cumulative risk assessment - society of toxicology

Office of Research and Development National Center for Environmental Assessment

The views expressed in these presentations are those of the authors and do not necessarily reflect the views or policies of the U.S. EPA.

Cumulative Risk Assessment: Approaches and Case Study

A webinar sponsored by the Risk Assessment Specialty Section of the Society of Toxicology

3 PM–4:30 PM June 12, 2013

Glenn E. Rice1, Amanda M. Evans2 and Linda K. Teuschler 1

1U.S. Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment, Cincinnati, OH; 2Oak Ridge Institute for Science and Education

2 Office of Research and Development National Center for Environmental Assessment

Order of Talks 1. Grouping Chemical and Non-chemical Stressors for Cumulative

Risk Assessment: Potential Applications to Stressor Combinations Associated with Cardiovascular Disease Glenn Rice 3:10–3:30

2. Adapting Chemical Mixture Risk Assessment Methods to Assess Chemical and Non-Chemical Stressor Combinations Linda Teuschler 3:30–3:50

3. Cumulative Exposure to Neurodevelopmental Stressors in Women of Reproductive Age: 2003–2004 NHANES Amanda Evans 3:50–4:10

4. Questions 4:10–4:30

Office of Research and Development National Center for Environmental Assessment

Grouping Chemical and Non-chemical Stressors for Cumulative Risk Assessment: Potential

Applications to Stressor Combinations Associated with Cardiovascular Disease

Glenn Rice

National Center for Environmental Assessment U.S. EPA

June 12, 2013

The views expressed in this presentation are those of the author and do not necessarily reflect the views or policies of the U.S. EPA.

Presenter

Presentation Notes

�


Talk Outline 1. Grouping stressors to simplify and focus cumulative chemical mixture risk assessments

a) Forming chemical groups based on exposure data b) Forming chemical groups based on toxicity data c) Forming integrated chemical exposure and toxicity groups

2. Extending grouping approaches to non−chemical stressors

3. Preliminary application of grouping approaches to non-chemical stressors associated with cardiovascular diseases


U.S. EPA, 2007

• Goals of EPA’s “Cumulative Risk Resources Document” – Provide simplifying methods for conducting cumulative risk assessments (CRAs)

increases feasibility of conducting CRAs • Grouping chemicals by potential for co−occurrence and joint toxic action

simplifies conduct of CRAs & increases feasibility – Helps focus what could be overwhelming effort

Concepts, Methods, and Data Sources for Health Risk Assessment of Multiple Chemicals, Exposures, and Effects

Part 1: Grouping Stressors


Analytic Steps in Chemical Cumulative Risk Assessment (U.S. EPA, 2007)

Generate chemical list

Identify links between chemicals and

population

Quantify population exposures and form

exposure groups

Quantify dose-response relationships and form

toxicity groups

Integrate exposure and dose-response groups, refining

exposure and toxicity assessments

Adapted from U.S. EPA (2007)


Analytic Steps in Chemical Cumulative Risk Assessment (U.S. EPA, 2007)

Generate chemical list

Identify links between chemicals and

population

Quantify population exposures and form

exposure groups


toxicity groups

Integrate exposure and dose-response groups, refining

exposure and toxicity assessments

Adapted from U.S. EPA (2007)

Focus


Forming Chemical Exposure Groups

Chemical Groupings Based on Exposure Information

Exposure Medium

Time Same Different

Same Group 1 Group 3

Different Group 2 Group 4

• Simple approach • Classify all chemicals of concern into initial groups by

their potential to occur in the same or different media and at the same or different time

Adapted from U.S. EPA, 2007


Forming Chemical Exposure Groups

Chemical Groupings Based on Exposure Information

Exposure Medium

Time Same Different

Same Group 1

PCBs and MeHg in local fish

Group 3


Same Medium/Same Time • Example: PCBs & Methyl mercury (MeHg) co-occur in local fish,

consumed by population. • Likely co-exposures to chemicals occurring in the same medium at

same time; likely place into same chemical exposure group. • Importance of knowing populations’ consumption habits (e.g., parts of

fish eaten); MeHg accumulates in muscle tissue, but PCBs in fat.


Refining Chemical Exposure Groupings: Different Media/Same or Different Time

• To form exposure groups for other situations must examine:

– Timing of pollutant occurrence in the different contaminated media

– Timing of associated exposure events (including the time between temporally separated exposures), and

– Frequency and duration of encountering contaminated exposure media


Refining Chemical Exposure Groupings: Different Media/Same or Different Time II

• Ideally, consider internal doses when forming groups: – Half-life of chemicals inside the body – Some highly persistent pollutants may always be present

• Ideally, consider whether tissue changes persist after chemical exposure ends

– Induction of metabolism – Altered tissue sensitivity (e.g., liver induction persists after chemical

eliminated)

• Does order in which the exposures occur affect outcome?

– E.g., tumor initiators and promoters


Timing of Exposure: Tissue Concentrations and Effect

Exposure Chemical A

Persistence of Compound A in Body

Persistence of Biological Effect

Time

Inte

rnal

Dos

e

Seve

rity

of T

oxic

ity

Persistence of Compound B in Body

Exposure Chemical B


Pharmacokinetic Interaction: Two Chemicals

This illustrates only one of many possibilities

Hypothetical pharmacokinetic interaction: compound B increases the persistence of both compound A and biologic effect

Exposure Chemical A


Persistence of Biological Effect

Time

Inte

rnal

Dos

e


Seve

rity

of T

oxic

ity

Exposure Chemical B


Additivity: Two Chemicals

Time

Inte

rnal

Dos

e

Hypothetical additivity: toxicity of compounds A + B increases persistence and severity of toxicity

Seve

rity

of T

oxic

ity

This illustrates only one of many possibilities


Persistence and Severity of Biological Effect


Exposure Chemical A

Exposure Chemical B


Forming Chemical Exposure Groups: Extending to Chemical Buffers (protective)

Exposure Medium

Time Same Different

Same Group 1

MeHg and omega-3 fatty

acid in local fish

Group 3


• The exposure grouping concept can be extended to chemical buffers.

• For example, fish consumption also source of fish fatty acids (e.g., omega-3 fatty acids) that might enhance cognitive development.

Fish fatty acid intake may confound MeHg-IQ relationship, biasing downward observed regression coefficient estimates from epi studies

MeHg Fetal IQ

Fish Fatty Acids +

_


Forming Toxicity Groups • For each individual chemical collect all relevant toxicological data:

– Identify primary effect (adverse effect observed at lowest dose [LOAEL]) and secondary/tertiary effects (effects above LOAEL)

– Form groups initially by tissue or organ systems affected • Refine groups based on the following:

– Examine mode of action information – Identify effects (e.g., continuous responses) not adverse alone, could

lead to adverse effect in combination with other factors (e.g., other chemical exposures, nutritional status)

– Collect pharmacokinetic information including metabolic pathways • Example: Group: MeHg, Pb, and PCBs as neurodevelopmental

toxicants and omega-3 fatty acid (FA) as a possible chemical buffer − MeHg, Pb, omega-3 FA, and PCBs exposures may result from contact

with different sources − Could refine group further with additional PK or PD information


Integrating Exposure and Toxicity Groups

• Exposure Group Results: • MeHg • PCBs • Omega-3 fatty acids

• Toxicity Group Results: • MeHg • PCBs • Pb • Omega-3 fatty acid

• Integrated Group: • MeHg • PCBs • Omega-3 fatty acids • No Pb because not in exposure group in this population


Part 2. Extending Grouping Approaches to Non-chemical Stressors

• EPA’s Cume Risk Resources Doc does not address this directly

• Feasible • Some non-chemical stressors considered in existing EPA

methods, but not necessarily called “Cume Risk” – EPA’s RfD methodology: UFH for susceptible human populations – EPA’s exposure assessment methods address vulnerable

populations that are differentially exposed


Extending Grouping Approaches to Non-chemical Stressors: Preliminary Considerations for Identifying Non-chemical Stressors that Could be Grouped with Chemical Stressors

1. Ongoing or previous Health Conditions and Disease States causing effects similar to those associated with a chemical’s effect

• Pharmacokinetic or pharmacodynamic changes due to ongoing or previous health conditions or disease states that change or exacerbate a chemical’s effect or body’s response (includes buffering)

2. Lifestages that may contribute to or buffer against a chemical’s effect 3. Genetic Factors that may contribute to or buffer against a chemical’s effect 4. Lifestyle Factors, Occupational Factors, Physical Stressors, and Biological

stressors that may contribute to or buffer against a chemical’s effect

– My intent: “further the dialogue” – I make no claim that these considerations constitute a complete set – Overlaps in these categories – Likely additional categories/Alternative Approaches to Categorizing – Different approaches could be used to organize data

Presenter

Presentation Notes


Ambient PM

• Illustrate non-chemical stressor groupings using examples from ambient particulate matter (PM) literature (where possible) and cardiovascular disease literature

• Several large U.S. cohort studies have associated elevated short-term and long-term PM10 and PM2.5 exposures with cardiovascular morbidities and mortality

• It is not known whether Ambient PM exposures cause cardiovascular morbidities and mortality or accelerate these outcomes that are caused by other stressors

Part 3. Example


Adapted from U.S. EPA, 2009

Ambient PM Pulmonary Oxidative Stress

and Inflammation Pulmonary Reflexes

Autonomic Nervous System

Atherosclerosis

Altered Vasoreactivity of

Coronary Vessels

Endothelial Cell

Activation/ Dysfunction

Systemic Inflammation/

Oxidative Stress

Liver Acute Phase

Response

Altered Sympathetic/

Parasympathetic Tone

Thrombosis

Pro-coagulation Effects

Plaque Destabilization or Rupture Altered Conduction/

Repolarization

Myocardial Ischemia

Arrhythmia Myocardial Infarction

Potential Pathways: PM & Cardiovascular Diseases

Systemic Inflammation Pathway

Autonomic Nervous System Pathway


Ongoing or Previous Health Conditions and Disease States Causing Effects Similar to Those Associated with a Chemical’s Effect

• Diabetes―EPA (2009) judged collective evidence: susceptibility factor

– Basis included positive epi studies quantifying increased risks of the following outcomes among diabetics exposed to PM relative to non-diabetic populations:

• Cardiovascular disease-related emergency room visits • Hospitalization for cardiac diseases • All cause mortality • Sources: (Zanobetti & Schwartz, 2002; Peel et al. 2007; Zeka et al., 2006;

Goldberg et al., 2006) – Some epi studies did not observe effect modification in diabetic

populations following PM exposures (Pope et al., 2006; Wellenius et al., 2006; Zanobetti and Schwartz, 2005)

• Additional basis: pathophysiologic evidence of inflamatory outcomes associated with diabetes; also observed following PM exposure


Lifestages That May Contribute to or Buffer Against a Chemical’s Effect

• Lifestages - distinguishable time frames in individuals’ lives characterized by unique and relatively stable behavioral or physiological characteristics, associated with development, growth, and aging (e.g., fetus, birth to <1 month…elderly)

• EPA (2009) concludes: older adults = a susceptible population, possibly due to higher prevalence of pre-existing cardiovascular diseases and gradual decline in physiological processes compared to younger lifestages –Overlaps exist between potentially susceptible older adults and

populations with pre-existing diseases (Kan et al., 2008) –Barnett (2006) and Host (2007) quantified increased risk of

cardiovascular disease-related hospital admissions among >65 yr compared to <65 yr after short-term PM exposures

• Youth/middle-aged lifestages could be a buffer


Potential Revision Analytic Steps in Chemical Cumulative Risk Assessment

Quantify population exposures and Form

exposure groups


toxicity groups

Identify relevant non-chemical stressors, population exposures,

changes in dose-response relationships

Suggested additional steps in Cume Risk Methods: 1. After forming chemical groups, identify stressors that plausibly increase or decrease

specific disease risks and include these in the appropriate chemical groups 2. Seek quantitative data to support quantitative estimate of risk associated with stressors

-Epidemiology or toxicology data (encourage publishing such data! Ranges are important)

3. If data not available to quantify change in risk, qualitative analyses of multiple related stressors still useful for cumulative risk assessment and risk managers

-Provide opportunities to highlight important uncertainties (stressors that co-occur); needed research


• Linda Teuschler • Michael Wright • Rick Hertzberg • Amanda Evans • Jane Ellen Simmons • Jason Lambert • Gino Scarano

Acknowledgments


Adapting Chemical Mixture Risk Assessment Methods to Assess Chemical and

Non-Chemical Stressor Combinations

Linda K. Teuschler

U.S. Environmental Protection Agency Office of Research and Development (ORD)

National Center for Environmental Assessment - Cincinnati, Ohio

Society of Toxicology

Risk Assessment Specialty Section Webinar, June 12, 2013

The views expressed in this presentation do not necessarily reflect the views or policies of the U.S. EPA.


Mixtures Risk Assessment (MRA) Methods Proposed for Extension to Cumulative Risk Assessment (CRA)

• MRA methods generally use simple models to estimate risks/hazards for complex exposures

• Component based approaches are suggested for application to CRA for chemical and nonchemical stressor combinations

– Response Addition

– Effects Addition

– Hazard Index (HI)

– Weight of Evidence (WOE*) for Toxicological Interactions

• Tiering is important concept to optimize use of resources

*Mumtaz & Durkin. 1992.


Exposure

Toxicity

Event – Mechanism of Action Detailed understanding at biochemical and molecular level Key Event – Mode of Action Identification of key and required steps Outcome – Observable adverse effect

Toxicity

Graphic used with permission of Jason Lambert

General Schematic of Toxic Events


Response Addition Assumes Statistical and Toxicological Independence of Toxic Action Rm = f1(D1) + f2(D2) + f3(D3) = R1 + R2 + R3 , where the Ri are probabilities of adverse effects

Response Addition & Effects Addition

Effects Addition Assumes Toxicological Independence of Toxic Action Em = f1(D1) + f2(D2) + f3(D3) = E1 + E2 + E3 , where the Ei are biological measurements of an adverse effect - For a common health outcome, assumes the toxicity caused by the first chemical has no impact on the toxicity caused by the second chemical (and so on). - Use at levels/exposures where joint toxic action is not expected to occur. - For effects addition, constrained by the biological limit of the effect measurement.

Rm = mixture probabilistic risk; Em = total effect measurement for the mixture; Di = dose of ith stressor; fi = dose response function of the ith stressor


Exposure to Stressors 4,5,6

Toxicity

Stressor 5 Toxic Action Stressor 6 Toxic Action

Response Addition and Effects Addition via Independent Toxic Action

Event – Mechanism of Action Key Event – Mode of Action Outcome – Adverse Effect Toxicity

Assumes same type of adverse effect is caused via independent toxic modes of action

Stressor 4 Toxic Action


r2 r1 r1*r2

r1 = 0.01, risk of heart attack for chemical 1 r2 = 0.02, risk of heart attack for nonchemical stressor 2 then r1*r2 = 0.0002, and we get, Rm = 0.01 + 0.02 – 0.0002 = 0.0298 or ~ 0.03 For small risks, the risk intersection (not an interaction term) has virtually no impact.

Heart Attack Probability Space

Response Addition- Statistical Law of Independent Events


e2

e1

e1 = incremental increase in DP (4 mmHg) caused by chemical 1 e2 = incremental increase in DP (3 mmHg) caused by nonchemical stressor 2 So, E* (mmHg) > baseline DP + e1 + e2 (constrained by biological upper limit) e.g., DP in presence of chemical 1 and nonchemical stressor 2 = 88 + 4 + 3 = 95 (mmHg)

Diastolic Blood Pressure (DP) Space, Upper Limit of 110 mmHg = E*

Effects Addition – Biological Upper Limit to Consider

baseline (88 mmHg)


• Given independence of toxicological action

– Response Addition: For some stressors, it may be reasonable to sum probabilistic risks of the same effect across diverse stressors – subtract off risk intersections

– Effects addition: May be possible to sum biological effects – need estimate of “normal levels” and a way to account for biological limits of the effect

• Assumption of independence of toxic action may be difficult to show; no established criteria for doing this

• Need to evaluate data on potential toxicological interactions that may preclude additivity assumption

Extension of Response & Effects Addition Methods to CRA


Based on dose addition, assuming toxicological similarity Interpreted as an indication of potential risk when HI > 1

U.S. EPA, 2000

MRA Hazard Index (HI)

• Scaling factor = ( 1 / RfVi ) for each stressor i • n = total number of stressors causing the same effect • RfVi is a Reference Value, representing an allowable level/intake of

stressor i • Ei is estimated Intake or exposure (in same units as the RfVi) • Use at levels/exposures where interaction effects are unlikely

∑=

=n

i i

iRfV

EHI1


Toxicity



Dose Addition via Common Mode of Action


Same toxic effect via a “Shared set of Key Events”



Toxicity



Dose Addition via Common Mode of Action


Same toxic effect via a “Shared set of Key Events”


Not Likely Scenario for

Diverse Stressors


Toxicity



More Likely: Dose Addition via Toxicological Similarity


Common target organ, tissue or system affected; may use Common adverse outcomes


Toxicity Toxicity

???

???


Example CRA HI

• RfVPM = allowable intake of particulate matter (PM) • EPM = exposure intake of PM • RfVPb = allowable intake of Lead (Pb) • EPb = exposure intake of Pb • RfVO3 = minimum recommended intake of Omega 3 Fatty Acids (O3) • EO3 = exposure intake of O3, limited by an effective dose upper bound • RfVAL = allowable number (3) of Allostatic Load (AL) markers of psychological

stress out of 10 cardiovascular, metabolic, immune system markers [AL=0, low stress; AL=1,2; intermediate stress; AL= 3-10; high stress] • EAL = number of AL markers in exposed population

RfVE

RfVE

RfVE

RfVE

AL

AL

O

O

Pb

Pb

PM

PM +−+=3

3nsion)HI(Hyperte

Interpreted as an indication of potential risk when HI > 1 Is “dose” addition still a realistic/valid assumption?


• Given similarity of toxicological action

– HI: For some stressors, it may be reasonable to sum hazard quotients for the same effect across diverse stressors

• Similar mode of action may not be easy to show; example criteria for the dioxin-like compounds and for some pesticide classes

• Similarity of toxicological action can be a very conservative assumption

• Population vulnerabilities can be accommodated – Intakes can be estimated for differential exposures – RfVs and PODs can be estimated for susceptible populations

• Need to evaluate data on potential toxicological interactions that may preclude additivity assumption

Extension of the HI Method to CRA Chemical and Nonchemical Stressors


Weight of Evidence (WOE*) Method for Considering Toxicological Interactions of Chemicals

• Systematic, consistent method for expressing the WOE* for toxicological interactions

• Reflects what we know about mechanisms of action and how they can or may apply to interactions

• Applicable using available data, typically found on binary combinations of chemicals

• Can be used to express uncertainty of, or qualitatively alter a mixtures risk assessment (e.g., modify the interpretation of a HI)

• Subject to evaluation with experimental data

Slides used with permission of Moiz Mumtaz. *Citation: Mumtaz, MM; Durkin, PR. (1992) A weight-of-evidence scheme for assessing interactions in chemical mixtures. Toxicol Ind Health 8:377-406.


<IIA Strength of

Mechanistic Data

Direction of Interaction

Strength of Data on Toxicological Significance

Slide shows only 3 of the 6 criteria used to evaluate interactions data on binary combinations of chemicals Mumtaz, M.M., De Rosa, C., and Durkin, P.R. (1994) "Approaches and Challenges in Risk Assessments of Chemical Mixtures", in Toxicology of Chemical Mixtures, Raymond S.H. Yang, ed.,Academic Press, pp. 565-597. *Mumtaz & Durkin. 1992.

WOE* Criteria of Importance & Nomenclature


Type of Joint Toxic Action - Chemicals

= Evidence of Additivity < Evidence of Toxicological Interactions Less than

Additive (e.g., antagonism) > Evidence of Toxicological Interactions Greater than

Additive (e.g, synergism) ? Inconclusive Evidence or No Data

<IIA


Toxicological Significance – Chemicals

A. Directly demonstrated • Change in toxicity observed as a result of the

interaction. B. Inferred from related compounds.

• Interaction observed in related compounds. C. Unclear

• No reliable data.

<IIA


ON TOXICITY OF

Pb Zn

Pb =IIB hematologic

Zn

<IA hematologic

EFFE

CT

OF

WOE* Evaluation - Chemicals: Considers Sequence of Exposure


<IA Effect of Zinc on Lead

< less than additive action I in vivo and in vitro

- Lead inhibits ALAD, a zinc-containing enzyme in the heme synthesis pathway - Zinc protects against the inactivation. - Zinc induces metallothionein - Zinc protects against lead absorption in GI tract

A oral zinc supplement protective of lead-induced hematopoietic effects in children


ON CARDIOVASCULAR TOXICITY OF

Ambient PM Methyl Mercury

Psychological Stress

+ I + I

Diabetes + A ER Visits

?

Elderly (> 65) + A Hosp Admissions

+ A Coronary Heart

Disease Omega 3 Oils - I - B

Coronary Heart Disease

EFFE

CT

OF

Example Evaluation of Nonchemical Stressors Effects on Chemical Toxicity

Nomenclature: + increases, - decreases Evidence: A strong, B weak, I Inferred, ? none


• Evaluation of the effect of nonchemical stressors/buffers on

chemical toxicity – Not a sequential evaluation – Need to articulate criteria/nomenclature for important

nonchemical stressors/buffers – Add epidemiological study input for associations, effect

modification – Add strength of collective data; inferred associations

– Categories of nomenclature may differ for various stressors/buffers

• Use to qualitatively modify a cumulative risk assessment – For e.g., say whether a HI combining both chemical and

nonchemical stressors would under or overestimate hazard

Extension of WOE* Methodology to Cumulative Risk Assessment



• Professional judgment is an important element

– Ensure results biologically defensible, present transparently – Confirm assumptions whenever possible

• Uncertainty/sensitivity analyses are important – Discuss data gaps, data quality differences among stressors – Describe exposure range(s) for which assessment is valid – Analyze influence of variables quantitatively

• Data needs from CRA researchers – Test environmentally-relevant doses and stressor proportions – Ensure sufficient statistical power to detect effects – Publish raw data if possible, otherwise include variance

estimates/standard errors/confidence intervals – Characterize chemical and nonchemical exposure ranges

Uncertainties in Chemical MRA may Apply to CRA for Combined Stressors

Cumulative Exposure to Neurodevelopmental Stressors in

Women of Reproductive Age: 2003–2004 NHANES

Amanda M. Evans*, Glenn E. Rice, Linda K. Teuschler, J. Michael Wright

*Oak Ridge Institute for Science and Education U.S. Environmental Protection Agency,

National Center for Environmental Assessment, Cincinnati, OH

The views expressed in this presentation are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency.

Presenter

Presentation Notes

A key step in risk assessment is the exposure assessment. In cumulative risk assessment this can involve not only chemical mixtures but also non-chemical stressors. Here, I will be describing a cumulative exposure assessment case study for neurodevelopmental toxicity using data from CDC’s national health and nutrition examination surveys. We found that a non-chemical stressor modified the association between race/ethnicity and exposure to neurodevelopmental toxicity hazards.

Why perform a Cumulative Risk Assessment: Differential Exposures and Effects

Multi-chemical body burdens (Woodruff et al., 2011)

06/12/2013 SOT RASS CRA Webinar 49

Differential vulnerability to chemical stressors by non-chemical stressors: o Socioeconomic status (Nelson et al., 2012)

o Gender (Cory-Slechta et al., 2004)

o Psychological stress (Clougherty et al., 2007)

o Lifestage (Host et al., 2008)

Presenter

Presentation Notes

There is a growing interest in CRAs because it is well recognized that we are all exposed to multiple chemical, biological, physical, and social stressors each day Recent analyses have found evidence of exposure to multiple chemical exposures. Although there is evidence of exposure to chemical mixtures, levels may not be high enough to cause harm, and/or chemicals may not have additive or joint effects—importance of CRA. Differential susceptibility to chemical exposures has been found across multiple non-chemical factors including: socioeconomic status, gender, psychological stress, and lifestage Recent analyses of 2003-04 National Health and Nutrition Examination Surveys (NHANES) data by Tracey Woodruff (University of California–San Francisco) found that regnant women in the U.S. are exposed to multiple chemicals. Although evidence of chemical mixtures were found, levels may not be high enough to cause harm, and/or chemicals may not have additive or joint effects—importance of CRA. http://ehp03.niehs.nih.gov/article/info%3Adoi%2F10.1289%2Fehp.1002727 Madeline Scammel and colleagues (Boston University) examined 2003-06 NHANES data and found that exposure patterns were different for various SES variables and actually found opposite patterns for the chemicals examined (BPA and PFCs). http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3312862/pdf/1476-069X-11-10.pdf Jane Cloughtery’s (University of Pittsburgh) research has revealed an association between traffic-related air pollution and asthma only among urban children exposed to violence. http://ehp03.niehs.nih.gov/article/fetchArticle.action?articleURI=info%3Adoi%2F10.1289%2Fehp.9863 Toxicological evidence from Deborah Cory-Slechta and colleagues has found elevated in stress hormones (corticosterone) by combined Pb and stress (maternal restraint) exposure. Interestingly, elevated stress levels were present in females only with both Pb and stress but for males elevates levels were found after Pb exposure alone. Cory-Slechta DA, Virgolini MB, Thiruchelvam M, et al. 2004. Maternal stress modulates the effects of developmental lead exposure. Environmental health perspectives 112(6):717 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1241967/pdf/ehp0112-000717.pdf Host et al (2007) quantified increased risk of cardiovascular disease-related hospital admissions among >65 yr of age compared to <65 yr of age following short-term particulate matter exposures. Host S, Larrieu S, Pascal L, et al. 2008. Short-term associations between fine and coarse particles and hospital admissions for cardio-respiratory diseases in six French cities. Occupational and environmental medicine 65(8):544-551

Cumulative Exposure Case Study: Neurodevelopmental Toxicity (NDT)

Fetal neurodevelopment is a critical window of vulnerability to many stressors


Neurodevelopmental stressors o Chemical

– Lead (Pb) (Lanphear et al., 2005)

– Methyl mercury (MeHg) (Grandjean et al., 2012)

o Non-chemical – Maternal stress (Bergman et al., 2010)

Maternal stress modifies lead-induced NDT (Cory-Slechta et al., 2004)

Presenter

Presentation Notes

One lifestage that has been identified as particularly vulnerable is fetal neurodevelopment. A main health endpoint of interest during this period neurodevelopmental toxicity, including cognitive deficits, visual or motor impairments, and neurobehavioral changes. Multiple chemicals have been identified as neurotoxic, including the metals lead and methylmercury. Although less recognized and understood, non-chemicals have also been associated with neurodevelopmental toxicity, specifically maternal or prenatal stress. Not only have studies identified these as independent risk factors for neurodevelopmental toxicity, but toxicological studies have found that maternal stress can modulate the effect of developmental lead exposure. Animal studies have observed interactions of Pb and stress that differed in relation to outcome measure and sex Prenatal stress is hypothesized to be associated with adverse fetal neurodevelopment through fetal programming pathways including dysregulation of the hypothalamic pituitary adrenal (HPA) axis-stress response regulation (fight or flight) ( Lead likely operates through numerous neurotoxic pathways including -dopaminergic dysfunction inducing alteration of dopamine release and dopamine receptor density Cadmium (Cd) (Kippler at al., 2012) Polychlorinated biphenyls (PCBs) (Park et al., 2010) Chlorpyrifos (Saunders et al., 2012) Baccarelli A, Giacomini SM, Corbetta C, et al. 2008. Neonatal thyroid function in Seveso 25 years after maternal exposure to dioxin. PLoS medicine 5(7):e161 Bergman K, Sarkar P, Glover V, et al. 2010. Maternal prenatal cortisol and infant cognitive development: Moderation by infant–mother attachment. Biological Psychiatry 67(11):1026-1032 Gedeon Y, Ramesh GT, Wellman PJ, et al. 2001. Changes in mesocorticolimbic dopamine and D< sub> 1</sub>/D< sub> 2</sub> receptor levels after low level lead exposure: a time course study. Toxicology letters 123(2):217-226 Kippler M, Tofail F, Hamadani J, et al. 2012. Early-Life Cadmium Exposure and Child Development in 5-Year-Old Girls and Boys: a Cohort Study in Rural Bangladesh. Environmental health perspectives Lanphear BP, Hornung R, Khoury J, et al. 2005. Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis. Environmental health perspectives 113(7):894 Lidsky TI, Schneider JS. 2003. Lead neurotoxicity in children: basic mechanisms and clinical correlates. Brain 126(1):5-19 Park HY, Hertz-Picciotto I, Sovcikova E, et al. 2010. Neurodevelopmental toxicity of prenatal polychlorinated biphenyls(PCBs) by chemical structure and activity: a birth cohort study. Environmental Health 9(51):1-13 Saunders M, Magnanti B, Carreira SC, et al. 2012. Chlorpyrifos and neurodevelopmental effects: a literature review and expert elicitation on research and policy. Environmental Health 11(Suppl 1):S5

Study Aims

1. Characterize cumulative exposure to chronic stress and neurodevelopmental toxicants, including Pb and MeHg

2. Identify potential maternal populations that may be at increased risk of NDT hazard


Presenter

Presentation Notes

Here, the study aims are to characterize the joint exposure to lead, methylmercury and chronic stress and to identify potential maternal populations that may have a higher likelihood of exposure to neurodevelopmental toxicity hazards.

Quantifying Chronic Stress: Allostatic Load (AL)

Allostasis is defined as “maintaining stability through change” (Sterling & Eyer, 1988)

Chronic stress may result in physiological dysregulation (McEwen & Wingfield, 2003 )

• Physiological dysregulation taxes the body and has been measured using the concept of allostatic load (AL)

AL has been operationalized as the sum of “elevated” physiological parameters Elevated physiological parameters are secondary

mediators to elevated cortisol in response to stress


Presenter

Presentation Notes

Although chemical exposures are readily quantified both in external and internal environments. Quantifying non-chemical stressors, such as chronic stress, is a little more difficult. We have chosen to use an index of cumulative physiological risk, known as allostatic load. Allostatic load is based on the concept of Allostasis, which has been defined as “stability through change”. That is, there are physiological alterations in response to either perceived or real environmental stress. While this works for acute and transient exposures to stress, Chronic stress may result in physiological dysregulation. Physiological dysregulation taxes the body and has been measured using the concept of allostatic load (AL) AL has been operationalized as the sum of physiological parameters that are considered higher risk Elevated physiological parameters are secondary mediators to elevated cortisol in response to stress Sterling, P. and Eyer, J., 1988, Allostasis: A new paradigm to explain arousal pathology. In: S. Fisher and J. Reason (Eds.), Handbook of Life Stress, Cognition and Health. John Wiley & Sons, New York. McEwen BS, Wingfield JC. 2003. The concept of allostasis in biology and biomedicine. Hormones and behavior 43(1):2-15 The allostatic load model highlights the health risk associated with even small degrees of biological dysregulation when cumulated across systems. The theory posits that such cumulated risk creates a common pathway to the onset and progression of disparate diseases. This way of framing risk runs counter to the conventional focus on disease-specific pathways and the study of singular risk factors analyzed independently. (Adler and Stewart, 2010) http://onlinelibrary.wiley.com/doi/10.1111/j.1749-6632.2009.05385.x/pdf Adler NE, Stewart J. 2010. Preface to the biology of disadvantage: socioeconomic status and health. Annals of the New York Academy of Sciences 1186(1):1-4

Methods: Dataset


2003−2004 National Health and Nutrition Examination Surveys (NHANES) (all women, n = 5,152)

Inclusion Criteria: Reproductive age (15 to 44 years) (n =3,331 excluded)

Completed both questionnaire and physical examination (n = 64 excluded)

Not pregnant (n = 347 excluded)

Self-identified as Non-Hispanic White or Black, or Mexican American (n = 100 excluded)

Measurements for all biomarkers/biometrics (n = 1,310 analytical sample)

• Neurodevelopmental Toxicants • Blood Pb (µg/dL) and blood MeHg (µg/L)

• 10 biomarkers/biometrics for AL

Presenter

Presentation Notes

110 women self-identified as other Hispanic or multi-racial Blood mercury (total and inorganic) MeHg = total mercury Less than 20% of inorganic mercury exposures were above the detection limit; therefore, total mercury was assumed to reasonably approximate MeHg exposures

Methods: Allostatic Load AL is an indicator of chronic

stress exposure AL biomarkers were dichotomized

as high or low based on clinical criteria (Bird et al., 2010)

AL score = Sum of high-risk biomarkers

• AL scores ≥3 ≈ high chronic stress exposure (Juster et al., 2010)

• AL scores 1-2 ≈ intermediate chronic stress exposure

• AL score = 0 ≈ low chronic stress exposure


Allostatic Load Biomarkers by System Biomarker Cut-point

Cardiovascular Heart Rate (beats/min) >100 Mean Systolic BP (mm Hg) >130 Mean Diastolic BP (mm Hg) >85 Homocysteine (umol/L) >15

Metabolic Body Mass Index (kg/m2) >25 HDL-Cholesterol (mg/dL) <50 Total Cholesterol (mg/dL) >200 Glycohemoglobin (%) >6.5

Immune C-reactive protein (mg/dL) >1.0 Albumin (g/dL) <3.4

Presenter

Presentation Notes

Juster RP, McEwen BS, Lupien SJ. 2010. Allostatic load biomarkers of chronic stress and impact on health and cognition. Neuroscience and biobehavioral reviews 35(1):2 Bird CE, Seeman T, Escarce JJ, Basurto-Dávila R, Finch BK, Dubowitz T, et al. 2010. Neighbourhood socioeconomic status and biological ‘wear and tear’in a nationally representative sample of us adults. Journal of Epidemiology and Community Health 64:860-865.

Methods: Hazard Measures


The Hazard Index (HI) is used here as an indicator of NDT

1. Calculate Hazard quotients (HQs) Individual blood concentration (E) divided by metal-

specific health reference value (HRV) • HRVPb = 1.76 µg/dL (Jedrychowski et al., 2009) • HRVMeHg = 5.8 µg/L (USEPA, 2001)

HQ = E/HRV

2. Calculate the HI: Sum HQs* HI-NDT = HQPb + HQMeHg

HI-NDT >1 ≈ higher NDT hazard HI-NDT ≤1 ≈ lower NDT hazard

*Dose-addition is assumed for the HI calculation because Pb and MeHg have similar neurodevelopmental endpoints

Presenter

Presentation Notes

Jedrychowski W, Perera F, Jankowski J, et al. 2009. Gender specific differences in neurodevelopmental effects of prenatal exposure to very low-lead levels: The prospective cohort study in three-year olds. Early Human Development 85(8):503-510 Health reference value for methyl mercury is 5.8 micrograms per liter; with is the blood mercury level equivalent to US EPAs current reference dose (RfD) (0.1 µg/kg-day) http://www.epa.gov/iris/subst/0073.htm http://www.epa.gov/hg/exposure.htm

Methods: Sociodemographic and Lifestyle Variables


Race/Ethnicity • Non-Hispanic White (White) • Non-Hispanic Black (Black) • Mexican American

Age (years) • 15-19, 20-26, 27-36, 37-44

Income ($1,000) • <15, 15-55, >55

Poverty-to-income Ratio • <2 or ≥2

Head of Household highest educational attainment • Less than high school graduate • High School/GED or some

college/AA degree • College graduate or more

Smoking status (serum cotinine)

• Non-smoker (<10 ng/mL) • Smoker (≥10 ng/mL)

Physical Activity •No or Yes

Methods: Data Analysis The association between race/ethnicity and higher NDT hazard (HI >1) was examined using logistic regression

• Sampling weights were used to account for complex survey design and to produce unbiased, national estimates

Covariates that changed the odds ratio (OR) between race/ethnicity and higher NDT hazard by ≥10% were included in the final model

The final model was stratified by AL groups (0, 1−2, ≥3)


RESULTS


Population Characteristics


More likely to have chronic stress (AL≥3) • Blacks

• Women 20-44 years

of age (compared with those 15-19 years of age)

• Lower socioeconomic status

• Smokers • No physical activity

More concern with NDT hazard (HI-NDT>1) • Blacks, Mexican

Americans • Women 20-44 years

of age (compared with those 15-19 years of age)

• Lower socioeconomic status

• Smokers

Presenter

Presentation Notes

Differences among by AL score compared with the overall population distribution, the percentage of blacks increases 67% and 37-44 year olds increases 50% among those with a higher AL score -<HS increases 31% -smoker increaess 26% -no PA 65% Differences by HI More than 25% for both Blacks (27%) and 37-44 year olds (33%) and nearly 50% for Mexicans (45%) -<HSincrease 50% -smoker 44%

Percentage of population by NDT Hazard


11

2

26

9 2

22 14

2

32 22

1

36

0 5

10 15 20 25 30 35 40

HQ-PB>1 HQ-MeHg >1 HI>1

Popu

latio

n (%

)

All Women White Black Mexican Americans

Presenter

Presentation Notes

This chart displays the percentage of women ages 15-44 years on the y-axis and groups with more concern for neurodevelopmental toxicity hazard on the x-axis. On the left you can see that 11% of all women had blood lead values greater than the chosen health reference value of 1.76 micrograms/deciliter. While the middle column shows that only 2% of all women had methylmercury blood values greater than the blood equivalent reference dose of 5.8 micrograms per liter. On the right is the hazard index representing the neurodevelopmental toxicity hazard to both Pb and MeHg. More than a quarter of women had a hazard index greater than one, indicating that although Pb and MeHg exposures alone may be below their respective health reference values that combined they are greater than one. Here is the same chart as the last with women stratified by race/ethnicity. On the left we can see that compared with White women, three times as many Mexican American women and two times as many black women had blood lead exposures greater than the chosen health reference value. In the middle, we can see there was little difference for methyl mercury exposure by race. On the right is the hazard index for the combined NDT hazard for Pb and MeHg. Here, we can see that compared to whites, there is a higher percentage of blacks and mexican americans with a hazard index greater than one-and further than blacks and mexican americans are similar. 16% of tot pop with low AL 25% with high AL (21 W, 42 B, 28 MA) AL>3=40% (37 W, 59 B, 41 MA), AL>4= 25% (21 W, 42 B, 28 MA), AL>5= 14 % (10 W, 27 B, 16 MA)—additional graphs on slide 22 Mean Pb = 1.1 (1.0 W, 1.4 B, 1.5 MA) MeHg= 1.2 (1.2 W, 1.4 B, 1.0 MA) What percentage of the population has HQ/HI>1 Pb=11% (8 W, 16 B, 25 MA) MeHg=2% (2 W, 2 B, 1 MA) HI=27% (23 W, 35 B, 38 MA)


Association between race/ethnicity and odds of having an HI >1 among women ages 15-44 years in NHANES 2003-04

*Adjusted for country of birth, age, education, and smoking

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

White Black Mexican American

White Black Mexican American

Unadjusted Adjusted*

Odd

s Rat

io (9

5% C

I)

1.7 (1.0, 2.6)

2.0 (1.3, 3.0)

2.2 (1.5, 3.4)

1.4 (0.8, 2.5)

Presenter

Presentation Notes

This chart displays the odds ratio and 95% confidence intervals for the association between race/ethnicity and odds of having a hazard ratio greater than 1 for neurodevelopmental toxicity of joint lead and methylmercury. On the y-axis is the odds ratio and 95% confidence interval stratified by race/ethnicity on the x-axis. In red are white women, the reference group; green=black, and purple= Mexican Americans. On the left we have the unadjusted association and can see that both blacks and Mexican Americans were more likely to have a hazard index >1, compared to white women. Additionally it can be noted that they had similar odds to each other. On the right is the association adjusted for age, head of household education, and smoking. First, note that the adjusted odds ratios for both Blacks and Mexican Americans are lightly larger than their respective unadjusted odds ratios. The odds of Blacks having a hazard index >1 were 2x the odds of Whites, and the odds of Mexican Americans having a hazard index >1 were nearly 3x the odds of Whites. OR for Blacks increases 22%, Mexicans increases 35% Age doesn’t change much but remains SS Education : -<HS decreases 43% (NS) Smoking increases 40%, remains SS


Possible effect measure modification of race/ethnicity and HI-NDT association by chronic stress

Adjusted* association between race/ethnicity and odds of having an elevated HI-NDT among non-pregnant reproductive-aged

women in the 2003–2004 NHANES stratified by chronic stress

*Adjusted for age, head of household education, and smoking

0 1 1 2 2 3 3 4 4 5 5

Adjusted (n=346)

Low (AL=0) n=86

Intermediate (AL 1–2)

n=194

High (AL ≥3)

n=67

Adjusted (n=312)

Low (AL=0) n=88

Intermediate (AL 1–2)

n=164

High (AL ≥3)

n=55

Black Mexican American

OR

(95%

CI)

for

HI>

1

9.5 14.1

Presenter

Presentation Notes

Here is a similar graph to the last but now I have stratified the adjusted association between race/ethnicity and odds of having a hazard index >1 by AL score groups. The blue boxes represent the un-stratified adjusted OR from the last graph, next we have the no chronic stress group, then the intermediate group and finally the chronic stress group. the green boxes are for blacks and the purple boxes are for mexican americans. For both blacks and mexican americans we found effect measure modification by AL score group. Specifically, we see the lowest odds ratios for the no chronic stress group and the highest odds ratios for the chronic stress group. Both blacks and mexican americans had similar odds to each other among the chronic stress group and their odds for having a hazard index >1 were nearly 4 times the odds of Whites. OR for Blacks increases 22%, Mexicans increases 35% Age doesn’t change much but remains SS Education : -<HS decreases 43% (NS) Smoking increases 40%, remains SS AL=0 (95 W, 40 B, 62 MA) AL>3 (99W, 108 B, 60 MA)

Pb was the main contributor to the HI-NDT • Mean HQPb 3-fold higher than the mean

HQMeHg (0.6 vs. 0.2)

Independent of country of birth, age, education, and smoking, Blacks were more likely than Whites to have an elevated HI-NDT Chronic stress modified the association

between NDT hazard and race/ethnicity


Summary

Use of exposures in non-pregnant women of reproductive age as surrogates of potential maternal/fetal exposures may not be representative of exposures during pregnancy

Other stressors that may be associated with neurodevelopmental outcomes were not included

Cross sectional data—One-time measurements performed on all stressors (AL biomarkers, Pb, MeHg)

Used HI to examine higher joint Pb and MeHg exposures May not reflect underlying mode of action Uncertainty in using dose-addition for similar endpoint


Discussion: Limitations

Presenter

Presentation Notes

Behavior and dietary changes can affect exposures Physiological changes occur that can affect blood concentrations (increased blood volume and plasma) Jones L, Parker JD, Mendola P (2010). NCHS Data Brief: Blood Lead and Mercury Levels in Pregnant Women in the United States, 2003–2008. CDC, available at: http://www.cdc.gov/nchs/data/databriefs/db52.pdf accessed on 25 October 2012. Other stressors may also be associated with neurodevelopmental toxicity that were not included We conducted sensitivity analyses including polychlorinated biphenyls Due to their low contribution to the neurotoxicity HI (<1%) and the small sample size (n=400), we chose to exclude them from the present analyses One-time measurements of all stressors (AL biomarkers, Pb, MeHg) Does not account for intra-individual variability or physiological and behavior changes that may occur at different lifestages that can modify exposures and/or doses Reasonable measures of current exposure due to relatively short half-lives (Pb~1 month9; MeHg~2 months10) AL is an indirect measure of chronic stress Due to the multi-system effects that chronic stress is hypothesized to cause, this is likely a better measure of long term/chronic stress than a single biomarker (e.g., cortisol) incorporating neuroendocrine AL biomarkers (cortisol, dehydroepiandrosterone sulfate, epinephrine, and norepinephrine) may strengthen its use as an indicator of chronic stress AL biomarkers are weighted equally-it is likely that specific physiological systems or biomarkers are more relevant to chronic stress during pregnancy and/or neurotoxicity Using the Hazard Index to represent higher joint Pb and MeHg exposures Although the reference dose is the basis for the MeHg blood reference value, this does not account for increasing MeHg concentrations associated with increased age11 There is no scientific consensus on whether or not there is a “safe” level of lead exposure. We have used a reference value identified from the literature that was not associated with lower IQ3 We conducted sensitivity analyses using a more health conservative value of 1.0 µg/dL identified by Cal EPA as a benchmark dose for public health safety12

Chronic stress, a non-chemical stressor, was found to modify the association between race/ethnicity and likelihood NDT hazard

This research highlights the importance of evaluating co-exposures (chemical and non-chemical) with a common endpoint

Results from these analyses could identify potentially susceptible populations for future epidemiological studies or be used by risk managers 06/12/2013 SOT RASS CRA Webinar 65

Conclusions

Presenter

Presentation Notes

Children born to older, non-White women who smoke may be a population sensitive to adverse neurodevelopmental outcomes due to elevated hazard from NDTs and chronic stress exposure

Thank you!

Questions . . .

Contact information:

Amanda M. Evans [email protected]

Glenn Rice

[email protected]

Linda Teuschler [email protected]

mailto:[email protected]�



cumulative risk assessment - society of toxicology

Documents