montreal lead study: association between cumulative lead exposure index and blood lead levels in 1-5...

Post on 11-Jan-2016

213 Views

Category:

Documents

1 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Montreal Lead Study: Association between

Cumulative Lead exposure index and Blood lead levels in

1-5 Years-old ChildrenGerard Ngueta, (PhD Candidate in Epidemiology)

Research Advisor: Dr. Patrick Levallois

GG

Gerard.Ngueta@crchuq.ulaval.ca1

G mt

Relevance of Montreal Lead StudyDifference from earlier association studies

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

2

Montreal Lead Study

Undertaken in 2009 and firstly designed to assess lead exposure in 1-5 years old children living in Montreal city

US E.P.A reports indicating the global decrease of environmental lead exposure and huge reduction of BLL in young children.

Montreal Study remains relevant because of poor data published in Canadian children (aged under 6).

G mt

Relevance of Montreal Lead StudyDifference from earlier association studies

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

3

Strong association between environmental lead and BLL

Reported in several previous studies and reports

In most cases, they were cross-sectional association studies

Most of children included were those with high risk of lead exposure [Black and Hispanic origin > 50% population study]

The present study differs from earlier: 1) Caucasian children ≈ 67% of population study2) Cumulative lead exposure index (CLEI) in addition to cross-

sectional measures of exposure

G mt

Relevance of Montreal Lead StudyDifference from earlier association studies

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

4

Motivation for using CLEI ? Previous studies indicated that lead has a half-life of approximately

30 days in the blood

Greenbalt reported that the time required to find steady-state is approximately four-to-five times the elimination half-life after a single exposure [Annual review of medicine 36, 421-427]

There is a scientific reason to believe that BLL at a given time is more related to lead exposure experienced during 3-5 months before that time.

G mt

Relevance of Montreal Lead StudyDifference from earlier association studies

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

5

Motivation for using CLEI ?

Days Environmental lead Blood lead

149 X149

148 X148

…. ….

3 X3

2 X2

1 X1

Day of visit X0 Y

G mt

Relevance of Montreal Lead StudyDifference from earlier association studies

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

6

Motivation for using CLEI ?

Days Environmental lead Blood lead

149 X149

148 X148

…. ….

3 X3

2 X2

1 X1

Day of visit X0 Y

G mt

Relevance of Montreal Lead StudyDifference from earlier association studies

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

7

Motivation for using CLEI ? Previous studies indicated that lead has a half-life of approximately

30 days in the blood

Greenbalt reported that the time required to find steady-state is approximately four-to-five times the elimination half-life after a single exposure [Annual review of medicine 36, 421-427]

There is a scientific reason to believe that PbS at a given time is more related to lead exposure experienced during 3-5 months before that time.We hypothesize that cumulative lead exposure would be more suitable to assess the association between Environmental lead and BLL

G mt

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

8

Objectives of study To estimate the effects of both CLEI and cross-sectional measures of

water lead and residential dust lead on BLL in children aged 1 to 5.

To assess the association between both measures of exposure and the likelihood of BLL greater or equal to 2 µg/dl.

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

9

Where are children from ? Selected from four large districts:

1) Mercier-Hochelaga-Maisonneuve2) Saint-Laurent3) Verdun4) Villeray-Saint-Michel-Parc-Extension

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

10

Who are they?

306 children aged under 6

Child should drink tap water regularly [1]be born in Canada, [2]live in the residence for at least one year, [3]not live outside home more than 2 days per week [4] no suffer from severe disease [5]

Families should not use a water filtration system [6]Families should live in single-house unit or multi-units apartments with 3 levels or less [7]

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

11

Who are they?

Age (Months)

12-23 50 (16%)

24-35 66 (22%)

36-71 190 (62%)

Girls 153 (50%)

Caucasians 207 (68%)

Day Care attendance 229 (75%)

Mother with university degree

221 (73%)

Hand-to-Mouth behaviour 148 (49%)

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

12

Who are they?

Age (Months)

N (%) BLL † BLL ≥ 2 µg/dL

12-23 50 (16.0%) 1.27 (0.91 – 1.76)

7 (14.3%)

24-35 66 (22.0%) 1.40 (1.00 – 1.91)

16 (24.6%)

36-71 190 (62.0%) 1.31 (0.95 – 1.72)

30 (16.3%)

Median (p25-p75): Expressed in µg/dL

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

13

Samples collection

Largely described elsewhere

Water samples

WLLF5 WLLS1

WLLS2

WLLS3

WLLS45 min

flushing30 min of

stagnation

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

14

Sample collection

Largely described elsewhere

Water samples

Dust samplesWindowsill dust : Floor dust: sampled in the center of the available floor space 1) Child’s room 2) Home entrance 3) Another room frequently used

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

15

Sample collection

Largely described elsewhere

Water samples

Dust samples

Paint samples

1) X-ray fluorescence evaluation2) Collection of paint chips and lab measurement of lead level

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

16

Cumulative lead exposure index

Firstly: Modeling seasonal changes of environmental lead

CLEI for water

CLEI = Qe*0.50*∑ [WLLi*exp(-(Ln 2/30)*(N-i))] (expressed in µg)

Daily amount of water consumed

mean water lead level observed for the day i before

day of visit

Gastrointestinal absorption rate

N from 0 to 149

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

17

Cumulative lead exposure index

Firstly: Modeling seasonal changes of environmental lead

CLEI for water

CLEI = Qe*0.50*∑ [WLLi*exp(-(Ln 2/30)*(N-i))] (expressed in µg)

Worst-case exposure scenario: refers to the highest water lead concentration in the samples collected after stagnation time of 30 minutes. Best-case exposure scenario: refers to the sample collected after 5 minutes of flushing

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

18

Cumulative lead exposure index

Firstly: Modeling seasonal changes of environmental lead

CLEI for water

CLEI = Qe*0.50*∑ [WLLi*exp(-(Ln 2/30)*(N-i))] (expressed in µg)

CLEI for dust

CLEI = Qe*0.30*DLL*∑ [exp(-(Ln 2/30)*(N-i))] (expressed in µg)

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

19

Statistical analysis

Linear regression Ordinary least square evaluation Effect estimates Standardized partial regression coefficients Independent contribution of each exposure

To estimate the effects of both CLEI and cross-sectional measures of water lead and residential dust lead on BLL in children aged 1 to 5.

G mt

Population StudySamples collectionCumulative exposure variablesOverview on Statistical analysis

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

20

Statistical analysis

Logistic regression Maximum-likelihood estimation Effect estimates

To assess the association between both measures of exposure and the likelihood of BLL greater or equal to 2 µg/dl.

Crude models

Semi-adjusted models

Adjusted models

Each exposure variable alone

Models including all exposure variables

Full model including all covariates

Propensity Score methods

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

21

Unadjusted linear effects of environmental lead on BLL

Cumulative lead exposure index † Cross-sectional measure of exposure ††

  Estimate 95% CI R-Square   Estimate 95%CI R-Square

Intercept 0.1459 0.0714; 0.22050.0986

  Intercept 0.2116 0.1393; 0.28380.0353

Water (Worst scenario) 0.0665 0.0431; 0.0899 Water (Worst scenario) 0.0170 0.0067; 0.0274

Intercept 0.1587 0.0866; 0.23080.0964

Intercept 0.1786 0.1050; 0.25210.0667

Water (Best scenario) 0.1231 0.0792; 0.1670 Water (Best scenario) 0.0591 0.0338; 0.0844

Intercept 0.2318 0.1650; 0.2987 0.0501 Intercept 0.2295 0.1613; 0.2977

0.0465Windowsill dust loading 0.0003 0.0001; 0.0004   Windowsill dust loading

0.0033 0.0015; 0.0051

Intercept 0.2412 0.1761; 0.30640.0294

Intercept 0.2440 0.1793; 0.30880.0277

Floor dust loading 0.0022 0.0007; 0.0036 Floor dust loading 0.0243 0.0078; 0.0408

1) R-Squares obtained with CLEI were greater than those by modeling cross-sectional exposure.

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

22

Unadjusted linear effects of environmental lead on BLL

Cumulative lead exposure index † Cross-sectional measure of exposure ††

  Estimate 95% CI R-Square   Estimate 95%CI R-Square

Intercept 0.1459 0.0714; 0.22050.0986

  Intercept 0.2116 0.1393; 0.28380.0353

Water (Worst scenario) 0.0665 0.0431; 0.0899 Water (Worst scenario) 0.0170 0.0067; 0.0274

Intercept 0.1587 0.0866; 0.23080.0964

Intercept 0.1786 0.1050; 0.25210.0667

Water (Best scenario) 0.1231 0.0792; 0.1670 Water (Best scenario) 0.0591 0.0338; 0.0844

Intercept 0.2318 0.1650; 0.2987 0.0501 Intercept 0.2295 0.1613; 0.2977

0.0465Windowsill dust loading 0.0003 0.0001; 0.0004   Windowsill dust loading

0.0033 0.0015; 0.0051

Intercept 0.2412 0.1761; 0.30640.0294

Intercept 0.2440 0.1793; 0.30880.0277

Floor dust loading 0.0022 0.0007; 0.0036 Floor dust loading 0.0243 0.0078; 0.0408

2) Low values of R-squares (<9%) indicate that an exposure source taking alone is not enough to explain the variability observed in the log(BLL).

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

23

Unadjusted linear effects of environmental lead on BLL

Cumulative lead exposure index † Cross-sectional measure of exposure ††

  Estimate 95% CI R-Square   Estimate 95%CI R-Square

Intercept 0.1459 0.0714; 0.22050.0986

  Intercept 0.2116 0.1393; 0.28380.0353

Water (Worst scenario) 0.0665 0.0431; 0.0899 Water (Worst scenario) 0.0170 0.0067; 0.0274

Intercept 0.1587 0.0866; 0.23080.0964

Intercept 0.1786 0.1050; 0.25210.0667

Water (Best scenario) 0.1231 0.0792; 0.1670 Water (Best scenario) 0.0591 0.0338; 0.0844

Intercept 0.2318 0.1650; 0.2987 0.0501 Intercept 0.2295 0.1613; 0.2977

0.0465Windowsill dust loading 0.0003 0.0001; 0.0004   Windowsill dust loading

0.0033 0.0015; 0.0051

Intercept 0.2412 0.1761; 0.30640.0294

Intercept 0.2440 0.1793; 0.30880.0277

Floor dust loading 0.0022 0.0007; 0.0036 Floor dust loading 0.0243 0.0078; 0.0408

3) Significant crude association between Log(BLL) and Water lead, Windowsill dust loading and Floor dust loading

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

24

Adjusted linear effects of CLEI on BLL

1) Water lead and windowsill dust remain markedly associated with Log(BLL) (β=0.0758, p=0.0002 and β=0.0002, p=0.0058 respectively).

  Unstandardized estimate

95% CI Standardized estimate

95% CI

Worst scenario        

Intercept -0.0482 -0.9391; 0.8427 0.1747 -0.6928; 1.0421

Water 0.0758 -0.0359; 0.1156 0.1819 -0.1483; 0.4565

Windowsill dust loading 0.0002 -0.0001; 0.0004 0.0967 -0.0665; 0.3158

Floor dust loading 0.0007 -0.0014; 0.0028 0.0277 -0.0959; 0.1995

Paint lead        

Reference level 0   0  

Level-1 exposure 0.0735 -0.0885; 0.2356 0.0735 -0.0885; 0.2356

Level-2 exposure 0.2789 0.0599; 0.4979 0.2789 0.0599; 0.4979

Adjusted R-Square 0.2115      

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

25

Adjusted linear effects of CLEI on BLL

2) paint-lead effect was also strongly significant in children exposed to paint chips with high lead level (paint chips ≥ 5000 mg/kg) compared with those living in homes without paint chips and low lead level in painted surface (XRF < 1 mg/cm2).

  Unstandardized estimate

95% CI Standardized estimate

95% CI

Worst scenario        

Intercept -0.0482 -0.9391; 0.8427 0.1747 -0.6928; 1.0421

Water 0.0758 -0.0359; 0.1156 0.1819 -0.1483; 0.4565

Windowsill dust loading 0.0002 -0.0001; 0.0004 0.0967 -0.0665; 0.3158

Floor dust loading 0.0007 -0.0014; 0.0028 0.0277 -0.0959; 0.1995

Paint lead        

Reference level 0   0  

Level-1 exposure 0.0735 -0.0885; 0.2356 0.0735 -0.0885; 0.2356

Level-2 exposure 0.2789 0.0599; 0.4979 0.2789 0.0599; 0.4979

Adjusted R-Square 0.2115      

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

26

Adjusted linear effects of cross-sectional exposure on BLL

1) Water lead effect was no longer statistically significant (β=0.0106, p=0.0884).

  Unstandardized estimate

95% CI Standardized estimate

95% CI

Worst scenario        

Intercept 0.3656 -0.5196; 1.2508 0.4888 -0.3889; 1.3664

Water 0.0106 -0.0016; 0.0227 0.0598 -0.0091; 0.1288

Windowsill dust loading 0.0034 -0.0013; 0.0055 0.1079 -0.0404; 0.1754

Floor dust loading 0.0094 -0.0147; 0.0336 0.0322 -0.0502; 0.1146

Paint lead        

Reference level 0   0  

Level-1 exposure 0.1179 -0.0463; 0.2820 0.1179 -0.0463; 0.2820

Level-2 exposure 0.3008 0.0710; 0.5306 0.3008 0.0710; 0.5306

Adjusted R-Square 0.1557      

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

27

Adjusted linear effects of cross-sectional exposure on BLL

2) Adjusted R-square remains low by modeling cross-sectional exposure (0.1557 versus 0.2115 for CELI) .

  Unstandardized estimate

95% CI Standardized estimate

95% CI

Worst scenario        

Intercept 0.3656 -0.5196; 1.2508 0.4888 -0.3889; 1.3664

Water 0.0106 -0.0016; 0.0227 0.0598 -0.0091; 0.1288

Windowsill dust loading 0.0034 -0.0013; 0.0055 0.1079 -0.0404; 0.1754

Floor dust loading 0.0094 -0.0147; 0.0336 0.0322 -0.0502; 0.1146

Paint lead        

Reference level 0   0  

Level-1 exposure 0.1179 -0.0463; 0.2820 0.1179 -0.0463; 0.2820

Level-2 exposure 0.3008 0.0710; 0.5306 0.3008 0.0710; 0.5306

Adjusted R-Square 0.1557      

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

28

Association between CLEI (worst scenario) and BLL ≥ 2 µg/dl

  Crude OR (95% CI) Adjusted OR (95% CI)

    Model 1a Model 2b

Water (Worst scenario) (µg/kg of bw)      < 1.12 1 1 1

1.12 - 2.92 1.20 (0.65 - 2.21) 0.69 (0.29 - 1.65) 1.11 (0.51-2.41)≥ 2.92 4.00 (2.41 - 6.65) 3.57 (1.74 - 7.32) 3.70 (1.93-7.11)

Windowsill dust loading (µg/kg of bw)      

< 87.03 1 1 187.03 - 239.67 0.92 (0.49 - 1.73) 0.42 (0.16 - 1.10) 0.76 (0.35-1.72)

≥ 239.67 2.53 (1.55 - 4.15) 1.03 (0.45 - 2.37) 1.98 (1.03-3.82)Floor dust loading (µg/kg of bw)      

< 8.12 1 1 18.12 - 20.47 2.09 (1.22-3.59) 3.64 (1.58 - 8.39) 1.97 (1.01-3.87)

≥ 20.47 2.44 (1.42-4.21) 3.44 (1.46 - 8.07) 2.25 (1.13-4.48)

Paint lead c      

Reference level 1 1 1

Level-1 exposure 1.16 (0.68-1.97) 1.48 (0.65 - 3.37) 1.10 (0.56-2.14)Level-2 exposure 3.23 (1.78-5.84) 6.08 (2.21 - 16.71) 3.04 (1.44-6.42)

-2Log(Likelihood) 516.85 343.82  

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

29

Association between CLEI (worst scenario) and BLL ≥ 2 µg/dl

  Crude OR (95% CI) Adjusted OR (95% CI)

    Model 1a Model 2b

Water (Worst scenario) (µg/kg of bw)      < 1.12 1 1 1

1.12 - 2.92 1.20 (0.65 - 2.21) 0.69 (0.29 - 1.65) 1.11 (0.51-2.41)≥ 2.92 4.00 (2.41 - 6.65) 3.57 (1.74 - 7.32) 3.70 (1.93-7.11)

Windowsill dust loading (µg/kg of bw)      

< 87.03 1 1 187.03 - 239.67 0.92 (0.49 - 1.73) 0.42 (0.16 - 1.10) 0.76 (0.35-1.72)

≥ 239.67 2.53 (1.55 - 4.15) 1.03 (0.45 - 2.37) 1.98 (1.03-3.82)Floor dust loading (µg/kg of bw)      

< 8.12 1 1 18.12 - 20.47 2.09 (1.22-3.59) 3.64 (1.58 - 8.39) 1.97 (1.01-3.87)

≥ 20.47 2.44 (1.42-4.21) 3.44 (1.46 - 8.07) 2.25 (1.13-4.48)

Paint lead c      

Reference level 1 1 1

Level-1 exposure 1.16 (0.68-1.97) 1.48 (0.65 - 3.37) 1.10 (0.56-2.14)Level-2 exposure 3.23 (1.78-5.84) 6.08 (2.21 - 16.71) 3.04 (1.44-6.42)

-2Log(Likelihood) 516.85 343.82  

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

30

Association between CLEI (worst scenario) and BLL ≥ 2 µg/dl

  Crude OR (95% CI) Adjusted OR (95% CI)

    Model 1a Model 2b

Water (Worst scenario) (µg/kg of bw)      < 1.12 1 1 1

1.12 - 2.92 1.20 (0.65 - 2.21) 0.69 (0.29 - 1.65) 1.11 (0.51-2.41)≥ 2.92 4.00 (2.41 - 6.65) 3.57 (1.74 - 7.32) 3.70 (1.93-7.11)

Windowsill dust loading (µg/kg of bw)      

< 87.03 1 1 187.03 - 239.67 0.92 (0.49 - 1.73) 0.42 (0.16 - 1.10) 0.76 (0.35-1.72)

≥ 239.67 2.53 (1.55 - 4.15) 1.03 (0.45 - 2.37) 1.98 (1.03-3.82)Floor dust loading (µg/kg of bw)      

< 8.12 1 1 18.12 - 20.47 2.09 (1.22-3.59) 3.64 (1.58 - 8.39) 1.97 (1.01-3.87)

≥ 20.47 2.44 (1.42-4.21) 3.44 (1.46 - 8.07) 2.25 (1.13-4.48)

Paint lead c      

Reference level 1 1 1

Level-1 exposure 1.16 (0.68-1.97) 1.48 (0.65 - 3.37) 1.10 (0.56-2.14)Level-2 exposure 3.23 (1.78-5.84) 6.08 (2.21 - 16.71) 3.04 (1.44-6.42)

-2Log(Likelihood) 516.85 343.82  

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

31

Association between CLEI (worst scenario) and BLL ≥ 2 µg/dl

  Crude OR (95% CI) Adjusted OR (95% CI)

    Model 1a Model 2b

Water (Worst scenario) (µg/kg of bw)      < 1.12 1 1 1

1.12 - 2.92 1.20 (0.65 - 2.21) 0.69 (0.29 - 1.65) 1.11 (0.51-2.41)≥ 2.92 4.00 (2.41 - 6.65) 3.57 (1.74 - 7.32) 3.70 (1.93-7.11)

Windowsill dust loading (µg/kg of bw)      

< 87.03 1 1 187.03 - 239.67 0.92 (0.49 - 1.73) 0.42 (0.16 - 1.10) 0.76 (0.35-1.72)

≥ 239.67 2.53 (1.55 - 4.15) 1.03 (0.45 - 2.37) 1.98 (1.03-3.82)Floor dust loading (µg/kg of bw)      

< 8.12 1 1 18.12 - 20.47 2.09 (1.22-3.59) 3.64 (1.58 - 8.39) 1.97 (1.01-3.87)

≥ 20.47 2.44 (1.42-4.21) 3.44 (1.46 - 8.07) 2.25 (1.13-4.48)

Paint lead c      

Reference level 1 1 1

Level-1 exposure 1.16 (0.68-1.97) 1.48 (0.65 - 3.37) 1.10 (0.56-2.14)Level-2 exposure 3.23 (1.78-5.84) 6.08 (2.21 - 16.71) 3.04 (1.44-6.42)

-2Log(Likelihood) 516.85 343.82  

G mt

Unadjusted linear effectsAdjusted CLEI effectsAdjusted Cross-sectional effectsLikelihood of BLL ≥ 2 µg/dL

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS AND PERSPECTIVES

32

Association between CS (worst scenario) and BLL ≥ 2 µg/dl

  Crude OR (95% CI) Adjusted OR (95% CI)

    Model 1a Model 2b

Water (Worst scenario) (µg/L)      

< 2.16 1 1 1

2.16 - 4.96 1.51 (0.82 - 2.76) 1.48 (0.64 - 3.44) 1.40 (0.64 - 3.03)≥ 4.96 4.47 (2.64 - 7.57) 3.17 (1.50 - 6.71) 4.21 (2.16 - 8.19)

Windowsill dust loading (µg/ft2)      

< 7.15 1 1 17.15 – 20.69 1.19 (0.65 - 2.17) 0.96 (0.41 - 2.26) 1.23 (0.53 - 2.38)

≥ 20.69 2.65 (1.61 - 4.36) 1.89 (0.88 - 4.04) 2.47 (1.30 - 4.69)Floor dust loading (µg/ft2)      

< 0.70 1 1 1

0.70 - 1.62 2.20 (1.28 - 3.78) 1.60 (0.74 - 3.44) 1.95 (0.99 - 3.85)≥ 1.62 2.26 (1.31 - 3.88) 1.60 (0.73 - 3.51) 1.92 (0.96 - 3.84)

Paint lead c      

Reference level 1 1 1

Level-1 exposure 1.14 (0.68 - 1.97) 0.93 (0.42 - 2.08) 1.08 (0.56 - 2.01)Level-2 exposure 3.23 (1.78 - 5.84) 4.48 (1.73 – 11.60) 3.04 (1.45 - 6.37)

-2Log(Likelihood) 517.95 352.50  

G mt

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS & PERSPECTIVES

33

In the absence of cohort studies or Bone lead measures, cumulative lead exposure index is more appropriate than cross-sectional measures to estimate the association between lead exposure and Blood lead

In young children (aged under 6) included in this study:

If all other factors are kept stables, then for each additional increase in water cumulative exposure, BLL is expected to increase 1.08 µg/dl (p=0.0002).

If all other factors are kept stables, then for each additional increase in cumulative exposure of windowsill dust loading, BLL is expected to increase 1.00 µg/dl (p=0.0058).

G mt

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS & PERSPECTIVES

34

In the absence of cohort studies or Bone lead measures, cumulative lead exposure index is more appropriate than croiss-sectional measures to estimate the association between lead exposure and Blood lead

In young children (aged under 6) included in this study:

The mean BLL in children exposed to paint chips with high lead level (paint chips ≥ 5000 mg/kg) is 1.32 µg/dl higher than mean BLL observed in those living in homes without paint chips and low lead level in painted surface (XRF < 1 mg/cm2)

G mt

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS & PERSPECTIVES

35

In the absence of cohort studies or Bone lead measures, cumulative lead exposure index is more appropriate than croiss-sectional measures to estimate the association between lead exposure and Blood lead

In young children (aged under 6) included in this study:

1 unit of change in Cumulative floor dust loading does not markedly influence the BLL…However, floor dust loading is significantly associated with the odds of BLL ≥ 2 µg/dl from 8.12 µg/kg of body weight

G mt

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS & PERSPECTIVES

36

In the absence of cohort studies or Bone lead measures, cumulative lead exposure index is more appropriate than croiss-sectional measures to estimate the association between lead exposure and Blood lead

In young children (aged under 6) included in this study:

The odds of BLL ≥ 2 µg/dl is 4 times higher in children with water cumulative lead ≥ 2.92 µg/kg of bw when compared with those with water cumulative lead < 1.12 µg/kg of bw In the children under the present study, water lead is the first largest contributor to BLL, followed by paint-lead and windowsill dust loading.

G mt

Gerard.Ngueta@crchuq.ulaval.ca

INTRODUCTIONOBJECTIVES

METHODOLOGYRESULTS

CONCLUSIONS & PERSPECTIVES

37

Future challenges :

Compare results obtained from cross-sectional measures of exposure with those reported in previous studies

Assess the departure from odds-ratio multiplicativity (and additivity) that may be due to nutritional factors , children’s characteristics and/or guardian’s socioeconomic position.

Translate cumulative lead exposure index in terms of public health languages (recommendations)

G mt

Gerard.Ngueta@crchuq.ulaval.ca38

Acknowledgment:

Canadian Water Network

Ministry of Environment and sustainable development (MDDEP)

Canada Health

G mt

Gerard.Ngueta@crchuq.ulaval.ca39

Research Team:Patrick LevalloisJulie St-LaurentDenis GauvinMarilène CourteauMichèle PrévostSchokoufeh NourFrance LemieuxMonique D’ArmourPat Rasmussen

Celine CampagnaElise DeshommesSuzanne GingrasAlain LeBlancAnnick Trudelle

Gerard.Ngueta@crchuq.ulaval.ca

THANK YOU !

G mt

Gerard.Ngueta@crchuq.ulaval.ca41

In the database we received, Paint lead was a categorical variable :

Reference level XRF < 1 mg/cm2

Level-1 exposure

XRF ≥ 1 mg/cm2 OR Paint chips < 5000 mg/kg

Level-2 exposure

Paint chip ≥ 5000 mg/kg

G mt

Gerard.Ngueta@crchuq.ulaval.ca42

Descriptive analysis of Environmental Exposure

Median (p25 – p75)

Water Lead (µg/dL)

WLLF5 1.24 (0.26 – 2.68)

WLLS1 2.33 (0.76 – 4.25)

WLLS2 2.24 (0.62 – 4.05)

WLLS3 1.99 (0.46 – 4.49)

WLLS4 1.90 (0.41 – 4.83)

Dust Lead (µg/ft2)

Floor dust 0.70 (0.35 – 1.62)

Windowsill dust 7.15 (2.56 – 20.70)

top related