttps://hawcp roiect.org/resources A Over 100 distinct extraction fields are collected for each
animal study and endpoint (for list of data extraction fields, see Downloads > Animal Bioassay Data
> Complete Export at the HAWC Vanadium and Compounds (Inhalation) Project
fhttps://hawc.epa.gov/assessment/100500286/). An additional resource used to implement use of
a consistent vocabulary to summarize endpoints assessed in animal studies is available in the
HAWC project "IRIS PPRTV SEM Template Figures and Resources" (see "Attachments," then select
the "Environmental Health Vocabulary (EHV)— a recommended terminology for
outcomes/endpoints" file).

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.

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
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 language 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

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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 dose-response to facilitate
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.

Exposures are standardized to common units when possible. For hazard characterization,
exposure levels are typically presented as reported in the study and standardized to common units
(e.g., ppm or mg/m3 for inhalation studies) as an initial phase in evidence synthesis and integration.
For inhalation exposures to vanadium, concentration in air will be reported as mg vanadium/m3.

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

7-3	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

8.EVIDENCE SYNTHESIS AND INTEGRATION

Evidence synthesis9 is a within-stream analysis, conducted separately for human, animal,
and mechanistic evidence. Findings from human and animal evidence for each unit of analysis are
separately judged to reach an expression of certainty in the evidence for a hazard (robust, moderate,
slight, indeterminate, or compelling evidence of no effect). Within-stream evidence synthesis
conclusions directly inform the integration across the evidence streams to draw overall conclusions
for each of the assessed health effect categories (evidence demonstrates¦, evidence indicates; evidence
suggests; evidence inadequate, or strong evidence supports no effect). A structured framework
approach is used to guide both evidence synthesis and integration. While there are circumstances
where specific mechanistic evidence (typically biological precursors) is included in the unit of
analysis for human or animal evidence synthesis, in most cases mechanistic findings are presented
separately from the human and animal evidence and used to inform conclusions on (1) the
coherence, directness of outcome measures, and biological significance of findings within the
animal or human evidence streams during evidence synthesis and, (2) evidence integration
judgments on the human relevance of findings in animals, coherence across evidence streams
("cross-stream coherence"), information on susceptible populations or lifestages, understanding of
biological plausibility and MOA, and possibly other critical inferences (e.g., read-across analyses).
The structured framework also accommodates consideration of supplemental information (e.g.,
ADME, non-PECO route of exposure) that can inform evidence synthesis and integration judgments.

• Evidence synthesis: A summary of findings and judgment(s) regarding the certainty in the
evidence for hazard for each unit of analysis from the human and animal studies are made
in parallel, but separately. A unit of analysis is an outcome or group of related outcomes
within a health effect category that are considered together during evidence synthesis.

These judgments can incorporate mechanistic and other supplemental evidence when the
unit of analysis is defined as such (see Section 3). The units of analysis can also include or be
framed to focus on precursor events (e.g., biomarkers). In addition, this can include 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

9The 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.

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

8-1 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

streams, potential susceptibility, understanding of biological plausibility and MOA and other

critical inferences 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, after synthesis a certainty in the 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, these judgments are used to inform evidence integration 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 units of analysis are
synthesized, the main evidence integration judgments typically focus on the unit of analysis with
the strongest evidence synthesis judgments, although exceptions may occur.10 Health outcomes or
endpoints where the unit of analysis is considered to present slight, indeterminant or compelling
evidence of no effect can inform the evidence integration hazard judgment but would typically not
be used as the basis for deriving a toxicity value. Structured evidence profile tables are used to
summarize these analyses and foster consistency within and across assessments. Instructions for
using HAWC to create these tables are available at the HAWC project "IRIS PPRTV SEM Template
Figures and Resources" (see "Attachments," then select the "Creating Evidence Profile Tables in
HAWC").

10In some cases, it may 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 pre-defined
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 IPCS ("201211 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 (U.S. EPA. 1991)).
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 Vanadium and Compounds (Inhalation) IRIS Assessment

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 (certainty 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)

Evidence from human studies

Unit of analysis #1
Studies considered
and study
confidence

Description of
the primary
results

Evidence from animal studies

Unit of analysis #1
Studies considered
and study
confidence

Description of
the primary
results

All/Mostly medium or high
confidence studies

Consistency

Dose-response gradient

Large or concerning
magnitude of effect

Coherence*

•	All/Mostly low confidence
studies

•	Unexplained inconsistency

•	Imprecision

•	Concerns about biological
significance*

•	Indirect outcome
measures*

•	Lack of expected
coherence*

Judgment reached for
each unit of analysis*

Robust
Moderate
©QQ Slight
QQQ Indeterminate

	Compelling

evidence of no effect

All/Mostly medium or high
confidence studies

Consistency

Dose-response gradient

Large or concerning
magnitude of effect

Coherence*

•	All/Mostly low confidence
studies

•	Unexplained inconsistency

•	Imprecision

•	Concerns about biological
significance*

•	Indirect outcome
measures*

•	Lack of expected
coherence*

Judgment reached for
each unit of analysis

Robust
Moderate
©QQ Slight
QQQ Indeterminate

	Compelling

evidence of no effect

Describe overall evidence
integration judgment(s):

©ffiffi Evidence demonstrates
Evidence indicates (likely)

QQQ Evidence suggests
QQQ Evidence inadequate

	Strong evidence supports no

effect

Highlight the primary supporting
evidence for each integration
judgment*

Present inferences and conclusions
on:

Human relevance of
findings in animals*

Cross-stream coherence*

Potential susceptibility*

Biological plausibility*

Other critical inferences
(e.g., from ADME or other
supplemental
information)*

*Can 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 Vanadium and Compounds (Inhalation) IRIS Assessment

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 mav be presented
separately, e.g., bv exposure route or
key uncertainty addressed.

Each analysis mav 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)

May include separate summaries, for example by study type (e.g.,
new approach methods vs. in vivo 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)

Overall summary of expert interpretation across
the assessed set of biological events, potential
mechanisms of toxicity, or other analysis
approach (e.g., AOP).

• Includes the primary evidence supporting
the interpretation(s).

• Describes and substantiates 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.

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

8-4 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 (see Table 8-1). These factors
are adapted from considerations for causality introduced by Austin Bradford Hill (Hill. 1965) 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 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 8-3; see additional discussion in
fU.S. EPA. 2005al: fU.S. EPA. 19941: and fU.S. EPA. 2020al]. These factors are similar to the domains
considered in the GRADE Quality of Evidence framework fSchiinemann et al.. 20131. Each of the
considered factors and the certainty of evidence judgments require 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).

Biological understanding (e.g., knowledge of how an effect manifests or progresses) or
mechanistic inference (e.g., dependency on a conserved key event across outcomes) can be used to
define which related outcomes are considered as a unit of analysis. The units of analysis may also
include predefined categories of mechanistic evidence (typically precursor events). When
mechanistic evidence is included in the units of analysis, it is evaluated against all evidence
synthesis factors. Mechanistic and other supplemental evidence not included in the units of analysis
can be analyzed to inform select evidence synthesis factors (i.e., coherence, directness of outcome
measures, or biological significance) within the animal and human evidence synthesis. Additional
mechanistic evaluations (e.g., biological plausibility) as considered as part of across stream
evidence integration (see Section 8.2).

Five levels of certainty in the evidence for 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 Tables 8.4 and 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. Level of certainty analyses are conducted

for each unit of analysis within an evidence stream. Observations that increase certainty are having

an evidence base exhibiting a signal of an effect on the health outcome based on evaluation of

consistency across studies or experiments, the presence of a dose or exposure-response gradient,

observing a large or concerning magnitude of effect, and coherent findings for closely related

endpoints (can include mechanistic endpoints). These patterns are more compelling when

observed among high or medium confidence studies. Observations that decrease certainty are
This document is a draft for review purposes only and does not constitute Agency policy.

8-5 DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

1	having an evidence base of mostly low confidence studies, unexplained inconsistency, imprecision,

2	concerns about biological significance, indirect measures of outcomes, and lack of expected

3	coherence. Study sensitivity considerations can be expressed as a factor that can either increase or

4	decrease certainty in the evidence, depending on whether an association is observed. An evidence

5	base of mostly null findings where insensitivity is a serious concern decreases certainty that the

6	evidence is sufficient to support a lack of health effect or association. Conversely, there may be an

7	increase in the evidence certainty in cases where an association is observed although the expected

8	impact of study sensitivity is toward the null.

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Table 8-3. Considerations that inform judgments of the certainty of the evidence for hazard for each unit of
analysis

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
certainty. 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 where an association is identified
because the expected impact of study insensitivity is toward the null.

• An evidence base of mostly null findings where 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 (COD related to
funding source can be considered as a factor to explain apparent
inconsistency. For small evidence bases, it may be hard to assess
consistency. An evidence base of a single or a few studies where
consistency cannot be accurately assessed does not, on its own, 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 Vanadium and Compounds (Inhalation) IRIS Assessment

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 certainty (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 may be decreased if the findings are considered not likely to be
biologically significant. Effects that are small in magnitude might not be
considered to be biologically significant (adverseb) based on information
such as historical responses and variability. However, effects that appear
to be of small magnitude may be meaningful at the population level (e.g.,
IQ shifts); in such cases, certainty would not be decreased.

•	Certainty may 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 may be demonstrated
across studies or within studies, and it can be
dose- or duration-dependent. It may also not be
a monotonic dose-response (monotonicity
should not necessarily be expected as different
outcomes may be expected at low vs. high doses
or long vs. short durations 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 may increase certainty in a
relationship between exposure and outcome

•	A lack of dose-response when expected based on biological
understanding can decrease certainty in the evidence. If the data are not
adequate to evaluate a dose-response pattern, however, then 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)1, endpoint severity, judgments regarding
the potential for delayed or secondary effects, the underlying
mechanism(s) involved, as well as 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 Vanadium and Compounds (Inhalation) IRIS Assessment

Consideration

Increased evidence certainty
(of the human or animal evidence for hazard3)

Decreased evidence certainty
(of the human or animal evidence for hazard3)



(this is primarily applicable to epidemiology
studies because of their observational nature).



Directness of

outcome/endpoint

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 that have an
unclear linkage to an apical or clinical (adverseb) outcome. Scenarios
where 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.

•	Related to indirectness, certainty in the evidence may 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.

• An observed lack of expected coherent changes (e.g., in well-established
biological relationships) within or across biologically related units of
analysis typically decrease evidence certainty. This includes mechanistic
changes when included in the unit of analysis. However, as described for
decisions to increase certainty in 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
depends on the availability of evidence across multiple related endpoints
for which changes would be anticipated, and it considers factors (e.g.,

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Consideration

Increased evidence certainty
(of the human or animal evidence for hazard3)

Decreased evidence certainty
(of the human or animal evidence for hazard3)

• Coherence across findings within a unit of
analysis (e.g., consistent changes in disease
markers and biological precursors in exposed
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.

dose and duration of exposure, strength of expected relationship) across
the studies of related changes.

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.c

aWhile 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 may 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 Vanadium and Compounds (Inhalation) IRIS Assessment

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
example-based criteria that guide how to draw the certainty 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 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 a stronger certainty of evidence judgment. 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 to be 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.

Table 8-4. Framework for evidence synthesis 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 confidence but are not required. Supplemental evidence included in
the unit of analysis (e.g., mechanistic studies in exposed humans or human cells) may raise

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Evidence
synthesis
judgment

Description



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 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

(0©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 the certainty of evidence. For multiple studies, there is
primarily consistent evidence of an association with reasonable support for adversity, but
there may be some uncertainty due to potential chance, bias, or confounding or because of
the indirectness of some measures.

For a single study, 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. Supporting evidence could include associations with related endpoints,
including mechanistic evidence from exposed humans when included within the unit of
analysis.

When available and included in the unit of analysis, mechanistic data in humans that address
the above considerations may raise the certainty of evidence to Moderate for a set of studies
that otherwise would be described as Slight. In exceptional cases, biological support from
mechanistic evidence in exposed humans may support raising the certainty of evidence to
Moderate for evidence that would otherwise be described as Indeterminate.

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 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 where 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
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 may include situations where 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
(...)

A set of high confidence studies examining a reasonable spectrum of endpoints showing null
results (for example, 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

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Evidence
synthesis
judgment

Description

...in human studies

(Strong signal for
lack of an effect
with little
uncertainty)

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.

Mechanistic data in humans that address the above considerations or that provide
information supporting the lack of an association between exposure and effect with
reasonable confidence may provide additional support for this judgment.

Table 8-5. Framework for evidence synthesis 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 the certainty
of evidence: coherent effects across multiple related endpoints (within or across biologically
related units of analysis and may include mechanistic endpoints); 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. Mechanistic evidence from animals
included in the unit of analysis or used to assess coherence of findings in the animal evidence
may raise the certainty of evidence to robust for a set of studies that otherwise would be
described as moderate.

Moderate

(0©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
the certainty of 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 the
certainty of evidence, such as consistent effects across laboratories or species; coherent
effects across multiple related endpoints (may include mechanistic endpoints within the unit
of analysis); an unusual magnitude of effect, rarity, age at onset, or severity; a strong

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 Vanadium and Compounds (Inhalation) IRIS Assessment

Evidence
synthesis
judgment

Description



dose-response relationship; and/or consistent observations across exposure scenarios (e.g.,
route, timing, duration), sexes, or animal strains.

When available and included in the unit of analysis, mechanistic data in animals that address
the above considerations may raise the certainty of evidence to Moderate for a set of studies
that otherwise would be described as Slight. In exceptional cases, strong biological support
from mechanistic studies may raise the certainty of evidence to Moderate for evidence that
would otherwise be described as Indeterminate.

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 without additional information increasing the certainty of evidence
(e.g., coherent findings within the same study or from other studies).

Biological evidence from mechanistic studies may also be independently interpreted as Slight.
This category serves primarily to encourage additional study where 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 may include situations where 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.

Mechanistic data in animals that address the above considerations or that provide
information supporting the lack of an association between exposure and effect with
reasonable confidence may 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

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

8.2. EVIDENCE INTEGRATION

1	The phase of evidence integration combines animal and human evidence synthesis

2	judgments while also considering information on the human relevance of findings in animal

3	evidence, coherence across evidence streams ("cross-stream coherence"), information on

4	susceptible populations or lifestages, understanding of biological plausibility and MOA, and

5	possibly other critical inferences (e.g., read-across analyses) that generally draw on mechanistic

6	and other supplemental evidence (see Table 8-6). This analysis culminates in an evidence

7	integration judgment and narrative for each potential health effect (i.e., each noncancer health

8	effect and specific type of cancer, or broader grouping of related outcomes as defined in the

9	evaluation plan). To the extent it can be characterized prior to conducting dose-response analyses,
10	exposure context is provided.

Table 8-6. Considerations that inform evidence integration judgments

Judgment

Description

Human relevance
of findings

•	Used to describe and justify the interpretation of the relevance of the animal data to
humans. This can include consideration of mechanistic or other supplemental
information. When human evidence is lacking or has results that differ from animals,
analyses of the mechanisms underlying the animal response in relation to those
presumed to operate in humans, and the chemical's pharmacokinetics, can inform the
extent to which the animal response is likely to be relevant to humans and potentially
strengthen overall confidence in the evidence integration conclusion. Conversely,
evidence for a mechanistic pathway that is expected to only occur in animals and not in
humans can provide support for a conclusion that the animal evidence for an effect is
not relevant to humans.

•	In the absence of chemical-specific evidence informing human relevance, the evidence
integration narrative will briefly describe the interpreted comparability of experimental
animal organs/systems to humans based on underlying biological similarity (e.g.,
thyroid signaling processes are well conserved across rodents and humans). Generally, a
high-level systems summary should be possible for most encountered effects. In some
cases, however, it may be appropriate to use a statement such as, 'without evidence to
the contrary, [health effect described in the table] responses in animals are presumed
to be relevant to humans.' 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.

Cross-stream
coherence

• Addresses the concordance of findings known to be biologically related across human,
animal, and mechanistic studies, considering factors such as exposure timing and levels.
Notably, for many health effects (e.g., some nervous system and reproductive effects;
cancer), it is not necessary (or expected) that effects manifest in humans are identical
to those observed in animals, although this typically provides stronger evidence. For
example, tumors in one animal species can be predictive of carcinogenic potential in

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

8-15	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Judgment

Description



humans or other species, but not necessarily at the same site. EPA guidelines and other
resources (e.g., OECD guidelines) are consulted when drawing these inferences.

• Mechanistic support for, or biological understanding of, the relatedness between

different outcomes (and the manner in which they are manifest) in different species can
inform an understanding of coherence across evidence streams. Evidence supporting a
biologically plausible mechanistic pathway across species adds coherence (see below).

Potential
susceptibility
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 primarily on the influence of
intrinsic biological factors such as race/ethnicity, genetic variability, sex, lifestage, and
pre-existing health conditions (which can also have an extrinsic basis). Information on
extrinsic factors potentially influencing susceptibility (e.g., proximity to exposure;
certain lifestyle factors including subsistence living) are not considered in evidence
integration judgments on potential susceptibility; these exposure-focused factors are
considered by risk managers after the human health assessment is complete. Evaluation
of potential susceptibility can also include consideration of mechanistic and ADME
evidence.

Biological
plausibility or
MOA

understanding

•	Support for the biological plausibility of an association between exposure and the
health effect increases evidence certainty, particularly when observed across species.
This may be provided by data from experimental studies of mechanistic pathways,
particularly when support is provided for key events or is conserved across multiple
components of the pathway. Mechanisms or biological changes with broad scientific
acceptance for their relevance to chemical toxicity or the health effect (e.g., key
characteristics, hallmarks of cancer) may be used to organize the chemical-specific
evidence and identify key events leading from exposure to the health effect. For each
key event and key event relationship, the evidence is considered regarding the
consistency of experimental data and the generalizability, or likelihood of similarities
(e.g., in presence or function) across species, as well as the strength of the support for
the biological mechanism.

•	Mechanistic evidence from well conducted studies that demonstrates that the health
effect is unlikely to occur (i.e., species-specific effects, irrelevant exposure conditions)
can support a judgment that the effects from animal or human studies are not
biologically relevant, which weakens the summary evidence integration judgment. Such
a decision depends on an evaluation of the certainty of the information supporting vs.
opposing biological plausibility, as well as the certainty of the health effect specific
findings (e.g., stronger health effect data require more certainty in mechanistic
evidence opposing plausibility). Importantly, because understanding biological
plausibility is dependent on expert knowledge and canonical scientific knowledge, the
lack of such understanding does not provide a rationale to decrease the certainty of the
evidence for an effect (NTP, 2015); (NRC, 2014).

•	These analyses are typically conducted separately to establish MOA understanding and
referenced in the evidence integration judgment. If sufficiently supported, MOA
understanding can serve to increase (e.g., strong support for mutagenicity) or increase

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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Judgment

Description



(e.g., critical dependence on a key event not likely to be operant in humans) certainty in
the evidence integration judgments.

Other critical

inferences

(optional)

• Consideration of other evidence or nonchemical-specific information that informs
evidence integration judgments (e.g., read-across analyses, ADME understanding used
to inform other considerations; judgments on other health effects expected to be linked
to the health effect under evaluation; read-across analyses or inferences) may be
separately described as "other critical inferences."

Using a structured framework approach, one of five phrases is used to summarize the
evidence integration judgment based on the within evidence stream integration of the human and
animal evidence, and supplemental (mechanistic) evidence: evidence demonstrates, evidence
indicates, evidence suggests, evidence is inadequate, or strong evidence supports no effect (see
Table 8-7). The five integration judgment levels reflect the differences in the amount and quality of
the data that inform the evaluation of whether exposure may cause 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. More
than one descriptor 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.

8-17	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 humansc
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 concentration11].

Evidence demonstrates

• A strong evidence base demonstrating that [chemical] exposure causes [health effect] in
humans.

• This conclusion level is 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
(likely6)

• 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
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 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 js used if there is slight human evidence and indeterminate-to-slight
animal evidence.

• This conclusion level js also used with slight animal evidence and indeterminate-to-slight
human 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 Vanadium and Compounds (Inhalation) IRIS Assessment

Summary evidence integration
judgment3 in narrative

Evidence
integration
judgment level

Explanation and example scenarios'3





•	This conclusion level could also be used with moderate human evidence and sliaht 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).5

•	This conclusion level js used if there is indeterminate human and animal evidence.

•	This conclusion level js_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.

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 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 Vanadium and Compounds (Inhalation) IRIS Assessment

Summary evidence integration
judgment3 in narrative

Evidence
integration
judgment level

Explanation and example scenarios'3





•	This conclusion level js 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 js 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 comoellina 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 pathway (AOPs) 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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

For evaluations of carcinogenicity, consistent with EPA's cancer guidelines fU.S. EPA.
2005a) fU.S. EPA. 2005al. 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 guideline (U.S. EPA. 2005a).

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

8-21 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 fU.S. EPA. 2012b! EPA's Review of the
Reference Dose and Reference Concentration Processes (U.S. EPA. 2002): Guidelines for Carcinogen
Risk Assessment (U.S. EPA. 2005a). and Supplemental Guidance for Assessing Susceptibility from
EarlyLife 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 exposure11 to the chemical of interest, if supported by existing data. For noncancer hazards,
an inhalation reference concentration (RfC) and an oral reference dose (RfD) will be derived. In
addition to an RfC and RfD, this assessment will attempt to derive organ- or system-specific toxicity
values when the data are sufficiently strong (i.e., noncancer conclusions of evidence demonstrate or
evidence indicates [likely]). A reference value may also be derived for cancer effects in cases where
a nonlinear MOA is concluded that indicates a key precursor event necessaiy for carcinogenicity
does not occur below a specific exposure level (U.S. EPA. 2005a) (see Section 3.3.4). In addition,
when feasible and if the available data are appropriate for doing so, the assessment 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.

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 inhalation unit risk (IUR) facilitates estimation of human
cancer risks. Low-dose linear extrapolation is also used as a default when the data are insufficient

nDose-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 fU.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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

to establish the mode of action fU.S. EPA. 2005al An IUR is a plausible upper-bound lifetime cancer
risk from chronic inhalation of a chemical per unit of air concentration (expressed as ppm or
Hg/m3). In contrast with RfCs, an IUR 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]". EPA generally would not conduct a dose-response
assessment and derive a RfC or RfD when the noncancer outcome is not as strong (i.e., "evidence
suggests"). Cases where 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.

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

9-2 DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

1	cannot be located, semiquantitative analysis may be feasible (e.g., via NOAEL/LOAEL). In the draft

2	and final assessments, specific endpoints considered for dose-response are summarized in a tabular

3	format that includes rationales for decisions to proceed (or not) for POD derivation (see Table 9-2

4	for example format) selection.

5	In addition, mechanistic evidence that influences the dose-response analyses is highlighted,

6	for example, evidence related to susceptibility or potential shape of the dose-response curve (i.e.,

7	linear, nonlinear, or threshold model). Mode(s) of action is summarized including any interactions

8	between them relevant to understanding overall risk. For cancer dose-response, biological

9	considerations relevant to dose-response for cancer are:

10	• Is there evidence for direct mutagenicity?

11	• Does tumor latency decrease with increasing exposure?

12	• If there are multiple tumor types, which cancers have a longer latency period?

13	• Is incidence data available (incidence data are preferred to mortality data)?

14	• Were there different background incidences in different (geographic) populations?

15	• While benign and malignant tumors of the same cell of origin are generally evaluated

16	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 Vanadium and Compounds (Inhalation) IRIS Assessment

Table 9-1. Attributes used to evaluate studies for derivation of toxicity values (in addition to the health effect
category-specific evidence integration judgment)

Study attributes

Considerations

Human studies

Animal studies

Study confidence

High or medium confidence studies are highly preferred over low confidence studies. The available high and medium
confidence studies are further differentiated based on 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, relevance of specific health
outcomes to humans).

Animal studies provide supporting evidence when adequate human
studies are available and are considered principal studies 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, (U.S. EPA, 2012b), see Section 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. Attempts are made to highlight where it might
be possible to develop separate risk estimates for a specific population or life stage or determine whether evidence is available
to select a data-derived uncertainty factor (UF).

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 Vanadium and Compounds (Inhalation) IRIS Assessment

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.

Measurement of 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. Examples include: to characterize experimental variability more
realistically, 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 with greater biological significance. When
there are multiple endpoints for an organ/system, characterizing the overall impact on this organ/system is considered. For
example, if there are multiple histopathological alterations relevant to liver function changes, liver necrosis may be selected as
the most representative endpoint to consider for dose-response analysis. For cancer types, consideration is given to the overall
risk of multiple types of tumors. Multiple tumor types (if applicable) are discussed, and a rationale given for any grouping.

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 Heisey, 2001).

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

9-5 DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Table 9-2. Example table used in assessment to show endpoint consideration judgments for POD derivation.

Endpoint

Study reference/
confidence

Exposure route
duration

Human population
or strain/species

Sexes studied

POD
derivation

Rationale

Endpoint 1

Study citation and confidence
(endpoint-specific level)

e.g., Gestational
(route)

e.g., Wistar rats

males, females,
or both



e.g., Exposure-related increase

Endpoint 2

Study citation and confidence
(endpoint-specific level)

e.g., Gestational
(route)

e.g., Sprague-Dawley
rats

males, females,
or both

X

e.g., No exposure-related
effect; response not considered
biologically significant (<5%)

Endpoint 3

Study citation and confidence
(endpoint-specific level)

e.g., ongoing,
measured during
gestation

e.g., Children aged 7 yr

Both males and
females



e.g., Consistent associations
across studies, minimal
concerns for exposure
measurement

Table 9-3. Specific example of presenting endpoints considered for dose-response modeling and derivation of
points of departure.

Endpoint

Study reference/
confidence

Exposure route
and duration

Human
population or
test species and
strain

Lifestage and
sex

POD
derivation

Rationale

Endocrine Effects (hazard judgment of evidence indicates [likely])

Decreased
serum free
and total T4

(NTP, 2018); high
confidence

Gavage, 28 d

S-D rat

Adult female

Yes,

Dose-dependent effects in free and total
T4 in females and free T4 in males; large
magnitude of effect in both sexes (91%
reduction in free T4 in males at low dose
where body weight unaffected, and
36%-53% reduction in free and total T4
in females at >3.12 mg/kg-d); effects in
males were not prioritized due to
elevated weight loss at higher doses.

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

9-6	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Endpoint

Study reference/
confidence

Exposure route
and duration

Human
population or
test species and
strain

Lifestage and
sex

POD
derivation

Rationale

Endocrine Effects (hazard judgment of evidence indicates [likely])



(NTP, 2018); high
confidence

Gavage, 28 d

S-D rat

Adult male

No, X



Add a
second
endpoint,
maybe not
modeled due
to large
insensitivity
vs. T4







Adult males and
females

No, X



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

9-7	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 (U.S. EPA. 2005a): (U.S. EPA. 2012b). (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. When pharmacodynamic modeling is not available
to assess health effects associated with inhalation exposure to vanadium compounds, 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.

9-8 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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 (U.S. EPA. 2012b)J.

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. 2012b! For noncancer (and
nonlinear cancer), a 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/recordisplav.cfm?deid=308382
). 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.

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 c ses where 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).12

Version 3.2 and higher of BMD also provides an alternative modeling approach that uses
Bayesian model averaging for dichotom ous modeling average (DMA). BMDS also provide a BMA
modeling approach for dichotomous data. EPA is in the process of evaluating this approach for use
in assessments and may provide supplementary values derived from such modeling.

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 oral slope factor (OSF) or IUR, and for nonlinear extrapolation, the POD is used in
calculating an RfD or RfC.

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

12The 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.

9-9 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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. 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, 1/2 standard
deviation for more severe effects. 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). see Section 3.1.1; (U.S. EPA. 1991). see Section 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/kg3/4day 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 ((U.S. EPA. 2011a): (U.S. EPA. 2005a). see section 3.1.3). Inhalation
exposures are scaled using dosimetry models that apply species-specific physiologic and
anatomic factors and consider whether the effect occurs at the site of first contact or after
systemic circulation ((U.S. EPA. 1994): (U.S. EPA. 2012a). see Section 3).

• It can be informative to convert doses across exposure routes. If this is done, the assessment
describes the underlying data, algorithms, and assumptions ((U.S. EPA. 2005a). see Section
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 (U.S. EPA. 1988).

• The preferred approach for dosimetry extrapolation from animals to humans is through
PBPK modeling.

• Briefly, PBPK model simulations can be 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 (e.g., 6 hours/day, 5 days/week inhalation
exposure), the resulting internal metric effectively accounts for the difference between the
pattern and a nominal 24 hours/day, 7 days/week 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. The human version of
the PBPK model can then be used to estimate the exposure concentration in air which, given
continuous (24 hour/day, 7 day/week) inhalation exposure, would result in a given internal

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

9-10 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

dose POD. Any remaining uncertainty factors, including the factor of 10 for human inter-
individual variability (UFh), will then be applied for derivation of the human equivalent
concentration (HECs).

• If needed, a similar approach can be applied for oral-to-inhalation route extrapolation for
endpoints where toxicity data are available from oral dosimetry studies but not from
inhalation.

9.3.2. Extrapolation: Slope Factors and Unit Risk

An OSF or IUR 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. 2005al. 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 (U.S. EPA. 2005a).

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 ((U.S. EPA. 2005a). 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 (EGBE) assessment where the cancer RfD was based on
hemosiderin deposition in the liver vs. liver tumor incidence fU.S. EPA. 20101.

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. 20141 and Bayesian methods for characterizing
population variability fNRC. 20141 are feasible and may be distinguished from the 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
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

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

9-11 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

modeling, which may range from detailed chemical-specific to default approaches fU.S. EPA.
2014. 2011a). 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 databased adjustments for pharmacodynamics or
pharmacokinetics is considered 9,10 (U.S. EPA. 2014. 2002). When sufficient data are
available, an intraspecies UF either less than or greater than lOfold may be justified fU.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
nonsusceptible individuals; (see (U.S. EPA. 2002). Section 4.4.5; (U.S. EPA. 1998). Section
4.2; flJ.S. EPA. 19961. Section 4; fU.S. EPA. 19941. Section 4.3.9.1 :f!J.S. EPA. 19911. Section
3.4). When the use of such data or modeling is not supported, an 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 (U.S. EPA. 2002.1998.1996.1994.
19911.

• 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 fU.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 (U.S. EPA. 2002.1998.1996.1994.1991).
The size of the factor depends on the nature of the database deficiency. For example, the
EPA typically follows the recommendation that a factor of 10 be applied if both a prenatal
toxicity study and a two-generation reproduction study are missing and a factor of 101/2
(i.e., 3) if either one or the other is missing ((U.S. EPA. 2002). Section 4.4.5).

The POD for a reference value is divided by the product of these factors ((U.S. EPA. 2002).
Section 4.4.5), recommends that any composite factor that exceeds 3,000 represents excessive
uncertainty and recommends against relying on the associated reference value.

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

9-12 DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

10. PROTOCOL HISTORY

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

10-1	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

36

37

38

39

40

41

42

43

44

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

REFERENCES

Agrawal. H: Eden. R: Zhang. X: Fine. PM: Katzenstein. A: Miller. IW: Ospital. I: Teffera. S: Cocker. DR.
III. (2009). Primary particulate matter from ocean-going engines in the Southern California
Air Basin. Environ Sci Technol 43: 5398-5402. http://dx.doi.org/10.1021/es8035016.

AT SDR (Agency for Toxic Substances and Disease Registry). (2012). Toxicological profile for

vanadium [ATSDR Tox Profile], Atlanta, GA: U.S. Department of Health and Human Services,
Public Health Service, http: //www.atsdr.cdc.gov/toxprofiles/tp58.pdf.

Cohen. MP. (2007). Toxicity of Vanadium Compounds: Pulmonary and Immune System Targets. ACS
Symp Ser Am Chem Soc 974: 217-239.

Crans. DC: Amin. SS: Keramidas. AD. (1998). Chemistry of relevance to vanadium in the

environment. In JO Nriagu (Ed.), Vanadium in the environment Part 1: Chemistry and
biochemistry (pp. 73-95). New York: John Wiley & Sons.

Dickersin. K. (1990). The existence of publication bias and risk factors for its occurrence. JAMA 263:
1385-1389.

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/j.efsa.2017.4971.

Gonzalez-Villalva. A: Pinon-Zarate. G: De la Pena Diaz. A: Flores-Garcia. M: Bizarro-Nevares. P:
Rendon-Huerta. EP: Colin-Barenque. L: Fortoul. TI. (2011). The effect of vanadium on
platelet function. Environ Toxicol Pharmacol 32: 447-456.
http://dx.doi.Org/10.1016/j.etap.2011.08.010.

Goulle. TP: Mahieu. L: Castermant. I: Neveu. N: Bonneau. L: Laine. G: Bouige. D: Lacroix. C. (2005).

Metal and metalloid multi-elementary ICP-MS validation in whole blood, plasma, urine and
hair Reference values. Forensic Sci Int 153: 39-44.
http://dx.doi.Org/10.1016/j.forsciint.2005.04.020.

Gustafsson. TP. (2019). Vanadium geochemistry in the biogeosphere -speciation, solid-solution
interactions, and ecotoxicity. Appl Geochem 102: 1-25.
http://dx.doi.Org/10.1016/j.apgeochem.2018.12.027.

Gutierrez-Gonzalez. E: Garcia-Esquinas. E: de Larrea-Baz. NF: Salcedo-Bellido. I: Navas-Acien. A:
Lope. V: Gomez-Ariza. TL: Pastor. R: Pollan. M: Perez-Gomez. B. (2019). Toenails as
biomarker of exposure to essential trace metals: A review [Review], Environ Res 179:
108787. http: //dx.doi.org/10.1016/i.envres.2019.108787.

Hauser. R: Elreedv. S: Hoppin. TA: Christiani. DC. (1995). Airway obstruction in boilermakers

exposed to fuel oil ash. A prospective investigation. Am J Respir Crit Care Med 152: 1478-
1484. http: //dx.doi.Org/10.1164/airccm.152.5.7582280.

Higgins. TPT: Thomas. I: Chandler. 1: Cumpston. M: Li. T: Page. MI: 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.org/handbook.

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
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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

workbench for systematic review. Syst Rev 5: 87. http://dx.doi.org/10.1186/sl3643-Q16-
0263-z.

Huang. 1. enHow: Huang. F: Evans. L. es: Glasauer. S. (2015). Vanadium: Global (bio)geochemistry.
Chem Geol 417: 68-89. http://dx.doi.Org/10.1016/i.chemgeo.2015.09.019.

I ARC (International Agency for Research on Cancer). (2006). Vanadium pentoxide. In IARC

monographs on the evaluation of the carcinogenic risk of chemicals to humans Vol 86:
Cobalt in hard metals and cobalt sulfate gallium arsenide, indium phosphide and vanadium
pentoxide. Lyon, France.

IPCS (International Programme on Chemical Safety). (2001). Concise International Chemical

Assessment Document 29: Vanadium pentoxide and other inorganic vanadium compounds.
(CICAD 29). Geneva, Switzerland: World Health Organization.
http://www.inchem.org/documents/cicads/cicads/cicad29.htm.

IPCS (International Programme on Chemical Safety). (2012). Guidance for immunotoxicity risk

assessment for chemicals. (Harmonization Project Document No. 10). Geneva, Switzerland:
World Health Organization, https: //apps.who.int/iris/handle/10665 /330098.

Ito. K: Tohnson. S: Kheirbek. I: Cloughertv. 1: Pezeshki. G: Ross. Z: Eisl. H: Matte. TP. (2016).

Intraurban variation of fine particle elemental concentrations in New York City. Environ Sci
Technol 50: 7517-7526. http://dx.doi.org/10.1021/acs.est6b00599.

Kodros. IK: Bell. ML: Dominici. F: L'Orange. C: Godri Pollitt. KT: Weichenthal. S: Wu. X: Volckens. 1.
(2022). Unequal airborne exposure to toxic metals associated with race, ethnicity, and
segregation in the USA. Nat Commun 13: 6329. http: //dx.doi.org/10.1038/s41467-Q22-
33372-z.

Kucera. I: Sabbioni. E. (1998). Baseline vanadium levels in human blood, serum, and urine. In JO

Nriagu (Ed.), Vanadium in the environment Part 2: Health effects (pp. 75-90). New York, NY:
John Wiley & Sons.

Mamane. Y: Pirrone. N. (1998). Vanadium in the atmosphere. In JO Nriagu (Ed.), Vanadium in the
environment Part 1: Chemistry and biochemistry (pp. 37-71). New York, NO: John Wiley &
Sons.

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.

National Toxicology Program (NTP). NTP Technical Report on the Toxicity Studies of Sodium
Metavanadate (CASRN 13718-26-8) and Vanadyl Sulfate (CASRN 27774-13-6)

Administered in Drinking Water to Sprague Dawley (Hsd:Sprague Dawley® SD®) Rats and
B6C3F1/N Mice. (Toxicology Report 106). Research Triangle Park, NC.

NRC (National Research Council). (2014). Review of EPA's Integrated Risk Information System
(IRIS) process. Washington, DC: The National Academies Press.
http://dx.doi.org/10.17226/18764.

NTP (National Toxicology Program). (2002). NTP technical report on the toxicology and

carcinogenesis studies of vanadium pentoxide (CAS No. 1314-62-1) in F344/N rats and
B6C3F1 mice (inhalation studies). (NTP TR 507; NIH Publication No. 03-4441). Research
Triangle Park, NC. https: //ntp.niehs.nih.gov/publications/reports/tr/500s/tr507.

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
Program, Office of Health Assessment and Translation.
https://ntp.niehs.nih.gov/ntp/ohat/pubs/handbookjan2015 508.pdf.

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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

NTP (National Toxicology Program). (2018). 28-day evaluation of the toxicity (C04049) of

perfluorononaoic acid (PFNA) (375-95-1) on Harlan Sprague-Dawley rats exposed via
gavage [NTP], http: //dx.doi.org/10.22427/NTP-DATA-002-02655-0003-000Q-3.

NTP (National Toxicology Program). (2019). Handbook for conducting a literature-based health

assessment using OHAT approach for systematic review and evidence integration. Research
Triangle, NC: National Institute of Environmental Health Sciences.
https://ntp.niehs.nih.gov/ntp/ohat/pubs/handbookmarch2019 508.pdf.

Peltier. RE: Lippmann. M. (2010). Residual oil combustion: 2. Distributions of airborne nickel and
vanadium within New York City. J Expo Sci Environ Epidemiol 20: 342-350.
http://dx.doi.org/10.1038/jes.2009.28.

Rairiska. E: Biziuk. M: Sarbu. C: Szczepaniak. K: Frontasveva. MY: Culicov. 0: Bode. P: Astel. A.

(2005). Assessment of phosphatic fertilizer production impact on occupational staff based
on NAAof hair, nails, and inhald particles. J Environ Sci Health A Tox Hazard Subst Environ
Eng 40: 2137-2152. http://dx.doi.org/10.1080/10934520500234635.

Rehder. D. (2015). The role of vanadium in biology. Metallomics 7: 730-742.
http: / /dx.doi.or g/10.10 3 9 /c4mt0 0 3 0 4g.

Roberts. GK: Stout. MP: Savers. B: Fallacara. DM: Heitmancik. MR: Waidvanatha. S: Hooth. Ml.

(2016). 14-day toxicity studies of tetravalent and pentavalent vanadium compounds in
Harlan Sprague Dawley rats and B6C3F1/N mice via drinking water exposure. Toxicol Rep
3: 531-538. http://dx.doi.Org/10.1016/i.toxrep.2016.05.001.

Schlesinger. WH: Klein. EM: Vengosh. A. (2017). Global biogeochemical cycle of vanadium [Review],
Proc Natl Acad Sci USA 114: E11092-E11100.
http://dx.doi.org/10.1073/pnas.1715500114.

Schiinemann. H: Brozek. 1: Guvatt. G: Oxman. A. (2013). GRADE handbook. Available online at
https://gdtgradepro.org/app/handbook/handbook.html (accessed April 22, 2022).

Shafer. MM: Toner. BM: Overdier. IT: Schauer. 11: Fakra. SC: Hu. S: Herner. TP: Avala. A. (2012).

Chemical Speciation of Vanadium in Particulate Matter Emitted from Piesel Vehicles and
Urban Atmospheric Aerosols. Environ Sci Technol 46: 189-195.
http://dx.doi.org/10.1021/es200463c.

Smith. MT: Guvton. KZ: Gibbons. CF: Fritz. TM: Portier. CI: Rusvn. I: Pemarini. PM: Caldwell. TC:

Kavlock. RT: Lambert. PF: Hecht. SS: Bucher. 1. R.: Stewart. BW: Baan. RA: Cogliano. VI: Straif.
K. (2016). Key characteristics of carcinogens as a basis for organizing data on mechanisms
of carcinogenesis [Review], Environ Health Perspect 124: 713-721.
http: / /dx.doi.org/10.12 89 /ehp. 1509912.

Spada. NT: Cheng. X: White. WH: Hvslop. NP. (2018). Pecreasing Vanadium Footprint of Bunker Fuel
Emissions. Environ Sci Technol. http: //dx.doi.org/10.1021/acs.est.8b02942.

Sterne. TAC: Hernan. MA: Reeves. BC: Savovic. 1: Berkman. NP: Viswanathan. M: Henry. P: Altman.

PG: Ansari. MT: Boutron. I: Carpenter. TR: Chan. AW: Churchill. R: Peeks. IT: Hrobjartsson. A:
Kirkham. T: Tiini. P: Loke. YK: Pigott. TP: Ramsay. CR: Regidor. P: Rothstein. HR: Sandhu. L:
Santaguida. PL: Schiinemann. HT: 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 /bmj. i4919.

Sturini. M: Maraschi. F: Cucca. L: Spini. G: Talamini. G: Profumo. A. (2010). Petermination of
vanadium(V) in the particulate matter of emissions and working areas by sequential
dissolution and solid-phase extraction. Anal Bioanal Chem 397: 395-399.
http: //dx.doi.org/10.1007/s00216-009-3277-8.

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

R-3	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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Tao. L: Fairlev. D: Kleeman. MT: Harlev. RA. (2013). Effects of switching to lower sulfur marine fuel
oil on air quality in the San Francisco Bay area. Environ SciTechnol 47: 10171-10178.
http://dx.doi.org/10.1021/es401049x.

Thayer. KA: Shaffer. RM: Angrish. M: Arzuaga. X: Carlson. LM: Davis. A: Dishaw. L: Druwe. I: Gibbons.
C: Glenn. B: Tones. R: Kaiser. TP: Keshava. C: Keshava. N: Kraft. A: Lizarraga. L: Markev. K:
Persad. A: Radke. EG:... Yost. E. (2022). Use of systematic evidence maps within the US
environmental protection agency (EPA) integrated risk information system (IRIS) program:
Advancements to date and looking ahead [Comment], Environ Int 169: 107363.
http://dx.doi.org/10.1016 /i .envint.2 0 22.107363.

Tullar. IV: Suffet. IH. (1975). The fate of vanadium in an urban air shed: The lower Delaware River
Valley. J Air Pollut Control Assoc 25: 282-286.

U.S. Department of Commerce. (2021). The effect of imports of vanadium on the national security:
an investigation conducted under section 232 of the trade expansion act of 1962, as
amended.

U.S. EPA (U.S. Environmental Protection Agency). (1985). Health and environmental effects profile
for vanadium pentoxide [EPA Report], (EPA/600/X-85/114). Cincinnati, OH.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults/titleDetail/PB88182696.xhtml.

U.S. EPA (U.S. Environmental Protection Agency). (1987). Health effects assessment for vanadium
and compounds [EPA Report], (EPA/600/8-88/061).
http://nepis.epa.gov/exe/ZyPURL.cgi?Dockey=2000T901.txt.

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). (1991). Guidelines for developmental toxicity risk
assessment. Fed Reg 56: 63798-63826.

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=30004YOB.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-
neurotoxicity-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). (2005a). Guidelines for carcinogen risk
assessment [EPA Report], (EPA630P03001F). Washington, DC.
https://www.epa.gOv/sites/production/files/2 013-
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.

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

R-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

37

38

39

40

41

42

43

44

45

46

47

48

49

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

https://www.epa.gov/risk/supplemental-guidance-assessing-susceptibility-early-life-
exposure-carcinogens.

U.S. EPA (U.S. Environmental Protection Agency). (2008). Provisional Peer-Reviewed Toxicity

Values for vanadium pentoxide (CASRN 1314-62-1) [EPA Report], Cincinnati, OH.

U.S. EPA (U.S. Environmental Protection Agency). (2009). Provisional peer-reviewed toxicity values
for vanadium and its soluble inorganic compounds other than vanadium pentoxide (CASRN
7440-62-2 and others): Derivation of subchronic and chronic oral RfDs [EPA Report],
(EPA/690/R-09/070F). Cincinnati, OH.

https://cfpub.epa.gov/ncea/pprtv/documents/Vanadium.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2010). Toxicological review of ethylene glycol
monobutyl ether (EGBE) (CAS no. 111-76-2) in support of summary information on the
integrated risk information system (IRIS), march 2010. (EPA/635/R-08/006F).
https://iris.epa.gOv/static/pdfs/0500tr.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2011a). Recommended use of body 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). (2011b). Toxicological review of vanadium
Pentoxide (External review draft). (635R11004C).
http://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=236587.

U.S. EPA (U.S. Environmental Protection Agency). (2012a). Advances in inhalation gas dosimetry for
derivation of a reference concentration (RfC) and use in risk assessment (pp. 1-140).
(EPA/600/R-12/044). Washington, DC.

https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=244650&.CFID=50524762&.CFTOK
EN=17139189.

U.S. EPA (U.S. Environmental Protection Agency). (2012b). 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
[EPA Report], (EPA/100/R-14/002F). Washington, DC: Risk Assessment Forum, Office of
the Science Advisor, https://www.epa. gov/sites /production/files /2 015-
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.

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/2 017-
06/documents/tsca ra guidance final.pdf.

U.S. EPA (U.S. Environmental Protection Agency). (2018a). Chemistry Dashboard. Washington, DC.

Retrieved from https: //comptox.epa.gov/dashboard
U.S. EPA (U.S. Environmental Protection Agency). (2018b). An umbrella Quality Assurance Project
Plan (QAPP) for PBPK models [EPA Report], (ORD 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
http s: / /che mvie w. epa. gov /chemvie w

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

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

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). (2019c). Integrated Science Assessment (ISA) for
particulate matter (final report, Dec 2019) [EPA Report], (EPA/600/R-19/188).

Washington, DC. https://cfpub.epa.gov/ncea/isa/recordisplay.cfm7deid=347534.

U.S. EPA (U.S. Environmental Protection Agency). (2020a). ORD staff handbook for developing IRIS
assessments (public comment draft) [EPA Report], (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). (2021a). IRIS assessment plan for inhalation
exposure to vanadium and compounds (scoping and problem formulation materials).
(EPA/635/R-21/077).

https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=3 51650.

U.S. EPA (U.S. Environmental Protection Agency). (2021b). Systematic review protocol for the
vanadium and compounds (oral exposure) IRIS assessment (preliminary assessment
materials). (EPA/635/R-21/047).

U.S. EPA (U.S. Environmental Protection Agency). (2022). ORD staff handbook for developing IRIS
assessments [EPA Report], (EPA 600/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.

Weisskopf. MG: Seals. RM: Webster. TF. (2018). Bias amplification in epidemiologic analysis of

exposure to mixtures. Environ Health Perspect 126. http: //dx.doi.org/10.1289/EHP2450.
WHO (World Health Organization). (1988). Vanadium. Geneva, Switzerland.

http://www.inchem.org/documents/ehc/ehc/ehc81.htm.

WHO (World Health Organization). (2000). Air quality guidelines for Europe (2nd ed.). Copenhagen,
Denmark: World Health Organization, Regional Office for Europe.
http://www.euro.who.int/en/health-topics/environment-and-health/air-
qualitv/publications/pre2009/air-quality-guidelines-for-europe.

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.

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

R-6	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

APPENDIX A. ELECTRONIC DATABASE SEARCH
STRATEGIES

Table A-l. Database search strategies for vanadium and compounds

Source

Search Strategy

Number of
records

WOS

3/28/2019

3/9/2020

6/3/2021

((TS="Ammonium metavanadate" ORTS="Ammonium monovanadate" OR
TS="Ammonium trioxovanadate" ORTS="Monosodium trioxovanadate" OR
TS="Oxosulfatovanadium pentahydrate" ORTS="Sodium metavanadate" OR
TS="Sodium o-vanadate" ORTS="Sodium orthovanadate" ORTS="Sodium
pervanadate" ORTS="Sodium tetraoxovanadate" ORTS="Sodium
trioxovanadate" ORTS="Sodium vanadate" ORTS="Trisodium orthovanadate"
ORTS="Trisodium tetraoxovanadate" ORTS="Trisodium vanadate" OR
TS="Vanadic sulfate" OR TS="vanadium" OR TS="Vanadyl sulfate" OR
TS="Vanadic" OR TS="Vanadin" ORTS="sodium peroxyvanadate" ORTS="Vanadyl
sulfate pentahydrate" ORTS="Ammonium vanadate" ORTS="Divanadium
trioxide" ORTS="Sodium hexavanadate") AND PY=(2010-2019))

((TS="Sodium tetravanadate" OR TS="Sodium vanadite" OR TS="Sulfovanadic
acid" ORTS="vanadium salt" ORTS="Tetrachlorovanadium" ORTS="Trichlorooxo
vanadium" ORTS="Trichlorooxovanadium" ORTS="Trichlorooxovanadium oxide"
OR TS="Vanadic acid" OR TS="Vanadic oxide" OR TS="Vanadious" OR
TS="Vanadosulfuric acid" ORTS="Vanadyl chloride" ORTS="Vanadyl trichloride"
ORTS="Divanadium pentaoxide" ORTS="Divanadium pentoxide" OR TS="Vanadic
acid anhydride" OR TS="Vanadic anhydride" ORTS="Vanadin(V) oxide" OR
TS="Vanadium dust" ORTS="Vanadium fume" ORTS="Vanadium oxide" OR
TS="Vanadium pentaoxide" ORTS="Vanadium pentoxide") AND PY=(2010-2019))

((TS="Vanadium" AND (TS="chloride" ORTS="dichloride" ORTS="oxide" OR
TS="oxychloride" ORTS="oxytrichloride" ORTS="sesquioxide" ORTS="sulfate"
ORTS="sulphate" ORTS="tetrachloride" OR TS="trichloride" ORTS="trioxide"))
AND PY=2010-2019)

29,092

PUBMED

3/28/2019

3/9/2020

6/3/2021

(((7440-62-2[rn] OR 00J9J9XKDE[rn] OR 27774-13-6[rn] OR 6DU9Y533FA[rn] OR
13718-26-8[rn] OR 13721-39-6[rn] OR 7803-55-6[rn] OR FL85PX638G[rn] OR
12439-96-2[rn] OR "Ammonium metavanadate"[tw] OR "Ammonium
monovanadate"[tw] OR "Ammonium trioxovanadate"[tw] OR "Monosodium
trioxovanadate"[tw] OR "Oxosulfatovanadium pentahydrate"[tw] OR "Sodium
metavanadate"[tw] OR "Sodium o-vanadate"[tw] OR "Sodium
orthovanadate"[tw] OR "Sodium pervanadate"[tw] OR "Sodium
tetraoxovanadate"[tw] OR "Sodium trioxovanadate"[tw] OR "Sodium
vanadate"[tw] OR "Trisodium orthovanadate"[tw] OR "Trisodium
tetraoxovanadate"[tw] OR "Trisodium vanadate"[tw] OR "Vanadic sulfate"[tw]
OR vanadium[tw] OR "Vanadyl sulfate"[tw] OR Vanadic[tw] OR Vanadin[tw] OR
"sodium peroxyvanadate"[tw] OR "Vanadyl sulfate pentahydrate"[tw] OR 16785-

5,664

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

A-l	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Source

Search Strategy

Number of
records



81-2[rn] OR 12436-28-l[rn] OR 12058-74-l[rn] OR 64082-34-4[rn] OR 10580-52-
6[rn] OR 7718-98-l[rn] OR 1314-34-7[rn] OR 7632-51-l[rn] OR 11115-67-6[rn] OR
7727-18-6[rn] OR "Ammonium vanadate"[tw] OR "Divanadium trioxide"[tw] OR
"Sodium hexavanadate"[tw] OR "Sodium tetravanadate"[tw] OR "Sodium
vanadite"[tw] OR "Sulfovanadic acid"[tw] OR "vanadium salt"[tw] OR
Tetrachlorovanadium[tw] OR "Trichlorooxo vanadium"[tw] OR
Trichlorooxovanadium[tw] OR "Trichlorooxovanadium oxide"[tw] OR"Vanadic
acid"[tw] OR "Vanadic oxide"[tw] OR Vanadious[tw] OR "Vanadosulfuric acid"[tw]
OR "Vanadyl chloride"[tw] OR "Vanadyl trichloride"[tw] OR 1314-62-l[rn] OR
"Divanadium pentaoxide"[tw] OR "Divanadium pentoxide"[tw] OR "Vanadic acid
anhydride"[tw] OR "Vanadic anhydride"[tw] OR "Vanadin(V) oxide"[tw] OR
"Vanadium dust"[tw] OR "Vanadium fume"[tw] OR "Vanadium oxide"[tw] OR
"Vanadium pentaoxide"[tw] OR "Vanadium pentoxide"[tw]) OR (Vanadium[tw]
AND (chloride[tw] OR dichloride[tw] OR oxide[tw] OR oxychloride[tw] OR
oxytrichloride[tw] OR sesquioxide[tw] OR sulfate[tw] OR sulphate[tw] OR
tetrachloride[tw] ORtrichloride[tw] ORtrioxide[tw]))) AND ("2010"[PDAT] :
"3000"[PDAT]))



TOXLINE

3/28/2019

@SYN0+@AND+@OR+(@TERM+@rn+7440-62-2+@TERM+@rn+27774-13-

6+@TERM+@rn+13718-26-8+@TERM+@rn+13721-39-6+@TERM+@rn+7803-55-

6+@TERM+@rn+12439-96-2+@TERM+@rn+16785-81-2+@TERM+@rn+12436-

28-l+@TERM+@rn+12058-74-l+@TERM+@rn+64082-34-

4+@TERM+@ rn+10580-52-6+@TERM+@rn+7718-98-l+@TERM+@ rn+1314-34-

7+@TERM+@rn+7632-51-l+@TERM+@rn+11115-67-6+@TERM+@rn+7727-18-

6+@TERM+@rn+1314-62-

l)+@RANGE+yr+2010+2019+@NOT+@org+pubmed+pubdart+nih

@SYNO+@AND+@OR+(FL85PX638G+6DU9Y533FA+OOJ9J9XKDE+"Ammonium+m
etavanadate"+"Ammonium+monovanadate"+"Ammonium+trioxovanadate"+"M
onosodium+trioxovanadate"+"Oxosulfatovanadium+pentahydrate"+"Sodium+me
tavanadate"+"Sodium+o-

vanadate"+"Sodium+orthovanadate"+"Sodium+pervanadate"+"Sodium+tetraoxo

vanadate"+"Sodium+trioxovanadate"+"Sodium+vanadate"+"Trisodium+orthovan

adate"+"Trisodium+tetraoxovanadate"+"Trisodium+vanadate"+"Vanadic+sulfate

"+vanadium+"Vanadyl+sulfate"+Vanadic+Vanadin+"sodium+peroxyvanadate"+"V

anadyl+sulfate+pentahydrate"+"Ammonium+vanadate"+"Divanadium+trioxide"+

"Sodium+hexavanadate"+"Sodium+tetravanadate"+"Sodium+vanadite"+"Sulfova

nadic+acid"+"vanadium+salt"+"Trichlorooxo+vanadium"+Tetrachlorovanadium+T

richlorooxovanadium+"Trichlorooxovanadium+oxide"+"Vanadic+acid"+"Vanadiu

m+dust"+"Vanadium+fume"+"Vanadium+oxide"+"Vanadium+pentaoxide"+"Vana

dium+pentoxide"+"Vanadic+oxide"+Vanadious+"Vanadosulfuric+acid"+"Vanadyl

+chloride"+"Vanadyl+trichloride"+"Divanadium+pentaoxide"+"Divanadium+pent

oxide"+"Vanadic+acid+anhydride"+"Vanadic+anhydride"+"Vanadin+V+oxide")+@

RANGE+yr+2010+2019+@NOT+@org+pubmed+pubdart+nih

@SYNO+@AND+vanadium+@OR+(chloride+dichloride+oxide+oxychloride+oxytri
chloride+sesquioxide+sulfate+sulphate+tetrachloride+trichloride+trioxide)+@RA
NGE+yr+2010+2019+@NOT+@org+pubmed+pubdart+nih

15

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

A-2	DRAFT-DO NOT CITE OR QUOTE

-------
Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

Source

Search Strategy

Number of
records

ATSDR
Toxicological
Profile for
Vanadium
(2012)

References pulled from ATSDR document

363

2008 & 2009

PPRTV

Assessments

References pulled from PPRTV documents

75

2011 IRIS
External
Review Draft

References pulled from V2O5 IRIS document

49

2006IARC
Document

References pulled from IARC document

241

2019 PM
Integrated
Science
Assessment

References pulled from the Integrated Science Assessment for Particulate Matter

27

OAR

References provided by Office of Air and Radiation (OAR)

10

TOTAL

25,988 unique items were discovered using this search strategy.

30,332

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

A-3	DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

APPENDIX B. PROCESS FOR SEARCHING AND
COLLECTING EVIDENCE FROM SELECTED OTHER
RESOURCES

As noted in Section 4, reference lists from existing assessments (final or publicly available
draft) were manually screened. References were identified from: PPRTV assessment of vanadium
pentoxide fU.S. EPA. 20081. PPRTV assessment of vanadium and its soluble compounds other than
vanadium pentoxide fU.S. EPA. 20091. IRIS External Review Draft assessment of vanadium
pentoxide fU.S. EPA. 2011bl. International Agency for Research on Cancer (IARC) document on
vanadium pentoxide flARC. 20061 as well as references pertinent to vanadium from the most recent
Integrated Science Assessment for Particulate Matter fU.S. EPA. 2019cl. In addition, references
suggested by the Office of Air and Radiation (OAR) were included for screening. References were
annotated with respect to the source of the record and screened using the same methods applied to
the rest of the literature inventory.

Review of the citation reference lists is typically done manually because they are not
available in a file format (e.g., IRIS) 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
already identified from the other sources are formatted in an RIS file format, imported into
DistillerSR, annotated with respect to source, and screened as outlined in Section 4.5. For tracking
assessments or reviews, the name of the source citation and the number of records imported into
DistillerSR are noted. The reference list of any study included in the literature inventoiy is
reviewed manually to identify titles that appear relevant to the PECO criteria. These citations are
tracked in a spreadsheet, compared against the literature base to determine whether they are
unique to the project, and then added to DistillerSR to be screened at the title and abstract stage for
PECO relevance.

B.l. EPA COMPTOX CHEMICALS DASHBOARD (TOXVAL)

ToxVal is searched in the EPA CompTox Chemicals Dashboard fU.S. EPA. 2018al. and data
available from the "Hazard" tab is exported from the CompTox File Transfer Protocol site. Using
both the human health POD summary file and the Record Source file, citations are identified that
apply to human health PODs. A citation for each referenced study is generated in HERO and verified
that it is not already identified from the database search (or searches of "other sources consulted")
prior to moving forward to screening in DistillerSR. Full texts are retrieved where possible; if full

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

B-l DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

texts are not available, data from the ToxVal dashboard are entered and the citation is annotated
accordingly for Tableau and HAWC visualizations by adding "(ToxVal)" to the citation.

B.2. EUROPEAN CHEMICALS AGENCY (ECHA)

A search of the ECHA registered substances database is conducted using the CASRN. The
registration dossier associated with the CASRN is 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 are downloaded in Portable Document
Format (PDF), including all nested reports having unique URLs. In addition, the data are extracted
from each dossier page and used to populate an Excel tracking sheet. Extracted fields include data
from the general information page regarding the registration type and publication dates, and on a
typical study summary page the primary fields reported in the administrative data, data source, and
effect levels sections. Each study summary results in more than one row in the tracking sheet if
more than one data source or effect level is reported.

At this stage, each study summary is reviewed for inclusion based on PECO criteria. Study
summaries identified as without administrative data information are excluded from review, and
study summaries labeled "read-across" (if any) are screened and considered supplemental material.
When a study summary considered relevant reports data from a study or lab report, a citation for
the full study is generated in HERO and verified that it was not already identified from the database
search (or searches of "other sources consulted") prior to moving forward to screening. When
citation information is not available and a full text could not be retrieved, the generated PDF is used
as the full text for screening and extraction and the citation is annotated accordingly for Tableau
and HAWC visualizations by adding "(ECHA Summary)" to the citation.

B.3. EPA CHEMVIEW

The EPA ChemView database fU.S. EPA. 2019al using the chemical CASRN is searched. The
prepopulated CASRN match and the "Information Submitted to EPA" output option filter are
selected before generating results. If results are available, the square-shaped icon under the "Data
Submitted to EPA" column is selected, and the following records are 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)

• TSCA Section 8(d) (health and safety studies)

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

B-2 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

• TSCA Section 8(e) (substantial risk)

• FYI (voluntary documents)

All records for ecotoxicology and physical and chemical property entries are excluded.
When results are available, extractors navigate into each record until a substantial risk report link
is identified and saved as a PDF file. If the report cannot be saved, due to file corruption or broken
links, the record is excluded during full-text review as "unable to obtain record." Most substantial
risk reports contain multiple document IDs, so citations are derived by concatenating the unique
report numbers such as the (formerly) Office of Toxic Substances (OTS); TSCA Section 8(e)
submission (8EHQ Num); Document Control Number (DCN); Toxic Substances Control Act Test
Submissions (TSCATS RefID); and Chemical Information System (CIS) associated with each
document, along with the typical author organization, year, and title. Once a citation is generated,
the study moves forward to DistillerSR where it is screened according to PECO and supplemental
material criteria.

B.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 are 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.

B.5. OECD ECHEMPORTAL

The OECD eChemPortal fhttps://hpvchem icals.oecd.org/UI/Search.aspx] is searched using
the chemical CASRN. Only database entries from the following sources are included and entries
from all other databases are excluded in the search. Final assessment reports and other relevant
SIDS reports embedded in the links are captured and saved as PDF files.

• OECD HPV

• OECD SIDS IUCLID

• SIDS United Nations Environment Programme (UNEP)

B.6. ECOTOX DATABASE

EPA's ECOTOX Knowledgebase fhttps://cfpub.epa.gov/ecotox/search.cfm] is searched
using the CASRN. Results are refined to terrestrial mammalian studies by selecting the terrestrial
tab at the top of the search page and sorting the results by species group. A citation for each
referenced study is generated in HERO and verified that it is not already identified from the

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

B-3 DRAFT-DO NOT CITE OR QUOTE

-------
1

Protocol for the Vanadium and Compounds (Inhalation) IRIS Assessment

database search (or searches of "other sources consulted") search prior to moving forward to
screening in DistillerSR.

B.7. EPA COMPTOX CHEMICAL DASHBOARD VERSION TO RETRIEVE A
SUMMARY OF ANY TOXCAST OR TOX21 HIGH-THROUGHPUT
SCREENING INFORMATION

Version 3.0.9 of the CompTox Chemicals Dashboard fU.S. EPA. 2019blis accessed for
high-throughput screening (HTS) data by searching the Dashboard by CASRN. Next, the
"Bioactivity" section is selected and the availability of ToxCast/Tox21 HTS data for active and
inactive assays is examined in the "TOXCAST: Summary" tab. If active assays are reported, the
figure is copied for presentation in the systematic evidence map. This figure presents (1) a
scatterplot of scaled assay responses versus AC50 values for each active assay endpoint and (2) a
cytotoxicity limit as a vertical line. More detailed information on the results of ToxCast and Tox21
assays are available in the CompTox Chemicals Dashboard section "ToxCast/Tox21," which includes
chemical analysis data, dose-response data and model fits, and "flags" assigned by an automated
analysis, which might suggest false positivity/negativity or indicate other anomalies in the data.

This information is not summarized further for the purposes of the systematic evidence map, which
is focused on identifying the extent of available evidence.

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

B-4 DRAFT-DO NOT CITE OR QUOTE

-------