Neurocognitive Impairment Due to Early Life Arsenic …superfund.ciesin.columbia.edu/sfund_files/documents/events/... · Neurocognitive Impairment Due to Early Life Arsenic Exposure

Neurocognitive Impairment Due to Early Life Arsenic Exposure

Joseph Graziano, PhDProfessor of Environmental Health Sciences

And PharmacologyMailman School of Public Health

Columbia University

Seminar Outline

• Background:  Arsenicosis in Bangladesh

• Our Prospective Cohort Study in Adults

• Consequences of As Exposure in Children

History of Arsenic Problem in Bangladesh1960’s‐70’s 1 million tube‐wells installed by UNICEF

1970’s‐80’s 3‐4 million tube‐wells installed privately97% of population drink tube‐well water

1993‐94 Patients with arsenicosis diagnosed

1996‐97 Arsenic detected in tube‐well water.10 million tube wells now in use.25‐40 million people estimated to have been exposedRecognized as major public health problem

2000 Launch of the Health Effects of Arsenic Longitudinal Study (HEALS)

2009‐12 Bangladesh government adopts a rural deep well drilling program

Tube wells are the major source of drinking water in Bangladesh and many other developing countries

Arsenic in Groundwater: An International Problem

• Taiwan• China• Chile• Argentina• Mexico• United States

• India• Bangladesh• Nepal• Vietnam• Cambodia• Mongolia• Myanmar

Known Health Effects of Arsenic• “Early/Intermediate””

– Skin Lesions:• Melanosis• Keratosis

• Other Conditions– Diabetes mellitus– Non‐pitting edema– Non‐malignant respiratory disease

– Cognitive deficits**– Black‐foot Disease

• “Late”– Cancers:

• Skin• Bladder• Lung• Liver

– Cardiovascular:• Hypertension• Ischemic Heart Disease**

Well As (ug/L)

BGS/DPHE (2001)

0

200

400

600

800

1000

0 500 1000Well As (ug/L)

Dep

th (f

t)

Arsenic in 5,966 wells

Baseline

Cohort Recruitment and Follow-up:

The HEALS Cohort

Follow-up 1

7/2000-5/2002 9/2002-11/2004 12/2004 -9/2006 6/2007-1/2009Follow-up 3

HEALS Original Cohort

~12,000 adults 6/2006-8/2008 11/2008-11/2009

Follow-up 2

Baseline Follow-up 1

HEALS Expanded Cohort

~8,000 adults

2010-2012

~35,000 adults:HEALS original andExpanded cohort

Baseline

Cohort Recruitment and Follow-up

Follow-up 1

7/2000-5/2002 9/2002-11/2004 12/2004 -9/2006 6/2007-1/2009Follow-up 3

HEALS Original Cohort

~12,000 adults 6/2006-8/2008 11/2008-11/2009

Follow-up 2

Baseline Follow-up 1

HEALS Expanded Cohort

~8,000 adults

2010-2012

~35,000 adults:HEALS original andExpanded cohort

113 deaths 120 deaths 174 deaths

A validated verbal autopsy was used to classify deaths using WHO’s ICD-10

Table 2. Hazard ratio for mortality of HEALS participants in relation to baseline arsenic exposure, Bangladesh

Arsenic exposure No. of deaths

All-cause mortality* No. of

deathsChronic disease mortality*

HR 95% CI HR 95% CIWell water arsenic (μg/L)

0.1–10 74 1.00 Referent 58 1.00 Referent10.1–50 90 1.34 0.99, 1.82 69 1.33 0.94, 1.8750.1–150 98 1.09 0.81, 1.47 83 1.22 0.87, 1.70150.1–854 131 1.68 1.26, 2.23 101 1.68 1.21, 2.33

P for trend 0.003 0.005

Daily arsenic dose (μg/day)0.041–35.0 87 1.00 Referent 66 1.00 Referent35.1–163.0 97 1.10 0.83, 1.47 80 1.21 0.88, 1.67163.1–401.0 91 1.09 0.81, 1.46 76 1.22 0.88, 1.71401.1–4898.0 118 1.54 1.17, 2.04 89 1.58 1.15, 2.18

P for trend 0.004 0.007

Urinary total arsenic (μg/g Cr)7.0–105.0 83 1.00 Referent 64 1.00 Referent105.1–199.0 96 1.07 0.80, 1.43 80 1.17 0.84, 1.62199.1–352.0 100 1.22 0.91, 1.63 83 1.37 0.98, 1.90352.1–5000.0 105 1.45 1.09, 1.94 77 1.47 1.05, 2.06

P for trend 0.008 0.01HR=hazard ratio; CI=confidence interval; Cr=creatinine.* Multivariate estimates adjusted for age, sex, body mass index, systolic blood pressure, education, and smoking status.

All‐cause and chronic disease mortality

Argos et al, Lancet 376: 252-8, 2010

“… During the summer of 1955, mass arsenic poisoning of bottle‐fed infants occurred in the western part of Japan because of contaminated milk powder.”

Environmental Health, 2006

“…This experience provides clear‐cut evidence for….developmental neurotoxicity. At the present time, more than 600 surviving victims, now in their 50s, have been reported to suffer from….mental retardation, neurological diseases and other disabilities.”

Environmental Health, 2006

Source:  Bangladesh Rehabilitation Assistance Committee (BRAC)

Maternal Blood As (μg/L)

0 20 40 60 80

Um

bilic

al C

ord

Bloo

d As

(μg/

L)

0

20

40

60

80

r=0.93b=0.94

A

Hall et al, EHP 115: 1503‐1509, 2007

Relationship Between Water Arsenic Concentrations and Intellectual Function

Wasserman et al, EHP, 2004

Adjusted

 WISC III Raw

 Score

Wasserman et al, EHP 2004

“Mn by As interactions were not significant in adjusted or unadjusted models.  Findings are consistent with other reports documenting adverse impact of both As and 

Mn exposure on child development outcomes.”

Conclusions: A greater decrease in UAs/Cr was associated with significantly greater improvements in WISC IV Working Memory scores, but NOT with a greater improvement in other  scores.

The provision of deep, low‐arsenic tube wells, coupled with education, lowered biomarkers of exposure (measured after two years):

Wasserman et al, EHP, 2016

Elevated groundwater arsenic from natural sources

USGS, 2003 About 30,000 monitoring wells

Arsenic concentrations in 1,428 domestic wells from 17 towns in central Maine, with 31% of the wells containing > 10 ug/L [As] (source: Yang 2010).

Environmental Health, 2014

Estimating IQ points lost* attributable to water arsenic, comparing those with  WAs > 5 ppb vs WAs <5ppb

Full Scale IQ: ‐4.0 (CI ‐6.8, ‐1.2); p<0.01

Perceptual Reasoning:  ‐4.4 (CI ‐7.4, ‐1.4); p<0.01Working Memory:   ‐3.5 (CI ‐6.6, ‐0.3); p<0.05Verbal Comprehension:  ‐3.1 (CI ‐6.7, 0.5); p<0.10

Wasserman et al, Environ Health, 2014

*Adjusted for mother’s IQ and education, HOME scores, number of siblings and school district.

Arsenic in 5,966 wells

Ongoing: Lifetime Exposure and Health Outcomes in Adolescents

This studyInclusion criteria

HEALS

Baseline  Follow‐up 4Follow‐up 3Follow‐up 2Follow‐up 1

Mom

Child

Outcomes:  WISC IV, CANTAB, PFTs

0

50

100

150

200

250

Baseline Follow‐up 1 Follow‐up 2 Follow‐up3N Baseline Follow‐up 1 Follow‐up 2 Follow‐up 3

200 Low, 3.0 Low, 2.9 Low, 2.9 Low, 2.9

246 Mod, 26.5 Mod, 26.2 Mod, 26.3 Mod, 26.3

192 High, 148.2 High, 146.5 High, 146.2 High, 146.2

85 High, 200.8 Mod, 13.8 Mod, 12.1 Mod, 12.1

Water arsen

ic (u

g/L)

Mother’s water arsenic exposure over time

WAs<10 µg/Ln = 200

WAs 10‐50µg/Ln = 246

WAs>50 µg/Ln = 192

Measures Mean ± SD  or % (n)

Mean ± SD  or % (n)

Mean ± SD  or % (n)

p‐value

Child characteristics

Male (%) 46 (92) 43.9 (108) 45.83 (88) 0.8828

Child age (years) 14.77 ± 0.73 14.78 ± 0.67 14.72 ± 0.67 0.6403

Years in school  6.78 ± 2.05 7.01 ± 2.0 6.63 ± 2.28 0.1404

Currently attending school (%) 84.9 (169/199) 86.1 (210/244) 80.2 (154) 0.2385

Exposure characteristics

Prenatal well water As (µg/L)Concurrent well water As (µg/L)

3.2 ± 2.97.2 ± 16.3

27.7 ± 11.229.7 ± 35.9

148.5 ±104.383.3 ± 93.3

<.0001

Biological measures n = 194 n = 243 n = 187

UAs (µg/L) 37.1 ± 32.9  78.0 ± 110.6 138.3 ± 180.1 <.0001

UAs/Cr (µg/g Cr) 81.3 ± 102.7  139.0 ± 113.1  275.7 ±324.1 <.0001

n = 195 n = 243 n = 185

BMn (µg/L)  11.8 ± 4.5  11.3 ± 3.4  11.2 ± 3.3  0.8933

BAs (µg/L)  2.5 ± 1.6  4.5 ± 3.2  7.9 ± 6.5 <.0001

BPb (µg/dL)  10.6 ± 4.8 9.5 ± 4.1 9.3 ± 4.2 0.0285

BCd (µg/L)  0.7 ± 0.3  0.6 ± 0.3  0.5 ± 0.2  <.0001

BSe (µg/L)  130.8 ± 17.0  133.3 ± 18.0 133.0 ± 19.4  0.3547

Selected Characteristics of the Adolescent Study

WAs<10 µg/Ln = 200

WAs 10‐50µg/Ln = 246

WAs>50 µg/Ln = 192

Measures Mean ± SD  or % (n)

Mean ± SD  or % (n)

Mean ± SD  or % (n)

p‐value

Child characteristics

Male (%) 46 (92) 43.9 (108) 45.83 (88) 0.8828

Child age (years) 14.77 ± 0.73 14.78 ± 0.67 14.72 ± 0.67 0.6403

Years in school  6.78 ± 2.05 7.01 ± 2.0 6.63 ± 2.28 0.1404

Currently attending school (%) 84.9 (169/199) 86.1 (210/244) 80.2 (154) 0.2385

Exposure characteristics

Pre‐natal well water As (µg/L)Concurrent well water As (µg/L)

3.2 ± 2.97.2 ± 16.3

27.7 ± 11.229.7 ± 35.9

148.5 ±104.383.3 ± 93.3

<.0001<.0001

Biological measures n = 194 n = 243 n = 187

UAs (µg/L) 37.1 ± 32.9  78.0 ± 110.6 138.5 ± 180.1  <.0001

UAs/Cr (µg/g Cr) 81.3 ± 102.7  139 ± 113.1  275.7 ± 324.1 <.0001

n = 195 n = 243 n = 185

BMn (µg/L)  11.8 ± 4.5  11.3 ± 3.4  11.2 ± 3.3  0.8933

BAs (µg/L)  2.5 ± 1.6  4.5 ± 3.2  7.9 ± 6.5  <.0001

BPb (µg/dL)  10.6 ± 4.8 9.5 ± 4.1 9.3 ± 4.2 0.0285

BCd (µg/L)  0.7 ± 0.3  0.6 ± 0.3  0.5 ± 0.2  <.0001

BSe (µg/L)  130.8 ± 17.0  133.3 ± 18.0 133.0 ± 19.4  0.3547

Selected Characteristics of the Study Participants

WAs<10 µg/Ln = 200

WAs 10‐50µg/Ln = 246

WAs>50 µg/Ln = 192

Measures Mean ± SD  or % (n)

Mean ± SD  or % (n)

Mean ± SD  or % (n)

p‐value

Child characteristics

Male (%) 46 (92) 43.9 (108) 45.83 (88) 0.8828

Child age (years) 14.77 ± 0.73 14.78 ± 0.67 14.72 ± 0.67 0.6403

Years in school  6.78 ± 2.05 7.01 ± 2.0 6.63 ± 2.28 0.1404

Currently attending school (%) 84.9 (169/199) 86.1 (210/244) 80.2 (154) 0.2385

Exposure characteristics

Pre‐natal well water As (µg/L)Concurrent well water As (µg/L)

3.2 ± 2.97.2 ± 16.3

27.7 ± 11.229.7 ± 35.9

148.5 ±104.383.3 ± 93.3

<.0001<.0001

Biological measures n = 194 n = 243 n = 187

UAs (µg/L)     37.1 ± 32.9  78.0 ± 110.6 138.5 ± 180.1  <.0001

UAs/Cr (µg/g Cr) 81.3 ± 102.7  139 ±113.1  275.7 ± 324.1 <.0001

n = 195 n = 243 n = 185

BMn (µg/L)  11.8 ± 4.5 11.3 ± 3.4  11.2 ± 3.3  0.8933

BAs (µg/L)  2.5 ± 1.6  4.5 ± 3.2 7.9 ± 6.5  <.0001

BPb (µg/dL)  10.6 ± 4.8 9.5 ± 4.1 9.3 ± 4.2 0.0285

BCd (µg/L)  0.7 ± 0.3  0.6 ± 0.3  0.5 ± 0.2  <.0001

Selected Characteristics of the Study Participants

WAs<10 µg/Ln = 200

WAs 10‐50µg/Ln = 246

WAs>50 µg/Ln = 192

Measures Mean ± SD  or % (n)

Mean ± SD  or % (n)

Mean ± SD  or % (n)

p‐value

Child characteristics

Male (%) 46 (92) 43.9 (108) 45.83 (88) 0.8828

Child age (years) 14.77 ± 0.73 14.78 ± 0.67 14.72 ± 0.67 0.6403

Years in school  6.78 ± 2.05 7.01 ± 2.0 6.63 ± 2.28 0.1404

Currently attending school (%) 84.9 (169/199) 86.1 (210/244) 80.2 (154) 0.2385

Exposure characteristics

Pre‐natal well water As (µg/L)Concurrent well water As (µg/L)

3.2 ± 2.97.2 ± 16.3

27.7 ± 11.229.7 ± 35.9

148.5 ±104.383.3 ± 93.3

<.0001<.0001

Biological measures n = 194 n = 243 n = 187

UAs (µg/L)       37.1 ± 32.9  78.0 ± 110.6 138.5 ± 180.1  <.0001

UAs/Cr (µg/g Cr) 81.3 ± 102.7  139.0 ± 113.1 275.7 ± 324.1 <.0001

n = 195 n = 243 n = 185

BMn (µg/L)  11.8 ± 4.5  11.3 ± 3.4  11.2 ± 3.3  0.8933

BAs (µg/L)  2.5 ± 1.6  4.5 ± 3.2  7.9 ± 6.5  <.0001BPb (µg/dL)  10.6 ± 4.8 9.5 ± 4.1 9.3 ± 4.2 0.0285

BCd (µg/L)  0.7 ± 0.3  0.6 ± 0.3  0.5 ± 0.2  <.0001

Selected Characteristics of the Study Participants

Adjusted Full Scale WISC Raw Scores by Group

Group 1 Group 2 Group 3 Group 4

Adj

uste

d Fu

ll S

cale

Raw

Sco

re

140

150

160

170

180

190

(Low WAs) (Mod WAs) (Hi WAs) (Hi→Mod WAs)

Adjusted Change in Raw Scores by Quartiles of Urine Arsenic and Blood Arsenic

Urine Arsenic (µg/g) Creatinine Blood Arsenic (µg/L)

Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

Full S

cale

Raw

Sco

re

140

150

160

170

180

190

(15.8, 58.1)(58.1, 104.2) (104.2, 191.9)(191.9, 2886.9)

(0.75, 2.18)((2.18, 3.50)(3.50, 5.85)(5.85, 50.91)

UAs/Cr rangeby quartile

BAs  rangeby quartile

After adjustment for mother’s IQ, father’s education, BPb, BCd, BMn housing type, child’s years of education, head circumference, and group.“Working Memory” subscale drives the finding.

Summary and Conclusions1. We have repeatedly observed adverse associations between well 

water As exposure and child intelligence in young children in Bangladesh.

2. We observed similar findings in 3rd‐5th grade children drinking As contaminated water in Maine.

3. In Bangladesh, the provision of deep, low‐As tubewells, coupled with education, was effective in substantially lowering biomarkers of As exposure, but after two years, only Working Memory had improved.

4. Among adolescents whose early life water As exposure was well characterized, concurrent biomarkers of As exposure were significantly associated with decrements in Full Scale raw scores on the WISC‐IV.  This finding was driven by adverse associations between As biomarkers and the Working Memory subscale.

AcknowledgementsIn the United States• Gail Wasserman• Xinhua Liu• Diane Levy• Nancy LoIacono• Faruque Parvez• Pam Factor‐Litvak• Jennie Kline• Vesna Slavkovich• Jagoda Balac• David Santiago• Lex van Geen• Tyler Ellis• Yan Zheng• Habibul Ahsan (Univ. Chicago)• Yu Chen (NYU)

In Bangladesh• Tariqul Islam• Hasan Shahriar• Mohammed Uddin• Abu Siddique• Samiul Islam• Rebeka Sultana• Ruksana Sultana• Abul Kamal Samsuddin• Golam Sarwar• Study Participants

Superfund Research Program grant P42 ES 10349; NIEHS Center Grant P30 ES 09089Water Aid Bangladesh

Thank You