Exposure of children to second-hand smoke has been linked to a variety of health effects, including respiratory infections, middle ear disease, and development of asthma.1,2 Tobacco smoke contains over 4000 constituents, including heavy metals such as lead.3,4 Lead exposure in children is associated with decreased intelligence, impaired growth, anemia, and attention and behavior problems.5 Young children are particularly susceptible to lead poisoning because they absorb more lead from their environments than do adults and because their central nervous systems are still developing.5 Active tobacco smokers have higher blood lead levels than nonsmokers,6,7 but evidence for similar effects with second-hand smoke exposure is limited, especially in children.8,9
Studies that have examined the health effects of tobacco smoke exposure among children typically define exposure by parents’ reported smoke exposure or the presence of smokers in the child’s household.10-12 A limitation of these studies is that most children in the United States are exposed to tobacco smoke13; thus, children in the “unexposed” category in these studies can have exposures from nonparental sources or in places other than the home, resulting in the potential misclassification of some children. Use of the biomarker cotinine (a metabolite of nicotine and indicator of second-hand smoke exposure) can reduce misclassification, which allows more valid comparisons among exposure groups.14
Our study analyzed data from children aged 4–16 years who had blood lead levels measured in the Third National Health and Nutrition Examination Survey (NHANES III). We used serum cotinine levels as the main basis for classifying children into second-hand smoke exposure groups, comparing blood lead levels in children who had evidence of high and intermediate second-hand smoke exposure with those who had low exposure. We also compared results obtained using reported second-hand smoke exposure with results obtained using serum cotinine.
NHANES III was conducted from 1988-1994 by the National Center for Health Statistics of the Centers for Disease Control and Prevention (CDC), Atlanta, Georgia.15 NHANES III was approved by the National Center for Health Statistics’ Institutional Review Board. In this survey, a stratified multistage clustered probability design was used to select a representative sample of the civilian, noninstitutionalized U.S. population. The final sample comprised 81 geographic sites. Survey participants completed extensive questionnaires in the household, and a comprehensive physical examination was performed at a specially equipped mobile examination center. A knowledgeable proxy completed questionnaires for participants less than 17 years of age.
Subjects and Demographics
We limited the analysis to children aged 4–16 years for whom serum cotinine levels and blood lead levels were measured (cotinine levels were not obtained for children younger than 4 years old). Children were included in the analysis if data were complete on their race/ethnicity, age, sex, region of the country, poverty index, education level of the reference adult (one of the persons who owns the home or pays the rent), family size, number of rooms in the household, year of house construction, blood lead level, whether cigarettes were smoked in the home, and cotinine levels. Of the 10,084 children 4–16 years old who participated in NHANES III, 4255 did not have cotinine levels measured (usually because the sample of blood obtained was inadequate), 18 did not have blood lead measures, 156 had cotinine levels higher than 15 (indicating active smoking),13 and 48 were missing data on other variables; these exclusions resulted in 5592 subjects in the final dataset. The children who did not have cotinine measures were similar to the included participants for all covariates except age; those missing cotinine measures were overrepresented in the youngest age group (52% of 4–6 year olds compared with 21% of the 7–11 year olds and 18% of the 12–16 year olds).
Serum cotinine levels were determined using high-performance liquid chromatography atmospheric-pressure chemical ionization tandem mass spectrometry, as described elsewhere.13 The limit of detection was 0.050 ng/mL. We stratified the participants into the following tertiles of cotinine levels: ≤0.050–0.104 ng/mL, 0.105–0.562 ng/mL, and 0.563–14.9 ng/mL. We included participants with no detectable cotinine in the lowest tertile.
Blood Lead Measurements
A 1-mL sample of EDTA-anticoagulated whole blood was collected by venipuncture from participants during the physical examination. Blood samples were frozen and shipped on dry ice for analysis to the NHANES Laboratory, Division of Environmental Health Laboratory Sciences, National Center for Environmental Health, CDC. The blood samples remained frozen at −20°C until they were analyzed. Lead was measured by graphite furnace atomic absorption spectrophotometry (GFAAS) using the method of Miller et al.16 The GFAAS method included deuterium background correction and had a limit of detection of 1.0 μg/dL. The lead result is the mean of duplicate measurements. The blood lead measurements were calibrated using standards prepared from lead nitrate Standard Reference Material 928 obtained from the National Institute of Standards and Technology, Gaithersburg, Maryland.
Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Mexican-American, and Other and was determined by proxy-report on the questionnaire. Region was defined on the basis of the Bureau of the Census definitions as Northeast, Midwest, South, and West. Education level of the reference adult was classified as <12 years, 12 years, and ≥13 years. The poverty index was used to divide the sample into 4 strata, poverty index <1, poverty index 1-<2, poverty index ≥2, and poverty index unknown.15 Year of housing construction was determined by self-report and was defined as before 1946, 1946–1973, after 1973, and don’t know. Family size was categorized as 4 or less and 5 or more people. The number of rooms in a household was categorized as 5 or less and 6 or more. Reported second-hand smoke exposure was defined by the parent’s response to the question “Does anyone who lives here smoke cigarettes in the home?” We used 3 age strata in the analysis: 4–6 years, 7–11 years, and 12–16 years.
We performed statistical analyses using SAS17 and SUDAAN,18 a software package that accounts for complex sample design when calculating variance estimates. We calculated all estimates using the sampling weight to represent U.S. children 4–16 years old. The purpose of the sampling weight calculations was to adjust for unequal probabilities of selection and to account for nonresponse. The weights were poststratified to the U.S. population as estimated by the Bureau of the Census.
Geometric mean blood lead levels were calculated as the antilog of the mean of log of the blood lead levels. We modeled the relation between second-hand smoke exposure and lead levels using linear and logistic regression in both univariate and multivariate models adjusting for covariates. The natural log of the blood lead levels was used as the dependent variable in the linear regressions, and the results were exponentiated to yield a percentage increase from the values. A blood lead level of 10 μg/dL or higher, which the CDC considers elevated,19 was the categorical dependent variable in the logistic regressions. We used cotinine levels and reported second-hand smoke exposures as our determinant of tobacco smoke exposure, but we did not include both in the same model because of colinearity. Covariates included age, sex, race/ethnicity, region, education level, poverty index, age of housing, number of rooms in the household, and family size. Multivariate models were evaluated for interaction between second-hand smoke exposure and other covariates.
Our final dataset of 5592 children represents an estimated 37 million children 4–16 years old in the United States. The distribution of the covariates and second-hand smoke exposure are presented in Table 1. As expected and previously reported, factors associated with increased blood lead levels and a higher proportion of elevated blood lead levels included younger age, black race, lower education level, poverty, older housing, and residing in the Northeast or Midwest.20
Higher levels of serum cotinine were related to higher mean blood lead levels in 4–12 year olds (Fig. 1A). Although we found a similar pattern in children classified on the basis of reported second-hand smoke, the difference was less (Fig. 1B).
In the univariate regression models, factors associated with an increase in the log blood lead level included cotinine level, younger age, 12 years or fewer of education, a poverty index <2, black or Mexican-American race/ethnicity, male sex, older age of housing, 5 or fewer rooms in the home, and living in the Northeast or Midwest (Table 2). Except for of number of rooms in the home and education level, these same factors were associated with increased log lead levels in the adjusted model (Table 2). In the multivariate model the geometric mean blood lead levels of children with high cotinine levels were approximately 38% higher (95% confidence interval [CI] = 25–52%) than children with low cotinine levels. When reported second-hand smoke in the home was used to indicate smoke exposure, blood lead levels were increased in both the univariate (26%; CI = 16–36%) and multivariate (13%; CI = 8–22%) models. Models that included additional levels of poverty index, family size, and number of rooms in the home (to evaluate residual confounding) did not change this estimate (data not shown).
Overall, 2.2% of the population had blood lead levels of 10 μg/dL or higher (Table 1). As would be expected, a higher proportion of children 4–6 years old had high blood lead levels than children 12–16 years old (4.6% vs. 1.1%; adjusted odds ratio [OR] = 3.8; CI = 1.8–8.4). Children with high cotinine levels were more likely to have high blood lead levels than were those with low cotinine levels (3.9% vs. 0.6%; adjusted OR = 4.4; CI = 1.9–10.5). Among children with the lowest current cotinine levels, the proportion of children with blood lead levels of 10 μg/dL or higher increased with increasing age, from 0.2% among 4–6 year olds to 0.5% among 7–11 year olds and 0.9% among 12–16 year olds; among children with the highest cotinine levels, this proportion decreased with increasing age (7.5%, 3.6%, and 2.0%, respectively, for the same age groups).
In the age-stratified multivariate logistic regression models, the relation between high cotinine levels and blood lead levels of 10 μg/dL or higher varied by age group. This effect was strongest in the 4–6-year-old children (OR = 22.1; CI = 3.9–124) and weaker in 7–11 year olds (OR = 5.3; CI = 1.8–15.9) and 12–16 year olds (OR = 3.8; CI = 0.7–20.2; Table 3). Models that included interaction terms suggested that the effect of a high cotinine level on blood lead was stronger among children living in older homes. In multivariate models that used reported second-hand smoke exposure as the dependent variable, the null value was included in the confidence intervals for the overall group (OR = 1.1; CI = 0.6–1.9) and for all the age strata (OR = 1.7, CI = 0.9–3.3 for 4–6 year olds; OR = 1.2, CI = 0.6–2.8 for 7–11 year olds; and OR = 0.6, CI = 0.3–1.3 for 12–16 year olds).
In a separate analysis limited to 4–6 year olds, the proportion of children with blood lead levels of 10 μg/dL or higher was associated with increasing cotinine levels in almost every subgroup. This association was weaker when reported second-hand smoke exposure was used (Table 4). Similarly, geometric mean blood lead levels were higher in children with increased cotinine levels in almost every subgroup (Table 5).
Tobacco smoke exposure appears to be associated with elevated blood lead levels among children in the United States. This association was strongest in younger children and was less apparent when the exposure was measured by reported tobacco smoke exposure rather than serum cotinine level.
As many previous studies have noted, including a previous analysis of this dataset, factors such as older age of housing, black race, lower educational level, and poverty are associated with higher blood lead levels.20-22 These associations could be seen even when looking only at children with high cotinine levels (Tables 4 and 5), demonstrating that these risk factors might have an additive influence on increased blood lead levels in children.
Lead is present in both tobacco23 and tobacco smoke.24 Lead levels in tobacco have decreased with decreasing ambient air lead levels.23 Current estimates from Canada are that each cigarette contains approximately 600–800 ng of lead but during the time NHANES III was conducted might have contained as much as 1400 ng of lead.23 The recently completed Massachusetts Benchmark Study estimates that mainstream tobacco smoke contains 60 ng of lead per cigarette, and that sidestream smoke contains 5–10 ng of lead per cigarette.24 This lead is likely to be associated with the particulate fraction of tobacco smoke and absorbed through the respiratory system. A recent study detected higher mean lead levels in the indoor air of homes where smoking occurs, compared with that in homes where no smoking occurs (21.8 ng/m3 vs. 7.8 ng/m3).25 In addition, lead in the particulate fraction could settle onto surfaces and food where it has the potential to reexpose people through either ingestion or inhalation.26
Our findings might represent an exposure directly related to inhaled lead, in that the geometric mean increase in blood lead level was fairly consistent in children up to age 12 (Fig. 1A), and ingestion of lead by eating paint chips or other hand-to-mouth activities should decrease with increasing age. Another possibility, however, is that inflammation associated with second-hand smoke exposure results in increased absorption of lead from other sources in the environment. Evidence from our data supporting an indirect effect of second-hand smoke was that the observed increase was modified in some subgroups, eg, the increase in the geometric mean blood level among 4 to 6-year-old children was 1.01 μg/dL in children living in homes built after 1973 but 3.11 μg/dL in children living in homes built before 1946 (Table 5). Exposure models that have estimated blood lead levels in children have not included second-hand smoke exposure, although lead in household dust probably is elevated when second-hand smoke is routinely present, and maternal lead levels probably are associated with primary smoking and second-hand smoke exposure; both of these factors are included in these models.27 Assuming that children breathe 5 m3 of air daily, a child could potentially be exposed to 39.8 μg of lead annually through inhalation (21.8 ng/m3 × 5 m3/day × 365 days/year, using the estimate of air lead provided by Bonanno et al.25) related to second-hand smoke in the home, although actual exposures probably would be lower (children might not be in the home all day and smoking presumably does not occur in the home 24 hours a day). Current exposure models estimate that air lead levels of this magnitude would increase blood lead levels by less than 0.1 μg/dL, much less than our observed increase of approximately 1.0 μg/dL.27 This suggests that the association of second-hand smoke with blood lead goes beyond direct contamination or that the current exposure models need to be reevaluated.
Cotinine is an excellent biomarker for recent tobacco smoke exposure.14 Previous analysis of NHANES III data have shown that most children without reported tobacco smoke exposure have detectable levels of cotinine in their blood, indicating unreported exposure to tobacco smoke.28 Use of cotinine as a biomarker of second-hand smoke exposure in this analysis was a stronger predictor of increased blood lead levels than reported second-hand smoke exposure. In our multivariate logistic regression analyses, high cotinine was a strong predictor of blood lead levels of 10 μ/dL or higher, whereas reported second-hand smoke exposure was not. This is likely related to children with exposure being misclassified as unexposed.
Cotinine has a biologic half-life of up to 40 hours.14 Thus, it is not a good biomarker for second-hand smoke exposure that occurred weeks to months before the measurement and is probably less likely to reflect “typical” exposure in older children than younger children. Conversely, blood lead has an estimated half-life of 200–300 days.29,30 Past exposures to second-hand smoke in the absence of recent exposure might explain why a higher proportion of 12–16 year olds in the low cotinine group had high blood lead levels (0.9%) than 4–6 year olds in the low cotinine group (0.2%). This finding also could reflect the overall decrease in lead exposure over time.
Others have reported increased blood lead levels in second-hand smoke-exposed children, although none has based exposure on measures of cotinine.8,9 In one study, mean blood lead levels in second-hand smoke-exposed children were 1.5–2.0 μg/dL higher than those in unexposed children, with overall blood lead levels much higher than those in our study.8
Our analysis is subject to several limitations. First, cotinine levels were not obtained on children younger than 4 years old when lead exposure has the greatest impact on neurologic development.22 In addition, cotinine levels were not obtained in 40% of the sample and other subjects were missing data on other covariates, raising the possibility of selection bias. An unmeasured confounder or residual confounding related to socioeconomic status or other factors might also contribute to these findings.
These data suggest that second-hand smoke could be associated with elevated blood lead levels among U.S. children 4–16 years old. The national goal to eliminate childhood lead poisoning as a public health problem by 2010 will require improved identification and elimination of all possible sources of lead in young children. Additional studies designed to investigate the association of second-hand smoke exposure and elevated blood lead levels should include children under 4 years old.
1. Cook DG, Strachan DP. Health effects of passive smoking-10: summary of effects of parental smoking on the respiratory health of children
and implications for research. Thorax
2. Infante-Rivard C. Childhood asthma and indoor environmental risk factors. Am J Epidemiol
3. California Environmental Protection Agency. Health effects of exposure to environmental tobacco smoke. Tob Control.
4.Jenkins RA. Occurrence of selected metals in cigarette tobacco and smoke. IARC Sci Publ
5.National Research Council. Measuring Lead Exposure in Infants, Children, and Other Sensitive Populations.
Washington, DC: National Academy Press; 1993.
6. Grasmick C, Huel G, Moreau T, et al. The combined effect of tobacco and alcohol consumption on the level of lead and cadmium in blood. Sci Total Environ
7. Shaper AG, Pocock SJ, Walker M, et al. Effects of alcohol and smoking on blood lead
in middle-aged British men. BMJ (Clin Res Ed
8. Willers S, Schutz A, Attewell R, et al. Relation between lead and cadmium in blood and the involuntary smoking of children
. Scand J Work Environ Health
9. Berglund M, Lind B, Sorensen S, et al. Impact of soil and dust lead on children
’s blood lead
in contaminated areas of Sweden. Arch Environ Health
10. Gergen PJ, Fowler JA, Maurer KR, et al. The burden of environmental tobacco smoke exposure on the respiratory health of children
2 months through 5 years of age in the United States: Third National Health and Nutrition Examination Survey, 1988 to 1994. Pediatrics
11. Cunningham J, O’Connor GT, Dockery DW, et al. Environmental tobacco smoke, wheezing, and asthma in children
in 24 communities. Am J Respir Crit Care Med
12. Fielder HM, Lyons RA, Heaven M, et al. Effect of environmental tobacco smoke on peak flow variability. Arch Dis Child
13. Pirkle JL, Flegal KM, Bernert JT, et al. Exposure of the US population to environmental tobacco smoke:the Third National Health and Nutrition Examination Survey, 1988 to 1991. JAMA
14. Benowitz NL. Cotinine
as a biomarker of environmental tobacco smoke exposure. Epidemiol Rev
15. Plan and operation of the Third National Health and Nutrition Examination Survey, 1988-94. Series 1: programs and collection procedures. Vital Health Stat 1.
16. Miller DT, Paschal DC, Gunter EW, et al. Determination of lead in blood using electrothermal atomisation atomic absorption spectrometry with a L’vov platform and matrix modifier. Analyst
17.SAS Institute I. SAS Language: Reference
, version 6. Cary, NC: SAS Institute, Inc; 1990.
18.Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual
, release 7.5. Research Triangle Park, NC: Research Triangle Institute; 1997.
19.Centers for Disease Control and Prevention. Preventing Lead Poisoning in Young Children: A Statement by the Centers for Disease Control and Prevention-October, 1991
. Atlanta, GA: US Department of Health and Human Services, Public Health Service; 1991.
20. Pirkle JL, Kaufmann RB, Brody DJ, et al. Exposure of the U.S. population to lead, 1991-1994. Environ Health Perspect
21. Lanphear BP, Byrd RS, Auinger P, et al. Community characteristics associated with elevated blood lead
levels in children
22. Kaufmann RB, Clouse TL, Olson DR, et al. Elevated blood lead
levels and blood lead
screening among US children
aged one to five years: 1988-1994. Pediatrics
23. Rickert WS, Kaiserman MJ. Levels of lead, cadmium and mercury in Canadian cigarette tobacco as indicators of environmental change:results from a 21 year study (1968-1988). Environ Sci Technol
24.Connally G. 1999 Massachusetts Benchmark Study
. Boston: Massachusetts Tobacco Control Program; 2000.
25. Bonanno LJ, Freeman NCG, Greenberg M, et al. Multivariate analysis on levels of selected metals, particulate matter, VOC, and household characteristics and activities from the Midwestern States NHEXAS. Appl Occup Environ Hyg
26. Lanphear BP, Weitzman M, Winter NL, et al. Lead-contaminated house dust and urban children
’s blood lead
levels. Am J Public Health
27.United States Environmental Protection Agency. Reference Manual: Documentation of Updates for the Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK)
, Windows version. Washington, DC: US EPA; 2001.
28. Mannino DM, Caraballo RS, Benowitz N, et al. Predictors of cotinine
levels in US children
. Data from the Third National Health and Nutrition Examination Survey. Chest
29. Delves HT, Sherlock JC, Quinn MJ. Temporal stability of blood lead
concentrations in adults exposed only to environmental lead. Hum Toxicol
30. Hryhorczuk DO, Rabinowitz MB, Hessl SM, et al. Elimination kinetics of blood lead
in workers with chronic lead intoxication. Am J Indust Med
Keywords:© 2003 Lippincott Williams & Wilkins, Inc.
tobacco smoke pollution; children; blood lead; cotinine