SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References

Maternal smoking during pregnancy has been associated with overweight and obesity in childhood and is strongly correlated with children's tobacco smoke exposure before and after pregnancy. We investigated the independent association of tobacco smoke exposure at various pre- and postnatal periods and overweight at age 6. A total of 1,954 children attending the 2001–2002 school entrance health examination in the city of Aachen, Germany, were included into this study. Height and weight were measured, BMI was calculated. Tobacco smoke exposure at various periods, other lifestyle and sociodemographic factors were ascertained by questionnaire. Multiple logistic regression models were used to assess the association between tobacco smoke exposure and overweight. Prevalence of overweight was 8.9%. Significant positive associations were found with maternal smoking before and during pregnancy and during the first and sixth year of life. When all smoking periods were included into one logistic model simultaneously, secondhand smoke exposure after birth remained positively associated with overweight at age 6 at either one of the two time periods (first year only: odds ratio (OR) (95% confidence interval (CI)): 2.94 (1.30–6.67), sixth year only: 2.57 (1.64–4.04), respectively) or at both (4.43 (2.24–8.76)). Exposure to tobacco smoke during the first years of life appears to be a key risk factor for development of childhood overweight.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References

It is well established that smoking in pregnancy results in intrauterine growth restriction and consequently in small for gestational age newborns (1). In recent years, maternal smoking during pregnancy has also been associated with overweight and obesity in childhood in a rapidly increasing number of epidemiological studies from various countries (2,3,4,5,6,7,8,9,10,11,12,13,14,15,16). Most of them were summarized in a recently published meta-analysis (17) which basically confirmed the positive association between maternal smoking during pregnancy and childhood overweight.

Complex mechanisms and biomolecular pathways are discussed to explain these patterns: thrifty phenotype theory (18), postnatal catch-up growth (19), neurotransmitter or endocrine imbalances (20), and behavioral differences between smoking and nonsmoking mothers, like breastfeeding frequency (21,22), physical activity and nutritional patterns (23). However, because maternal smoking occurs in combination with various social and behavioral factors that possibly influence child growth and development during fetal life, infancy and childhood, a complex and explanatory model of causal factors which might act at different periods and through different pathways still has to be established, while new factors are still identified (24).

Little attention has so far been given to the fact that exposure to tobacco smoke during pregnancy highly correlates with tobacco smoke exposure before pregnancy, after birth, and during childhood which may also have an impact on childhood overweight (3,12,25,26). The objective of our study therefore was to describe the independent association between tobacco smoke exposure at various pre- and postnatal periods and children's overweight at age 6.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References

Study design and study population

We used data from a cross-sectional study conducted in Aachen, Germany, in cooperation with the Aachen Public Health Service between December 2001 and July 2002. Aachen is a city with about 260,000 inhabitants and is located in the very West of Germany close to the border to the Netherlands and Belgium. All children who attended the 2001–2002 obligatory health examination before school entrance in the city of Aachen, were born between 1 July 1995 and 30 June 1996 and who were accompanied by a parent with sufficient knowledge of German were enrolled in this investigation. Participation was voluntary and the study was approved by the ethics board of the University of Aachen. Written informed consent was obtained in each case. The survey was completed by 1,647 mothers (84.4%), 270 fathers (13.8%) and 35 other relatives (1.8%).

The children's anthropometric data including height and weight were measured according to a standardized protocol as part of the routine examination conducted at the prospective primary school of the children. Body weight in underwear was measured to the nearest 0.1 kg and height to the nearest 0.1 cm using a digital scale (Seca column scale 910 with telescopic measuring rod; Seca, Hamburg, Germany). During the examination the parents were asked to complete a self-administered standardized 42-item questionnaire, which included detailed questions on sociodemographic and lifestyle factors of the parents and children (including parental body weight and height, parental smoking habits).

Outcome measurement

BMI was calculated as weight (kg)/height (m²). Because BMI changes substantially with age and developmental stages during childhood, BMI percentiles rather than fixed cut points were used to define overweight and obesity (27). The age- and sex-specific reference values of the German population (28) were used as reference for being overweight (defined as BMI ≥90th percentile) or obese (defined as BMI ≥97th percentile).

Measurement of exposure

The accompanying parent was asked if the mother smoked before pregnancy, during pregnancy, and during the 1st year after pregnancy (yes or no). If maternal smoking was indicated at any of these periods, parents were further asked about the average number of cigarettes smoked per day. Additionally, parents were asked if the child was exposed to secondhand smoke at home at the time of the school entrance examination (at age 6) and if so, how many cigarettes were smoked per day in the household in total (by all smoking persons). To operationalize exposure to smoking after birth taking into account the different time periods (maternal smoking during 1st year after pregnancy and secondhand smoke at home at age 6), a combined variable was constructed (no exposure at both time periods, exposure during 1st year only, exposure at age 6 only, exposure at both time periods).

Measurement of confounding factors

The following factors were considered as potential confounding factors and therefore included in the analyses.

Birth and infancy. Birth weight in grams (<2,500 g, ≥2,500 g), breast feeding (yes/no), parity (0, 1, >1).

Children's current lifestyle factors. Watching TV (≤1 h/day, >1 h/day), sports (“How often is your child physically active (playing outdoors or visiting a sport club?”): ≤1 time/week, >1 time/week), fast food consumption (<1 time/week, ≥1 time/week).

Parental factors. Highest parental education (<10 years, 10 years, >10 years, when information was only known from one parent, this educational level was taken), maternal BMI (mother obese, BMI ≥30 kg/m2).

Statistical analysis

Mothers and children were characterized according to maternal smoking status in the year before pregnancy, during pregnancy, in the first year after birth and secondhand smoke at age 6. Group differences to children who were not exposed to tobacco smoke at any respective time period were analyzed using Student's t-test for continuous variables and χ2 test for categorical variables.

Logistic regression models with overweight at age 6 as dependent variable were used to assess the association with tobacco smoke exposure at the described time periods. Several models were calculated. First, separate models were calculated for each period of exposure to tobacco smoke, while controlling for birth and infancy factors, children's current lifestyle factors and parental factors described above as potential confounding variables. Second, one model was calculated in which all tobacco smoke exposure variables were simultaneously included into the model, again controlling for additional potential confounding variables.

Multiple imputation was employed to impute missing covariate values for the logistic regression models. Variables considered in the imputation model included parity, birth weight, breastfeeding, watching TV, sports, visits to fast food restaurants, highest parental education, maternal BMI, maternal age. An α-level of 0.05 was chosen as criterion for statistical significance. Multicollinearity was assessed by calculating variance inflation factors. Variance inflation factors values above 2.5 were considered indicative for multicollinearity (29).

All analyses were carried out using the SAS software package, version 9.2 (SAS Institute, Cary, NC). The procedures PROC LOGISTIC and PROC MI were used for multiple logistic regression and multiple imputation, respectively.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References

Of the 2,020 children who fulfilled the inclusion criteria, 1,979 parents (98.0%) agreed to participate in the study. A complete maternal smoking history regarding tobacco smoke exposure before, during, and after pregnancy was available for 1,954 children. Information on current secondhand smoke exposure was missing for two of these children and BMI was missing for one child. Mean age of the children at the time of investigation was 5.8 years. One hundred seventy five children (8.9%) were overweight (BMI ≥90th percentile). Of those overweight, 83 children, 4.2% of the total study population, were obese (BMI ≥97th percentile). One thousand six hundred and thirty one children (83.9%) were of German nationality.

Figure 1 illustrates exposure to tobacco smoke during the various time periods: 681 mothers (34.9%) reported smoking before pregnancy, 322 (16.4%) during pregnancy, and 467 (23.9%) were smoking in the first year after birth. At the time of the school entrance examination, one-third of the children (n = 652, 33.4%) were exposed to secondhand smoke. Among all but nine (2.7%) of the children exposed to prenatal smoking, the mother had also smoked before pregnancy. Furthermore, of those children with intrauterine tobacco exposure, 92.1% (n = 297) were also exposed to maternal tobacco smoke during their first year of life. Two hundred and sixty children (80.7%) exposed to tobacco smoke during those two periods were also exposed to tobacco smoke at home at the time of the health examination. No tobacco exposure at any time was documented for 1,047 children.

image

Figure 1. Exposure to tobacco smoke before, during, and after pregnancy.

Download figure to PowerPoint

Children not exposed to tobacco smoke at any time had the highest birth weight, were more often breast fed, were watching TV less often, were more physically active and consumed fast food less frequently than children who were exposed to tobacco smoke. Their parents also were highly educated, less frequently obese and their mothers were older (Table 1).

Table 1.  Distribution of birth, infancy, current child and parental factors according to maternal smoking before, during, after pregnancy, secondhand smoke exposure at age 6
inline image

Associations of birth and infancy, current lifestyle factors, and parental factors with overweight are shown in Table 2. Breast feeding was significantly associated with a reduced risk of overweight at age 6, whereas watching TV ≥1 h/day, sports ≤1 time/week and fast food consumption ≥1 time/week were associated with an increased risk for overweight at age 6. It was less likely for children of higher educated parents to be overweight but more likely when the mother was obese.

Table 2.  Association of birth and infancy, children's lifestyle factors and parental factors with overweight at age 6
inline image

The association between exposure to tobacco smoke at the various time periods and overweight at the school entrance examination is shown in Table 3. In a first step, all time periods were considered separately. In crude analyses, significant positive associations were found for all periods of tobacco smoke exposure. After adjusting for confounding factors, a significantly increased risk for children being overweight remained for children of mothers who smoked in the 1st year after pregnancy only (odds ratio (OR): 2.08 (95% confidence interval (CI): 1.02–4.24)), for children who were exposed to secondhand smoke at age 6 only (OR: 2.38 (95% CI: 1.53–3.71)) and for children who were exposed at both time periods after birth (OR: 2.90 (95% CI: 1.86–4.54)).

Table 3.  Association between tobacco smoke exposure before, during or after pregnancy and at age 6 and overweight at the age 6
inline image

When all four exposure periods were included into one logistic model simultaneously to estimate the independent exposure effects for of each time period (see Table 3), secondhand smoke exposure after birth remained positively associated with overweight at age 6: children whose mother smoked only 1 year after birth had a almost threefold risk to be overweight at age 6 (OR: 2.94 (95% CI: 1.30–6.67)), children who were exposed to secondhand smoke at age 6 only had a 2.5 times higher risk (OR: 2.57 (95%.CI: 1.64–4.04)) and children who were exposed at both times had the highest risk to be overweight at age 6 (OR: 4.43 (95%.CI: 2.24–8.76)) compared to children who were not exposed.

Variance inflation factors for all smoking variables at the various exposure periods did not exceed 2.5 when included into the model simultaneously; hence there is no indication of relevant multicollinearity.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References

Our analyses of the association between tobacco smoke exposure before, during and after pregnancy and childhood overweight at the age of 6 years indicated that maternal smoking in the first year of life as well as secondhand smoke exposure at age 6 are independently associated with childhood overweight. In multivariate regression, children whose mothers smoked in their first year of life and/or at age 6 had a significantly higher risk to be overweight at school entry compared to children whose mothers did not smoke or were not exposed at the respective time period. By contrast, the positive associations with maternal smoking before and during pregnancy seen in bivariate analyses disappeared after the control for tobacco smoke exposure during the other periods.

There is a growing number of studies reporting maternal smoking during pregnancy as a risk factor for childhood overweight and obesity, including several studies from Germany (11,12,16) which are based on comparable study populations to ours, i.e., children enrolled at the compulsory school entry health examinations. Additionally, there are studies from other European countries (4,7,9,13,15), the United States (5,8), Canada (6,30), Japan (14,31), and Australia (2). Although there were some differences regarding methodological aspects (data reported retro- or prospectively, definition of exposure/outcome, type and number of potential confounders adjusted for), all studies and a meta-analysis (17) have come to the conclusion that children whose mothers smoked during pregnancy have an approximately twofold increased risk for childhood overweight and obesity. The results of our study obtained without control for tobacco smoke exposure during other periods are consistent with these findings.

In accordance with our study, there is increasing evidence that not only maternal smoking during pregnancy is a risk factor for childhood obesity, but also tobacco smoke exposure at other periods increases the risk for childhood obesity: Leary et al. (32) included partner's smoking during pregnancy and found an association in the same direction than that of mother's although weaker. In a study by von Kries et al. (25), when additional adjustment for paternal smoking at the time of interview (age 6 years) was done, the odds ratio for smoking before and in pregnancy decreased from 1.7 to 1.3 for overweight, but remained still significant. Kleiser et al. (12) also took smoking during pregnancy and parental smoking at the time of the interview into account. The ORs decreased substantially when both variables were taken into the model simultaneously, although the influence of smoking during pregnancy remained statistically significant whereas current smoking did not. In a study by Apfelbacher et al. (3) smoking during pregnancy remained marginally associated with overweight at age 6 (OR (95% CI): 1.11 (1.00–1.23)), but not with obesity (OR (95% CI): 1.13 (0.95–1.36)), whereas “smoking in the living place” at the time of interview remained significantly associated with both, overweight and obesity in a model where both variables among other confounders were simultaneously included into the analysis. In a recently published analysis from the UK Millennium Cohort Study (26), maternal smoking during pregnancy (OR (95% CI): 1.32 (1.19–1.47)) and smoking in the same room as the child at age 3 (OR (95% CI): 1.43 (1.28–1.60)) were associated with rapid weight gain between ages 3 and 5 years and remained statistically significant when simultaneously included in the model and adjusted for other factors as well (OR (95% CI): 1.23 (1.09–1.38) and 1.32 (1.16–1.49), respectively).

Therefore, it seems crucial, to comprehensively control for tobacco smoke exposure during all relevant time periods when examining the influence of smoke exposure on childhood obesity to delineate exposure effects during various periods of time.

Although multicollinearity could be ruled out statistically, maternal smoking status was not independent at the various time points as 92% of children who were exposed to tobacco smoke during pregnancy were also exposed to maternal smoking during their first year of life, and 80% of children exposed to tobacco smoke during and after pregnancy were also exposed to household environmental smoking at age 6. Nevertheless, our multivariate analyses indicate that the smoking exposure during childhood rather than maternal smoking during pregnancy may be the key risk factor for childhood overweight and probably acts as a surrogate marker for certain environmental factors increasing the risk of overweight and obesity for the exposed child.

The detrimental health effect of secondhand smoke exposure during the first several years of life on various medical conditions like respiratory problems (infections, asthma), sudden infant death syndrome, behavioral and neuro-cognitive deficits (33) are known. In addition, Weitzman et al. (34) found an association of secondhand smoke exposure with metabolic syndrome in adolescents (12–19 years of age) and Houston et al. (35) reported an increased risk of developing glucose intolerance for passive smoking among young adults. Our results appear to be consistent with these observations, given the well known association between overweight and metabolic disorders.

In the interpretation of our results, a number of limitations have to be kept in mind. Maternal smoking status before, during, and after pregnancy was ascertained retrospectively using a self-administered questionnaire. There was no chance to validate reported smoking during pregnancy using objective measures. However, mothers appear to remember pregnancy-related issues including smoking rather well (36) and despite social demands to refrain from smoking during pregnancy, self-reported smoking habits during pregnancy seem fairly valid (37,38). In our study, the proportion of mothers who admitted having smoked during pregnancy (16.4%) was higher compared with figures from another study conducted in Germany in 1999–2000 (11.2%) (25). Nevertheless, some mothers may have falsely denied smoking during pregnancy or relapsing after pregnancy. Due to a higher number of missing values for number of cigarettes smoked per day, only the dichotomous variable regarding smoking at the different time periods (yes/no) was used. Our results are based on a study from a single city (city of Aachen) in Western Germany. However, almost complete coverage was achieved of all eligible children in the city, and the prevalence of overweight and obesity in our population was very similar as reported in the German National Health Interview and Examination Survey for Children and Adolescents for the respective age group (39).

Although we carefully controlled for known potential confounding factors in our analysis, the potential for residual confounding due to imperfect measurement of covariates or additional factors not controlled for cannot be ruled out, as maternal smoking and exposure to secondhand smoke are likely to be related to a broad range of social and behavioral factors that might be related to overweight (24). It appears very unlikely though that such residual confounding could completely explain the very distinct patterns of associations of children's overweight and tobacco smoke exposure during the various periods of time.

Sharma et al. (40) found effect modification by maternal race/ethnicity in a low-income US cohort for the association between smoking during pregnancy and childhood obesity which might reflect genetic polymorphisms, variations in enzyme activity, differences in tobacco products used or preferences for certain types of cigarettes by race/ethnicity or smoking as a marker for obesogenic home environments. Since the fast majority of our study participants were German, effect modification by race/ethnicity could not be further explored.

In conclusion, tobacco smoke is one of the most ubiquitous environmental health hazards not only for adults but also likewise for children (33). The observation that tobacco smoke exposure in early childhood is associated with an increased risk for childhood overweight, is but one albeit very important addendum to the already long list of harmful health effects of tobacco smoke. The prevalence of overweight and obesity in children and adolescents is increasing worldwide at an alarming rate and thus poses as a public health problem which requires immediate interventions (39,41). While the importance of reducing maternal smoking during pregnancy for reasons other than childhood overweight is out of question, our results emphasize the importance to extend public health programs for preventing childhood overweight by including the reduction of tobacco smoke exposure during the early years of life, beyond focusing on nutritional and physical activity. Such efforts would be to the benefits of both children and their parents.

ACKNOWLEDGEMENT

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References

This work was supported by the German Society for Advancement of Scientific Research (Grant numbers HE 1809/5-1, BR 1704/4-1); and by the START-Program of the Medical Faculty, RWTH Aachen, Aachen, Germany.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGEMENT
  8. DISCLOSURE
  9. References