Objectively Measured Secondhand Smoke Exposure and Mental Health in Children:  Evidence From the Scottish Health Survey FREE

Approximately 2 million children in Britain currently live in a household where they are exposed to secondhand smoke (SHS)^1- 2 and approximately 66% of children aged 3 to 11 years are exposed to SHS in the United States.³ In children, exposure to SHS is related to elevated rates of mortality and respiratory illnesses, such as lower respiratory tract infections, wheezing, middle ear infection, and asthma.^1,4 However, to date, less attention has been paid to the impact, if any, of SHS exposure on their mental health. Mental illness is increasingly being recognized as the most significant health concern for children and adolescents in developed countries, with an estimated prevalence of 8% to 23% of the child and adolescent population in European countries.⁵

There are good reasons to anticipate an influence of SHS exposure on mental health in children. First, in animal studies, tobacco smoke can induce negative mood and a depression-like state manifested in decreased sensitivity to natural reward and enhanced sensitivity to stress and anxiety-eliciting situations later in life.⁶ Second, in humans, data from the United States have shown that prenatal and postnatal tobacco exposure are associated with higher risks of conduct disorders.⁷ Nevertheless, in another study, high cotinine levels were not associated with depression in nonsmoking adolescents, but there was an inverse association between cotinine level and depression among the smokers.⁸ Given the paucity of evidence in this area, it is difficult to draw any firm conclusions from the existing data. The aim of the present study was therefore to examine the association between objectively assessed SHS exposure and mental health in a representative sample of British children.

The present study used data from the 2003 Scottish Health Survey that consisted of a nationally representative household-based sample of the general population.⁹ Seventy-seven percent of eligible households with children took part in the survey. The parents of the children (aged 4-12 years; mean [SD], 8.3 [2.5] years) gave full informed consent to participate in the study and ethical approval was obtained from the London Research Ethics Council. Trained interviewers visited households and interviewing was conducted using computer-assisted personal interviewing.

Psychological distress was assessed using the parental version of the Strengths and Difficulties Questionnaire (SDQ), which has demonstrated good reliability and validity.^10- 11 The main outcome was total difficulties score (ranging from 0-40), with a higher score representing worse mental health. This was calculated by adding the scores from the SDQ subscales of hyperactivity, emotional symptoms, conduct problems, and peer problems.

In all children, objective exposure to SHS was assessed using salivary cotinine level, which is a reliable and valid circulating biochemical marker of nicotine exposure.^12- 13 A dental roll saturated with the child's saliva was placed in a tube and later analyzed with a rapid liquid chromatography technique using a Hewlett Packard (Palo Alto, California) hp5890 gas chromatograph. The technique had a coefficient of variation less than 7%. All valid cotinine values were obtained from saliva samples with sufficient volume to undertake the assay and with no evidence of contamination. Data on self-reported smoking and exposure to SHS were only collected from children 8 years and older since self-reported measures are considered unreliable in younger children. None of these children reported smoking in the previous week. Questions about SHS exposure first inquired about whether the children felt that they were exposed to SHS and if so, where (at home, other people's homes, or public transportation). Children reporting nonsmoking status but with a salivary cotinine level more than 15 ng/mL were removed from all analyses (<1% of the sample) because such values may be indicative of recent smoking.

Interviewers measured height and weight of the children for the calculation of body mass index. Parents reported the weekly frequency of sports and active play (for at least 15 minutes) of their children and any chronic health conditions. Demographic data were also recorded including single-parent status and questions on deprivation. We used The Scottish Index of Multiple Deprivation, which was assessed from 31 indicators in 6 individual domains including parents' employment, current income, housing, health, education, and skills and training.⁹

Based on previously used cutoffs,^14- 15 salivary cotinine level was categorized into quartiles: low SHS exposure (salivary cotinine level ≤0.05 ng/mL or below the detectable limit [the referent group]), low-moderate SHS exposure (0.06-0.30 ng/mL), moderate SHS exposure (0.31-0.70 ng/mL), and high SHS exposure (0.71-14.99 ng/mL). The SDQ score was normally distributed and we therefore used general linear models to examine associations with cotinine categories. The models were adjusted for age, sex, quintile of deprivation index, single-parent status, body mass index, chronic illness, and physical activity (number of weekly episodes of sports and active play lasting at least 15 minutes). A dose-response association was examined using a test for linear trend across groups. There were no clear differences in our results between boys and girls, so the data were pooled and adjusted by sex. All analyses were conducted using SPSS (version 14; SPSS Inc, Chicago, Illinois).

The initial sample with available cotinine data consisted of 1015 children, although after excluding those with missing SDQ scores (n = 34) and further demographic information (n = 80), the analytical sample was based on 901 children (mean [SD] age, 8.3 [2.5] years). The excluded children did not differ in age (8.4 vs 8.3 years) compared with those included, although fewer excluded children were from the least deprived quintile (12.3% vs 25.0%; P = .007). The mean value for salivary cotinine level was 1.29 ng/mL (95% confidence interval [CI], 1.16 to 1.42).

The relation between children's salivary cotinine levels and their other characteristics in shown in Table 1. Children with higher cotinine levels were more likely to live in areas of greater socioeconomic deprivation, be resident in a single-parent household, and report asthma. Among the older children who provided information on self-reported SHS exposure (n = 524), cotinine values were most strongly related to exposure at home (Figure), suggesting that exposure to parental smoke is the most important contributor to SHS exposure in children.

Using existing bandings,¹⁰ an abnormally high SDQ total difficulties score (≥20) was found in 3.0% and 2.7% of children aged 4 to 8 years and 9 to 12 years, respectively. The proportion of children with borderline scores (SDQ score of 16-19) was 5.0% and 6.6% in 4- to 8-year olds and 9- to 12-year olds, respectively. These proportions are largely comparable with the normative data for British children.¹⁶ Independent predictors of SDQ total difficulties score included sex (girls had lower SDQ scores than boys, regression coefficient β = −0.78 SDQ units; 95% CI, −1.33 to −0.22 SDQ units), age (12-year-olds had lower scores than 4-year-olds, β = −1.61; 95% CI, −2.70 to −0.52), socioeconomic deprivation (more highly deprived children had higher scores, β = 3.2; 95% CI, 2.0 to 4.4), single-parent status (children of single-parent families had higher scores, β = 1.7; 95% CI, 0.9 to 2.5), medical conditions (children with existing medical conditions had higher scores, β = 1.8; 95% CI, 1.0 to 2.7), and physical activity (least physically active tertile had higher scores, β = 1.2; 95% CI, 0.4 to 2.1).

The relation between cotinine values and SDQ scores is shown in Table 2. Participants in the highest cotinine quartile (>0.70 ng/mL) had significantly higher total SDQ scores compared with those in the lowest quartile (age- and sex-adjusted mean difference = 2.8; 95% CI, 1.6 to 3.9) and there was evidence of a dose-response effect across the cotinine group (P trend = .001). After possible confounding factors were added to the multivariable model, there was some evidence of attenuation (mainly through socioeconomic factors) but the SHS exposure–SDQ association remained statistically significant (Table 2). Of the SDQ subscales, the strongest associations with cotinine levels emerged for hyperactivity and conduct disorder; again, there was some evidence of attenuation of these gradients in the multiply adjusted models.

In the subsample of older children (8- to 12-year-olds) with available data on self-reported SHS exposure, there was an association between self-reported exposure in the home and SDQ score that was similar in magnitude to that observed in the highest cotinine exposure category (Table 3).

Results from this representative sample of British children suggest an association between increased SHS exposure and worse mental health, as reflected by higher scores on the SDQ questionnaire. Although smoke-free legislation has contributed to a successful secular decline in SHS exposure among children, the evidence suggests that those children living with smokers and in disadvantaged households continue to experience substantial contact with SHS.¹ Our data (collected before smoke-free legislation) confirm that objectively assessed SHS exposure is most closely associated with reported exposure in the home, which is consistent with previous reports.¹⁷ Indeed, self-reported exposure to SHS in the home was also robustly associated with SDQ scores in the older children. Previous reports have suggested that the accuracy of parental and youth reporting of SHS exposure might vary considerably according to the reporting source and age and sex of youth respondent.^18- 19

The financial burden of passive smoking in children is estimated to be at least £9.7 million (US $15.2 million) each year in UK primary care visits and asthma costs and £13.6 million (US $21.4 million) in UK hospital admissions.¹ These figures are based entirely on physical illness since, as described, the effects of objective SHS exposure on mental health have not been well examined. Previous studies have produced conflicting evidence. Data from the National Health and Nutrition Examination Survey have shown that prenatal and postnatal tobacco exposure are associated with higher risks of conduct disorders.⁷ Nevertheless, in another study, high cotinine levels were not associated with depression in nonsmoking adolescents, but there was an inverse association between cotinine level and depression among the smokers.⁸ In a small cohort of adolescents, active smoking was associated with a higher risk of depressive episodes over 5 years' follow-up, which was partly explained by dysfunction of the hypothalamic-pituitary-adrenal axis.²⁰ Other mechanisms might involve the dopaminergic system. Smokers who are genetically predisposed to a low resting intrasynaptic dopamine level have heightened smoking-induced dopamine release,²¹ which has been associated with greater depression and anxiety²² that are often comorbid with conduct disorder in childhood.²³ Thus, this genetic predisposition may also operate in relation to SHS exposure, especially in children of smoking parents who might be more likely to possess this genetic vulnerability. Since it has been reported that more than 80% of children are aware that passive smoking is harmful,¹ an involuntary exposure to a known harmful effect might also adversely affect mental health. We did not obtain data on prenatal smoking. However, since smoking is a lifelong habit often taken up in adolescence, it is possible that young children with smoking parents were also exposed to tobacco during the prenatal period. Low birth weight and decreased in utero brain growth are 2 of multiple potential etiologic pathways proposed as mediating the effects of prenatal tobacco smoke exposure on child neurodevelopment,²⁴ which is likely to influence mental health.

The strengths of the study include the sampling of a relatively large, representative general population-based group of children, the objective measurement of SHS exposure, and the well-characterized study members, which facilitates insights into the role of potential confounding factors. The limitations of the present study should also be recognized. Given the cross-sectional nature of this study, we cannot determine the causal nature of the association between SHS exposure and mental health, although it is unlikely that reverse causation is operating. The possibility remains that the results could be explained by residual confounding. We did, however, adjust our analyses for several important confounders, including a strong measure of deprivation that was assessed from 31 indicators in 6 individual domains. The data on mental health are based on parental reports; however, the SDQ is a highly validated instrument designed for this purpose and parental or teacher reports are more accurate in predicting psychiatric disorder than self-report.²⁵ Further, although self-report measures of internalizing emotional problems may be more valid than parent reports, the SDQ has been shown to have high concordance between child and parent for all scales.²⁶ Smokers are more likely to have common mental disorders, such as depression,²⁷ so this might have biased parental reports, although we did not have information about the mental health of the parents. The children excluded from our analyses were more likely to be from more deprived homes and thus probably exposed to greater SHS and more likely to have poorer mental health,²⁸ which may have led to an underestimation of the strength of the SHS exposure–mental health association herein. Lastly, although the SDQ reliably identifies children and adolescents with psychopathology, and increasing total difficulties scores are associated with an increasing likelihood of psychiatric disorder, it is a dimensional rather than a diagnostic measure.²⁹ Nevertheless, the SDQ has been shown to have clinical relevance, since in a previous study in the community, the questionnaire identified more than 70% of children with conduct, hyperactivity, depressive, and some anxiety disorders with a specificity of 94.6% and a sensitivity of 63.3%.²⁵

In conclusion, we found an association between SHS exposure and poor mental health, as reflected by higher scores on the SDQ. This study provides additional reasons to continue to educate parents about the multiple health dangers of exposing their children to tobacco smoke, particularly within the home.

Correspondence: Mark Hamer, PhD, Department of Epidemiology and Public Health, 1-19 Torrington Pl, University College London, London WC1E 6BT, England (m.hamer@ucl.ac.uk).

Accepted for Publication: September 19, 2010.

Published Online: December 6, 2010. doi:10.1001/archpediatrics.2010.243

Author Contributions: Dr Hamer is the guarantor of this article. Study concept and design: Hamer and Stamatakis. Acquisition of data: Stamatakis. Analysis and interpretation of data: Hamer, Ford, Stamatakis, Dockray, and Batty. Drafting of the manuscript: Hamer, Stamatakis, and Batty. Critical revision of the manuscript for important intellectual content: Hamer, Ford, Stamatakis, Dockray, and Batty. Statistical analysis: Hamer, Stamatakis, and Dockray. Administrative, technical, and material support: Ford. Study supervision: Batty.

Financial Disclosure: None reported.

Funding/Support: The Scottish Health Survey is funded by the Scottish Executive. Dr Hamer is funded by the British Heart Foundation; Dr Stamatakis is funded by the National Institute for Health Research; Dr Ford is a Medical Research Council (MRC) Clinician Scientist; and Dr Batty is a Wellcome Trust Career Development Fellow (WBS U.1300.00.006.00012.01). The MRC Social and Public Health Sciences Unit receives funding from the UK MRC and the Chief Scientist Office at the Scottish Government Health Directorates. The Centre for Cognitive Ageing and Cognitive Epidemiology is supported by the Biotechnology and Biological Sciences Research Council, the Engineering and Physical Sciences Research Council, the Economic and Social Research Council, the MRC, and the University of Edinburgh as part of the cross-council Lifelong Health and Wellbeing initiative.

Role of the Sponsors: The funding organizations played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.

Disclaimer: The views expressed in this article are those of the authors and not necessarily of the funding bodies.