Association of Maternal Lifestyles Including Smoking During Pregnancy with Childhood Obesity
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Department of Health Sciences, School of Medicine, University of Yamanashi, 1110 Shimokato, Chuo, Yamanashi 409-3898, Japan. E-mail: email@example.com
Objective: To examine the extent to which maternal smoking during early pregnancy and other prepregnancy lifestyle habits are associated with obesity and overweight in 5-year-old Japanese children.
Research Method and Procedures: We studied 1417 mother-child pairs enrolled in Project Enzan—a prospective cohort study. The dependent variables, childhood overweight and obesity, were defined with an international cut-off value. Maternal smoking during early pregnancy and other prepregnancy lifestyle habits were used as independent variables.
Results: Maternal smoking habits were associated with overweight in the 5-year-old children [adjusted odds ratio (OR): 2.15; 95% confidence interval (CI): 1.12 to 4.11]. Maternal sleep duration of ≥8 h/d negatively affected childhood overweight (adjusted OR: 0.71; 95% CI: 0.49 to 1.04). Children whose mothers skipped breakfast were likely to become overweight (adjusted OR: 1.78; 95% CI: 1.14 to 2.77). The results of childhood obesity analysis were similar to those of childhood overweight analysis.
Discussion: The results of this study suggest that there are effects of smoking during early pregnancy and other maternal lifestyle habits on the onset of childhood obesity in Japan. Therefore, interventions in maternal lifestyle habits are required to prevent childhood obesity, and these interventions should be initiated before pregnancy.
In industrialized countries, overweight and obesity are the most common nutritional disorders (1)(2)(3). Matsushita et al. (4) clearly showed the increasing trends in the prevalence of obesity in Japanese children. Overweight children have a high risk of being overweight in adulthood (5)(6), and they experience typical obesity-related morbidity (7). Because therapeutic interventions for overweight in children are expensive and the results are far from satisfactory (8), the development of strategies for preventing childhood overweight and obesity is a major challenge for health care professionals.
Many studies have suggested some risk factors for obesity, such as genetic and environmental factors, eating behavior, and physical activity. Fetal growth retardation has been suggested as one of the risk factors for adult obesity (9), and the association between low birth weight and increased risk of obesity has been reported by several studies (9)(10).
Moreover, several investigators have reported an increased risk of childhood or adulthood obesity after intrauterine exposure to maternal smoking (11)(12)(13)(14)(15)(16)(17). In developed countries, smoking during pregnancy is the principal environmental cause of intrauterine growth retardation (16)(18)(19). This relationship might be caused by the fact that the timing of catch-up growth is well characterized; birth weight was reported to be mainly affected by the second half of pregnancy (20).
To our knowledge, previous studies on the relationship between smoking during pregnancy and child outcomes were mainly retrospective or cross-sectional studies. Although some prospective studies were conducted (21)(22), the relationship between maternal smoking habits during pregnancy and childhood overweight in Japan has not been clarified.
In this study, we used the data from a dynamic prospective birth cohort study to verify the hypothesis that children of mothers who smoked during early pregnancy have an increased risk of childhood overweight and obesity. In addition, we aimed to clarify the relationship between other maternal lifestyle habits just before pregnancy and childhood overweight.
Research Methods and Procedures
Participants and Study Design
The study population was comprised of children born between January 1, 1991 and December 31, 1999 in Enzan City, Yamanashi Prefecture, Japan, and their mothers; these children underwent a medical checkup at 5 years of age. These subjects are the participants of Project Enzan, a dynamic ongoing prospective cohort study of pregnant women and their children in a Japanese rural area that started in 1988. The population of Enzan City is 27,000, with ∼200 births each year. We expected a high follow-up rate in this project because most of the people of this city had not migrated elsewhere.
In Japan, pregnant women must register their pregnancy at the city office, and children must undergo a medical checkup at the age of 1.5, 3, and 5 years. First, to ascertain the lifestyle habits of the expectant mothers, we conducted a questionnaire-based survey with the expectant mothers who visited the city office to register their pregnancy; informed consent was obtained before the survey. In Enzan City, >95% of the expectant mothers registered their pregnancy before week 16 of pregnancy. Next, during each medical checkup of the children born to these mothers, we surveyed the lifestyle habits of the children and their mothers using a questionnaire. During the checkups, we also obtained the growth and physical data of the children.
To ensure confidentiality, the mothers and children were identified by unique numbers to match the data obtained from the early pregnancy survey and the medical checkup at 5 years.
This study was approved by the Ethical Review Board, Yamanashi University School of Medicine on the basis of the “Guidelines Concerning Epidemiological Research” (the Ministry of Education, Culture, Sports, Science, and Technology and the Ministry of Health, Labor, and Welfare); the study was conducted with the cooperation of the Enzan City administration office.
The data on maternal lifestyle habits just before pregnancy, including the smoking status during early pregnancy, were obtained by administering a self-report questionnaire at the time of pregnancy registration. In this study, we used the following items as independent variables: maternal age, height, body weight, family structure (a family consists of either two, three, or more generations), duration of exercise before pregnancy (walking for <30 or ≥30 min/d), sleep duration (<8 or ≥8 h/d), smoking habits during early pregnancy (smoking, had quit smoking, or never smoked), alcohol consumption during early pregnancy (consuming alcohol, had quit alcohol, or never consumed alcohol), dietary habits (maintaining a good diet or not), and breakfast habits (having or skipping daily breakfast). Maternal body height and weight at the time of pregnancy registration were measured and recorded in the Maternal and Child Health Handbook by an obstetrician or a midwife. We used the BMI for the evaluation of maternal obesity. The maternal BMI was calculated according to the World Health Organization standards (body weight in kilograms divided by height in meters squared). The mothers were divided into two groups based on their BMI: an overweight group with a BMI of ≥25 kg/m2 and a non-overweight group with a BMI of <25 kg/m2.
The height and body weight data of the children were obtained from the physical measurements obtained during their medical checkup at 5 years of age. Height was measured using a conventional height bar (stadiometer; unit: 0.1 cm), whereas body weight was measured using a conventional weighing scale (unit: 100 grams).
Obesity and overweight in 5-year-old children are generally defined based on BMI (14)(23)(24). The definitions of childhood obesity and overweight have been previously established (25).
First, the relationship of maternal smoking habits during early pregnancy and other lifestyle habits just before pregnancy with childhood obesity and overweight were assessed using the χ2 test.
Next, we used a multiple logistic regression analysis to adjust the confounding factors, maternal age and BMI, and the variables that were significant in the univariate model. Breastfeeding and maternal education may be confounding factors; however, because a part of our data lacked information about these variables, we could not adopt these variables as confounding factors.
All analyses were conducted using SAS software, version 8.2 (SAS Institute, Inc., Cary, NC).
The mothers who completed the questionnaire gave birth to 1812 babies during the study period (the rate of infant medical checkup was 96%, and the questionnaire recovery rate was 99%). From these, questionnaires were collected from the mothers of 1443 children at the medical checkup of these children at 5 years (follow-up rate: 79.6%). Maternal lifestyle data were obtained for 1417 children (78.2%). Of these mothers, 76 (5.4%) had smoked during early pregnancy. The comparison of characteristics between smoking mothers and non-smoking mothers is shown in Table 1. Smoking mothers were significantly younger than non-smoking mothers. Moreover, registration of pregnancy from smoking mothers was significantly later than from non-smoking mothers. Non-smoking mothers were likely to be better educated than smoking mothers.
Table 1. . Comparison of characteristics between smoking mothers and non-smoking mothers
|Maternal age (yrs)||28.0 ± 4.9||29.2 ± 4.1||0.02|
|Registration of pregnancy (weeks)||12.3 ± 4.8||10.6 ± 3.1||<0.0001|
|Maternal BMI (kg/m2)||21.4 ± 4.4||20.8 ± 2.8||0.11|
|Birth weight of infant (grams)||2952 ± 471||3071 ± 410||0.01|
|Gestational week of infant (weeks)||38.9 ± 1.4||39.0 ± 1.4||0.80|
|Sex of infant|| || || |
| Female||35||657|| |
|Maternal education|| || || |
| Higher than high school||15||518||<0.0001|
| Up to high school||42||420|| |
Crude Relationship between Maternal Lifestyle and Childhood Overweight and Obesity
Crude odds ratios (ORs)1 and 95% confidence intervals (CIs) for maternal lifestyle factors in early pregnancy that affected childhood weight at 5 years of age are listed in Tables 2 (overweight) and 3 (obesity).
Table 2. . Crude and adjusted ORs and 95% CIs for maternal lifestyle factors in early pregnancy that affected childhood overweight at 5 years
|Smoking||1417|| || || || || || |
| Current smoker|| ||16||60||2.29||1.28, 4.08||2.15||1.12, 4.11|
| Ex-smoker and non-smoker|| ||140||1201|| || || || |
|Alcohol consumption||1395|| || || || || || |
| Current drinker|| ||14||125||0.89||0.50, 1.59|| || |
| Ex-drinker and non-drinker|| ||140||1116|| || || || |
|Sleep duration time||1417|| || || || || || |
| >8 h/d|| ||52||559||0.63||0.44, 0.89||0.71||0.49, 1.04|
| <8 h/d|| ||104||702|| || || || |
|Time of exercise||1393|| || || || || || |
| >30 min/d|| ||34||343||0.73||0.49, 1.09|| || |
| <30 min/d|| ||121||895|| || || || |
|Dietary habit||1411|| || || || || || |
| “My diet is unbalanced”|| ||90||649||1.30||0.93, 1.82|| || |
| “My diet is balanced”|| ||65||607|| || || || |
|Breakfast||1401|| || || || || || |
| “I sometimes skip”|| ||40||225||1.58||1.07, 2.32||1.78||1.14, 2.77|
| “I don't skip”|| ||115||1021|| || || || |
|Maternal psychological stress||1410|| || || || || || |
| “I feel psychological stress”|| ||10||74||1.09||0.55, 2.16|| || |
| “I don't feel psychological stress”|| ||146||1180|| || || || |
|Birth order of the infant||1417|| || || || || || |
| First|| ||63||514||0.98||0.70, 1.38|| || |
| Second or more|| ||93||767|| || || || |
|Breastfeed at 3 month||950|| || || || || || |
| Only breastfeeding|| ||34||357||0.62||0.41, 0.96||0.63||0.40, 1.01|
| Using bottle feeding|| ||74||485|| || || || |
|Maternal education||995|| || || || || || |
| Higher than high school|| ||47||486||0.58||0.39, 0.86||0.62||0.40, 0.96|
| Up to high school|| ||66||396|| || || || |
|Maternal occupation status||1406|| || || || || || |
| Working|| ||76||587||1.07||0.77, 1.50|| || |
| Not working|| ||80||663|| || || || |
|Birth weight||1416|| || || || || || |
| Low birth weight (<2500 grams)|| ||5||90||0.43||0.17, 1.08|| || |
| Normal birth weight (≥2500 grams)|| ||151||1170|| || || || |
The prevalence of childhood obesity was significantly higher among children whose mothers smoked during early pregnancy than among those whose mothers did not (crude OR: 5.14; 95% CI: 2.27 to 11.64). It was also significantly higher among children whose mothers skipped breakfast during early pregnancy than among those whose mothers did not (crude OR: 2.29; 95% CI: 1.16 to 4.54). On the other hand, the prevalence of childhood obesity was significantly lower among children whose mothers slept for at least 8 h/d during early pregnancy (crude OR: 0.29; 95% CI: 0.13 to 0.66) than among those whose mothers slept for <8 h/d. The results of childhood overweight analysis were similar to those of childhood obesity analysis.
Adjusted Relationship between Maternal Lifestyle and Childhood Overweight and Obesity
Next, we conducted a multiple logistic regression analysis to adjust the confounding factors, maternal BMI and maternal age, and to further analyze the significant relationships between the maternal lifestyle factors (smoking habits, sleep duration, and breakfast habits during early pregnancy) and childhood overweight and obesity. In this analysis, the maternal smoking habit was associated with the children being overweight at 5 years of age (adjusted OR: 2.15; 95% CI: 1.12 to 4.11). Maternal sleep duration of ≥8 h/d negatively affected childhood overweight; however, this relationship was not even slightly significant (adjusted OR: 0.71; 95% CI: 0.49 to 1.04). The children whose mothers skipped breakfast were likely to become overweight (adjusted OR: 1.78; 95% CI: 1.14 to 2.77). Furthermore, the children whose mothers smoked during early pregnancy showed an independent elevated risk for obesity compared with the children whose mothers did not smoke or who had quit smoking (adjusted OR: 3.93; 95% CI: 1.46 to 10.56). Similarly, we suggested that children whose mothers skipped breakfast were likely to become obese (adjusted OR: 3.00; 95% CI: 1.33 to 6.78), and long maternal sleep duration was a preventive factor for childhood obesity (adjusted OR: 0.37; 95% CI: 0.15 to 0.88).
The main finding of this study was the higher prevalence of obesity in children whose mothers smoked during early pregnancy in Japan. These results are similar to those of a previous study and consistent with those of Montgomery and Ekbom (11), who assessed the impact of maternal smoking during pregnancy on adult obesity (the National Child Development Study cohort). Although the point estimates were much higher in our results, the difference in the effect sizes might be caused by differences in the participants’ ages in these studies. Multiple factors can lead to obesity in young adults. Moreover, we suggested that childhood obesity was related to maternal lifestyle habits such as short sleep duration and skipping breakfast.
First, we discussed the relationship between maternal smoking and childhood obesity. It is postulated that smoking affects childhood obesity through either intrauterine exposure or maternal lifestyle habits.
Regarding intrauterine exposure, a newborn may be undernourished because of exposure to smoking during gestation; this condition is hypothesized to be associated with increased nutrient absorption, leading to post-natal obesity. It has been reported that undernutrition during early pregnancy increases the risk of adult obesity in a child (26) and that undernutrition causes intrauterine growth retardation and increases the risk of abnormal glucose tolerance (10). In addition, some animal studies have determined that smoking affects the hypothalamus (27) and the neurobehavioral system (28). Although smoking during pregnancy seems to have some impact on the hypothalamus and neurobehavioral system, this relationship has not been proven conclusively. The National Child Development Study cohort showed inconsistent results. Offspring of mothers who smoked during pregnancy were more likely to report a poor appetite, independent of several potential confounding factors (29). However, in this cohort, a relationship was observed between maternal smoking during pregnancy and adult obesity in offspring (11). Moreover, there was no relationship between low-birth-weight (LBW) infants and childhood obesity in our study. LBW was not a homogenous group. The etiologies of LBW were various, such as preterm birth, intrauterine growth retarded, and a combination of these factors. Because of these reasons, we were not able to clarify the relationship between LBW infants and childhood obesity.
One hypothesis suggests that, because mothers who smoke tend to have a less healthy lifestyle than their non-smoking counterparts, smoking is an indicator of other unhealthy lifestyle habits. The associated lifestyle habits might explain the tendency of the mothers who smoked to have obese children. This hypothesis suggests that other lifestyle habits are correlated with childhood obesity, and the findings of this study also indicate that smoking, sleep duration, and breakfast habits are correlated with childhood obesity. In this regard, it is possible that either the true influence of lifestyle habits, except smoking, was not accurately detected or that smoking was one of the unhealthy lifestyle habits having a significant impact.
Nevertheless, our results suggested that maternal alcohol consumption during pregnancy was not a risk factor for obesity or overweight in their offspring. This result was consistent with a previous report in that alcohol ingestion during pregnancy may decrease childhood body size (30).
A correlation has been reported between the lifestyle habits of the mother and those of the child (31), thus suggesting that less healthy maternal lifestyle habits affect childhood obesity.
This study had some advantages. First, because this was a prospective study, the possibility of recall bias for lifestyle habits such as smoking during pregnancy was low, and mothers were defined as the ones who smoked based on their reports of smoking during week 17 of pregnancy. Second, this study was the first to clarify the relationship of maternal lifestyle habits with childhood overweight and obesity in Japan. Finally, the follow-up rate of this study was high (78.2%), and the distribution of variables obtained from physical examinations among subjects was comparable to the national average. Therefore, the children enrolled in this study represented the Japanese children population.
However, this study also has some limitations. We used a questionnaire to obtain data on maternal lifestyle habits such as smoking during early pregnancy, but the validity of this questionnaire was not examined. A previous study on the accuracy of self-reported smoking habits among pregnant women showed that women reported their smoking very accurately (32). Based on this report, our results may be considered valid. Moreover, data on paternal height and weight were not obtained from the subjects. Therefore, it could not be statistically analyzed, and the effects of paternal genetic factors could not be sufficiently studied. Although we lacked data on paternal BMI, parental size, which reflected a genetic factor for childhood obesity and overweight, was partially addressed by inclusion of maternal BMI.
The results of this study suggested that there are effects of smoking during early pregnancy on the onset of childhood obesity in Japan. Furthermore, this study suggested that less healthy maternal lifestyle habits would increase the prevalence of childhood overweight and obesity. Our results indicated that, in the studied population, an unhealthy lifestyle was probably a more significant risk factor for childhood obesity than maternal smoking during pregnancy. Although there were some reports about the relationship between smoking during pregnancy and childhood obesity, maternal unhealthy lifestyles, such as skipping breakfast, have not been reported as risk factors of childhood obesity. Further studies are needed to confirm our results.
In conclusion, to prevent childhood obesity, intervention in early childhood lifestyle habits and maternal lifestyle habits is needed. This intervention should be started before pregnancy, because cessation of smoking after the detection of pregnancy may be too late to protect the offspring from potential obesity (15).
There was no funding/outside support for this study.
Nonstandard abbreviations: OR, odds ratio; CI, confidence interval; LBW, low birth weight.