1. Structural Variables Leading to Childhood Lead Poisoning, Teen Pregnancy, and Tobacco Use
Robert H. Keefe, PhD, ACSW
Sandra D. Lane, PhD, MPH
Robert A. Rubinstein, PhD, MSPH
Noah J. Webster, MA
Brooke A. Levandowski, MPH

2. Purpose of the Study
To investigate the persistent relationships between childhood lead exposure and:
Repeat teen pregnancy
Tobacco use

3. Hypotheses
Among females, childhood lead poisoning is associated with repeat teen pregnancy.
Among females, childhood lead poisoning is associated with cigarette smoking.

4. Suffer from additional health problems Have repeat pregnancies
In our prior research we found that teens who give birth are likely to:
Suffer from additional health problems
Have repeat pregnancies
Live in impoverished neighborhoods
Live in poorly maintained housing, increasing the risk of lead exposure
Have limited access to health care
Lane et al, Lane et al, in press Journal of Urban Health. Lane et al, in press, Health and Place.

5. Lead Poisoning in Syracuse
2nd highest prevalence of elevated blood level (EBL) in NY in
5 zip codes in Syracuse had
76% of the total number of lead poisoning cases
7.7% of the entire incidence of EBL in NY.
NYSDOH, 2003.

6. Highest Risk Exposures
Old lead-based paint in dilapidated buildings
Around windowsills
In the soil around buildings
In lead-tainted water
Around windowsills
Powdery dust emitted from opening and closing the windows
In the soil
Where paint chips accumulate
In lead-tainted water
From corroded lead solder in old copper pipes.
NYSDOH, Schilling & Bain, 1988.

7. Effects of Lead Exposure
Decreased
intelligence levels
Increased
neurological impairments
premature sexual maturation
impulsive or aggressive behavior
delinquency
Canfield et al, 2003a, 2003b. Espy, Needleman et al, Bellinger, 2001, 2003, 2004.

8. Neurotoxicity from lead poisoning affects the ability to:
Plan
Learn from prior experience
Control impulsive behavior
Use executive functioning skills
Canfield et al, 2003a, 2003b. Espy, Needleman et al, Bellinger, 2001, 2003, 2004

9. Lead Levels and Birth rates by Race/Ethnicity in Syracuse 15-19 year olds
(per mcb/dl)
Birth rates
(per 100,000)
White
7.38
32.9
African American
11.35
91.7
Hispanic
9.83
N/A

10. Syracuse Healthy Start Database
year olds
> 75% of all mothers < 20 years old
Routine screening included:
Childhood lead levels
Tobacco use
EBL reported to woman’s obstetrician
Note: The teens were born after the phase out of leaded gasoline and the ban on lead-based residential paint.
1998 to first quarter of 2002

11. Dichotomous measures Maternal race: White, African American
Childhood BLL: <20, >20mcg/dl
Tobacco use: none, any
Previous pregnancy: first, repeat pregnancy
Maternal age: , years
The pregnant teens were born between 1981 – 1989.

12. Data Analysis Univariate and bivariate associations
Multivariate logistic regression
Controlling for maternal race and Medicaid use
Descriptive statistics were run on teens in the SHS data based regarding
Childhood lead level
Maternal age
Medicaid insurance
Number of pregnancies
Tobacco use
Separate bivariate associations were assessed between all these same variables where “childhood lead level was the exposure variable.
Multivariate logistic regression was conducted using
Childhood lead level as the exposure
Repeat teen pregnancy or tobacco use as the outcomes
Maternal race and Medicaid use as confounders

13. Sample: 536 pregnant teens
64.6% were African American
47.0 were between years old
76.1% were on Medicaid
39.6% second or higher order pregnancy
37.5% were smokers
71.6% had childhood BLL less than 20mcg/dl
28.4% had childhood BLL greater than 20mcg/dl
13.8% had BLL between 20mcg/dl and 29/mcg/dl
9.0% had BLL between 30-39mcg/dl
5.6% had BLL between 4099mcg/dl
64.6% were African American
47.0 were between years old
76.1% were on Medicaid,
39.6% were experiencing their second or higher order pregnancy
37.5% were smokers.

14. Sample: 536 pregnant teens
71.6% had childhood BLL less than 20mcg/dl
28.4% had childhood BLL greater than 20mcg/dl
13.8% had BLL between 20mcg/dl and 29/mcg/dl
9.0% had BLL between 30-39mcg/dl
5.6% had BLL between 4099mcg/dl
64.6% were African American
47.0 were between years old
76.1% were on Medicaid,
39.6% were experiencing their second or higher order pregnancy
37.5% were smokers.

15. Sample: 536 pregnant teens
71.6% had childhood BLL less than 20mcg/dl
28.4% had childhood BLL greater than 20mcg/dl
13.8% had BLL between 20mcg/dl and 29/mcg/dl
9.0% had BLL between 30-39mcg/dl
5.6% had BLL between 4099mcg/dl
64.6% were African American
47.0 were between years old
76.1% were on Medicaid,
39.6% were experiencing their second or higher order pregnancy
37.5% were smokers.
The CDC considers 10 mcg/dl dangerous!

16. Baseline characteristics of pregnant teens by childhood BLL
% of the pregnant teens by number of pregnancies, smoking, maternal age, Medicaid vs. private insurance, and maternal race when dichotomized by BLL. Repeat pregnancy was associated with elevated childhood lead and mother’s age (older teens were more likely to have a repeat birth) (p < .05). Tobacco use was also significantly associated with maternal race (white teens smoked more than African American teens) (p < .05).

17. Baseline characteristics of pregnant teens by childhood BLL
Those with BLLs < 20 mcg/dl were pretty comparable in race, age, pregnancy, smoking and insurace

18. Baseline characteristics of pregnant teens by childhood BLL
So where the women with BLLs >=20 mcg/dl, so these two groups are comparable within the categories of BLL

19. Baseline characteristics of pregnant teens by childhood BLL
Repeat pregnancy was associated with elevated childhood lead and mother’s age (p < .05)
Tobacco use was significantly associated with maternal race (p < .05)
Repeat pregnancy was associated with elevated childhood lead and mother’s age (older teens were more likely to have a repeat birth) (p < .05). Tobacco use was also significantly associated with maternal race (white teens smoked more than African American teens) (p < .05).

20. Logistic regression results
One exposure: childhood BLL
Two outcomes:
repeat teen pregnancy
Controlled for race, age and insurance type
Tobacco use
Stratified by insurance type
Controlled for race & age

21. Outcome: repeat pregnancy
Predictor variables
Childhood lead level
(20+ mcg/dl vs. 0-19mcg/dl,)
1.59
(95% CI 1.04, 2.43)
Mother’s Race
(African American vs. white)
1.46
(95% CI 1.25, 1.71)
Mother’s Age
(18-19 vs years)
1.45
(95% CI 0.95, 2.21)
Medicaid
(Medicaid vs. Private insurance)
1.70
(95%CI 1.06, 2.73)
We performed two separate logistic regression analyses with childhood lead level as the exposure, or risk factor, as presented in Table 2. In the first analysis repeat birth was the outcome variable (Model 1) and in the second tobacco use was the outcome variable (Model 2). In each logistic regression, we controlled for confounders of maternal race and mother’s age. Receipt of Medicaid was a confounder of Model 1 and an effect modifier of Model 2, resulting in stratified effect estimates. The analyses found childhood lead poisoning to be a risk factor for subsequent repeat birth among teens and for tobacco use. Among those not covered by Medicaid, the odds of having an elevated childhood lead level among smokers was 4.25 (95%CI 1.89, 9.57) times the odds of having an elevated childhood blood lead among non-smokers, controlling for maternal race and age. Among those receiving Medicaid, the odds of elevated childhood lead level among smokers was 1.26 (95%CI 0.79, 2.03) times the odds of elevated childhood blood lead among non-smokers, controlling for maternal race and age. Controlling for maternal age, maternal race, and Medicaid use, the odds of elevated childhood lead level among those women having their second pregnancy was 1.59 (95%CI 1.04, 2.43) times the odds of elevated childhood blood lead among those having their first child.

22. Outcome: repeat pregnancy
Predictor variables
Childhood lead level
(20+ mcg/dl vs. 0-19mcg/dl,)
1.59
(95% CI 1.04, 2.43)
Mother’s Race
(African American vs. white)
1.46
(95% CI 1.25, 1.71)
Mother’s Age
(18-19 vs years)
1.45
(95% CI 0.95, 2.21)
Medicaid
(Medicaid vs. Private insurance)
1.70
(95%CI 1.06, 2.73)
We performed two separate logistic regression analyses with childhood lead level as the exposure, or risk factor, as presented in Table 2. In the first analysis repeat birth was the outcome variable (Model 1) and in the second tobacco use was the outcome variable (Model 2). In each logistic regression, we controlled for confounders of maternal race and mother’s age. Receipt of Medicaid was a confounder of Model 1 and an effect modifier of Model 2, resulting in stratified effect estimates. The analyses found childhood lead poisoning to be a risk factor for subsequent repeat birth among teens and for tobacco use. Among those not covered by Medicaid, the odds of having an elevated childhood lead level among smokers was 4.25 (95%CI 1.89, 9.57) times the odds of having an elevated childhood blood lead among non-smokers, controlling for maternal race and age. Among those receiving Medicaid, the odds of elevated childhood lead level among smokers was 1.26 (95%CI 0.79, 2.03) times the odds of elevated childhood blood lead among non-smokers, controlling for maternal race and age. Controlling for maternal age, maternal race, and Medicaid use, the odds of elevated childhood lead level among those women having their second pregnancy was 1.59 (95%CI 1.04, 2.43) times the odds of elevated childhood blood lead among those having their first child.

23. Outcome: tobacco use Predictor variables Medicaid non-users
Medicaid users
Childhood lead level
(20+ mcg/dl vs. 0-19mcg/dl)
4.25
(95% CI 1.89, 9.57)
1.25
(95% CI 0.78,2.00)
Mother’s Race
(African American vs. white)
0.32
(95% CI 0.15, 0.72)
0.29
(95% CI 0.19, 0.45)
Mother’s Age
(18-19 vs years)
0.93
(95% CI 0.42, 2.05)
1.35
(95% CI 0.88, 2.05)
We performed two separate logistic regression analyses with childhood lead level as the exposure, or risk factor, as presented in Table 2. In the first analysis repeat birth was the outcome variable (Model 1) and in the second tobacco use was the outcome variable (Model 2). In each logistic regression, we controlled for confounders of maternal race and mother’s age. Receipt of Medicaid was a confounder of Model 1 and an effect modifier of Model 2, resulting in stratified effect estimates. The analyses found childhood lead poisoning to be a risk factor for subsequent repeat birth among teens and for tobacco use. Among those not covered by Medicaid, the odds of having an elevated childhood lead level among smokers was 4.25 (95%CI 1.89, 9.57) times the odds of having an elevated childhood blood lead among non-smokers, controlling for maternal race and age. Among those receiving Medicaid, the odds of elevated childhood lead level among smokers was 1.26 (95%CI 0.79, 2.03) times the odds of elevated childhood blood lead among non-smokers, controlling for maternal race and age. Controlling for maternal age, maternal race, and Medicaid use, the odds of elevated childhood lead level among those women having their second pregnancy was 1.59 (95%CI 1.04, 2.43) times the odds of elevated childhood blood lead among those having their first child.

24. Outcome: tobacco use Predictor variables Medicaid non-users
Medicaid users
Childhood lead level
(20+ mcg/dl vs. 0-19mcg/dl)
4.25
(95% CI 1.89, 9.57)
1.25
(95% CI 0.78,2.00)
Mother’s Race
(African American vs. white)
0.32
(95% CI 0.15, 0.72)
0.29
(95% CI 0.19, 0.45)
Mother’s Age
(18-19 vs years)
0.93
(95% CI 0.42, 2.05)
1.35
(95% CI 0.88, 2.05)
We performed two separate logistic regression analyses with childhood lead level as the exposure, or risk factor, as presented in Table 2. In the first analysis repeat birth was the outcome variable (Model 1) and in the second tobacco use was the outcome variable (Model 2). In each logistic regression, we controlled for confounders of maternal race and mother’s age. Receipt of Medicaid was a confounder of Model 1 and an effect modifier of Model 2, resulting in stratified effect estimates. The analyses found childhood lead poisoning to be a risk factor for subsequent repeat birth among teens and for tobacco use. Among those not covered by Medicaid, the odds of having an elevated childhood lead level among smokers was 4.25 (95%CI 1.89, 9.57) times the odds of having an elevated childhood blood lead among non-smokers, controlling for maternal race and age. Among those receiving Medicaid, the odds of elevated childhood lead level among smokers was 1.26 (95%CI 0.79, 2.03) times the odds of elevated childhood blood lead among non-smokers, controlling for maternal race and age. Controlling for maternal age, maternal race, and Medicaid use, the odds of elevated childhood lead level among those women having their second pregnancy was 1.59 (95%CI 1.04, 2.43) times the odds of elevated childhood blood lead among those having their first child.

25. Outcome: tobacco use Predictor variables Medicaid non-users
Medicaid users
Childhood lead level
(20+ mcg/dl vs. 0-19mcg/dl)
4.25
(95% CI 1.89, 9.57)
1.25
(95% CI 0.78,2.00)
Mother’s Race
(African American vs. white)
0.32
(95% CI 0.15, 0.72)
0.29
(95% CI 0.19, 0.45)
Mother’s Age
(18-19 vs years)
0.93
(95% CI 0.42, 2.05)
1.35
(95% CI 0.88, 2.05)
We performed two separate logistic regression analyses with childhood lead level as the exposure, or risk factor, as presented in Table 2. In the first analysis repeat birth was the outcome variable (Model 1) and in the second tobacco use was the outcome variable (Model 2). In each logistic regression, we controlled for confounders of maternal race and mother’s age. Receipt of Medicaid was a confounder of Model 1 and an effect modifier of Model 2, resulting in stratified effect estimates. The analyses found childhood lead poisoning to be a risk factor for subsequent repeat birth among teens and for tobacco use. Among those not covered by Medicaid, the odds of having an elevated childhood lead level among smokers was 4.25 (95%CI 1.89, 9.57) times the odds of having an elevated childhood blood lead among non-smokers, controlling for maternal race and age. Among those receiving Medicaid, the odds of elevated childhood lead level among smokers was 1.26 (95%CI 0.79, 2.03) times the odds of elevated childhood blood lead among non-smokers, controlling for maternal race and age. Controlling for maternal age, maternal race, and Medicaid use, the odds of elevated childhood lead level among those women having their second pregnancy was 1.59 (95%CI 1.04, 2.43) times the odds of elevated childhood blood lead among those having their first child.

26. Outcome: tobacco use Predictor variables Medicaid non-users
Medicaid users
Childhood lead level
(20+ mcg/dl vs. 0-19mcg/dl)
4.25
(95% CI 1.89, 9.57)
1.25
(95% CI 0.78,2.00)
Mother’s Race
(African American vs. white)
0.32
(95% CI 0.15, 0.72)
0.29
(95% CI 0.19, 0.45)
Mother’s Age
(18-19 vs years)
0.93
(95% CI 0.42, 2.05)
1.35
(95% CI 0.88, 2.05)
We performed two separate logistic regression analyses with childhood lead level as the exposure, or risk factor, as presented in Table 2. In the first analysis repeat birth was the outcome variable (Model 1) and in the second tobacco use was the outcome variable (Model 2). In each logistic regression, we controlled for confounders of maternal race and mother’s age. Receipt of Medicaid was a confounder of Model 1 and an effect modifier of Model 2, resulting in stratified effect estimates. The analyses found childhood lead poisoning to be a risk factor for subsequent repeat birth among teens and for tobacco use. Among those not covered by Medicaid, the odds of having an elevated childhood lead level among smokers was 4.25 (95%CI 1.89, 9.57) times the odds of having an elevated childhood blood lead among non-smokers, controlling for maternal race and age. Among those receiving Medicaid, the odds of elevated childhood lead level among smokers was 1.26 (95%CI 0.79, 2.03) times the odds of elevated childhood blood lead among non-smokers, controlling for maternal race and age. Controlling for maternal age, maternal race, and Medicaid use, the odds of elevated childhood lead level among those women having their second pregnancy was 1.59 (95%CI 1.04, 2.43) times the odds of elevated childhood blood lead among those having their first child.

27. Discussion: Repeat Pregnancy
Community factors
Family factors
Individual factors
Developmental delays
All may lead to residence in older housing with peeling paint
Community factors: high unemployment, community stress.
Family factors: single parents
Individual factors: race, ethnicity, school dropout, greater number of sexual partners.
Developmental delays in cognitive ability and executive functioning.
All may lead to residence in older housing with peeling paint.

28. Discussion: Addiction
Smoking and race are determinants of EBL among women of reproductive age
Early lead exposure associated with decreased ability to delay gratification and alterations in dopamine system and other neurochemicals
Early lead exposure has been associated with decreased ability to delay gratification and alterations in dopamine system and other neurochemicals that suggest that early lead exposure could increase sensitivity to tobacco addition.

29. Limitations Small sample size
Other community factors as potential confounders
Poverty
Community stress
Residence location
Family level factors
Small sample size could mask an even larger effect estimate.
Other community factors could confound the relationship between childhood lead exposure and health outcomes:
Poverty
Community stress
Residence location
Family level factors such as single-parent families
These factors are unmeasured. How would/should we attempt to measure them accurately?

30. Further research Inclusion of community-level factors
Individual vs. community level responsibility