Skip to main content
NCBI home page
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2019 Aug 7.
Published in final edited form as: Int J Obes (Lond). 2018 Feb 28;42(7):1249–1264. doi: 10.1038/s41366-018-0050-0

Maternal smoking during pregnancy and offspring overweight: is there a dose–response relationship? An individual patient data meta-analysis

Lucia Albers 1, Christina Sobotzki 1, Oliver Kuß 2, Teresa Ajslev 3, Rosangela FL Batista 4, Heloisa Bettiol 5, Bernard Brabin 6,7,8, Stephen L Buka 9, Viviane C Cardoso 5, Vicki L Clifton 10, Graham Devereux 11, Stephen E Gilman 12,13,14,15, Luke E Grzeskowiak 10, Joachim Heinrich 16, Sandra Hummel 17,18, Geir W Jacobsen 19, Graeme Jones 20, Gibby Koshy 6, Camilla Schmidt Morgen 3, Emily Oken 21, Tomas Paus 22, Zdenka Pausova 23, Sheryl L Rifas-Shiman 21, Andrea J Sharma 24, Antônio AM da Silva 4, Thorkild IA Sørensen 3,25, Elisabeth Thiering 26, Stephen Turner 11, Torstein Vik 27, Rüdiger von Kries 1
PMCID: PMC6685293  NIHMSID: NIHMS983919  PMID: 29717267

Abstract

Background/objectives

A number of meta-analyses suggest an association between any maternal smoking in pregnancy and offspring overweight obesity. Whether there is a dose–response relationship across number of cigarettes and whether this differs by sex remains unclear.

Subject/methods

Studies reporting number of cigarettes smoked during pregnancy and offspring BMI published up to May 2015 were searched. An individual patient data meta-analysis of association between the number of cigarettes smoked during pregnancy and offspring overweight (defined according to the International Obesity Task Force reference) was computed using a generalized additive mixed model with non-linear effects and adjustment for confounders (maternal weight status, breastfeeding, and maternal education) and stratification for sex.

Results

Of 26 identified studies, 16 authors provided data on a total of 238,340 mother–child-pairs. A linear positive association was observed between the number of cigarettes smoked and offspring overweight for up to 15 cigarettes per day with an OR increase per cigarette of 1.03, 95% CI = [1.02–1.03]. The OR flattened with higher cigarette use. Associations were similar in males and females. Sensitivity analyses supported these results.

Conclusions

A linear dose–response relationship of maternal smoking was observed in the range of 1–15 cigarettes per day equally in boys and girls with no further risk increase for doses above 15 cigarettes.

Introduction

Several recent meta-analyses showed a strong associations between maternal smoking during pregnancy and offspring overweight and obesity with pooled odds ratios (ORs) ranging from 1.33 to 1.60 [1–4]. Therefore, smoking abstinence during pregnancy might have substantial benefit for prevention of offspring obesity in addition to the avoidance of multiple tobacco-related harms to the mother and the child (i.e., preterm delivery, sudden infant death (SIDS), or birth defects). Although plausibility of a causal association between maternal smoking in pregnancy is supported by some animal [5–9] and DNA methylation studies [10–13], there remains concern regarding residual confounding in the observational studies. For example, several studies have shown that children exposed to paternal, or other second-hand smoke in utero or following pregnancy, were at increased risk of overweight, although risk was lower than that for maternal smoking [14–17]. Although associations of both maternal and paternal smoking with offspring overweight remained present despite controlling for parental weight and social class, this may reflect residual confounding by unmeasured neighborhood or family factors accounting for both.

Addressing potential residual confounding, one study within families where one child was exposed to maternal smoking and the other was not yielded inconclusive results [18], whereas another study using conditional fixed-effect models among siblings to control for unmeasured con-founding confirmed an effect of maternal smoking on overweight [19]. A recent meta-analysis suggested a much smaller specific effect of maternal smoking in pregnancy than reported in previous meta-analyses when taking account of the effect of paternal smoking as a negative control reflecting unmeasured family factors [2]. The association with paternal smoking, however, might not only be a reflection of residual confounding. There might be a genuine effect of paternal smoking in pregnancy related to intrauterine exposure to small nicotine doses resulting from maternal inhalation of father’s smoke. This hypothesis would be supported by a dose–response relationship for maternal smoking in pregnancy, if even small doses of maternal smoking are associated with offspring overweight. Indeed cotinine has been detected in newborns’ hair with paternal smoking exposure alone, which could arise from passive inhalation by the mother and transfer to fetus. These cotinine concentrations were within the range seen with maternal smoking [20, 21]. A dose–response relationship of maternal smoking and offspring overweight or obesity was detected in some [22–33], but not in all studies [19, 34–36], which may be due to different confounders considered and difference in categorization of the dose of maternal smoking. An individual patient data (IPD) meta-analysis allows for uniform assessment of the dose–response in all included studies.

There are several meta-analyses of the association between maternal smoking in pregnancy and offspring overweight or obesity [1–4], however, none has previously explored the dose–response relationship between maternal number of cigarettes during pregnancy and offspring obesity/overweight. Information on whether the risk of over-weight/obesity increases with the level of fetal nicotine exposure or whether there is a threshold below which there is no association can provide needed insight into the etiology of offspring overweight/obesity and information to further refine smoking cessation efforts during pregnancy not only for the mother, but potentially all household members. A valid assessment of the dose–response requires meta-analysis with uniform assessment of the dose–response in all included studies. Since the reported studies on dose–response assessed the effect in different smoking categories, this is only possible in IPD meta- analyses and could be materialized as many studies ascertained maternal smoking exposures in more detail than reported in the published articles.

Here we undertook an IPD meta-analysis designed to test the hypothesis that there was a linear relationship between the number of cigarettes smoked during pregnancy and risk for child overweight. As animal studies suggested that changes in the intrauterine milieu affecting body composition in the offspring may be different by sex, we stratified by offspring sex [37].

Methods

Potentially eligible studies were identified in a systematic literature search [38] (Fig. 1) using the following search term: (offspring OR children OR toddlers OR child OR infant OR adolescen* OR adult*) AND (overweight OR obesity OR obese OR adipose OR adiposity) AND (maternal smoking during pregnancy OR maternal smoking in pregnancy OR mother smoked during pregnancy OR mother smoked in pregnancy OR in utero nicotine exposure OR in utero exposure OR nicotine exposure during pregnancy OR nicotine exposure in pregnancy OR cigarettes during pregnancy OR cigarettes in pregnancy) AND (dose–response OR dose–effect OR dose OR amount of cigarettes OR number of cigarettes OR volume of cigarettes OR volume of nicotine). All studies (retrospective and prospective) that included data on the number of cigarettes mothers smoked during pregnancy and the weight and height of children ≥3 years were considered for inclusion in our IPD. Outcome had to be reported as overweight or obesity or body mass index (BMI) differences in the off-spring of mothers who smoked during pregnancy compared with offspring of mothers who did not smoke during pregnancy. Studies were excluded if the manuscript language was neither English nor German, or if the study population was already reported in another included study. All studies published before May 2015 were considered. The literature search was performed independently by two investigators (CS and RvK).

Fig. 1.

Fig. 1

Open in a new tab

Flow chart displaying the process of literature search and study selection

Authors of the selected studies were sent an invitation letter via e-mail. If no response was received after about 2 months, a second reminder e-mail was sent. Collaboration and data transfer agreements were signed by authors cooperating in this project.

The study was approved by the Ethics Committee of the LMU Munich (UE Nr. 024–14). For all included studies, individual ethical approval is documented in the respective original publications.

The study is registered at PROSPERO international register of systematic reviews with registration number CRD4201502475.

Assessment of study quality

Study quality was assessed based on the quality assessment criteria for observational cohort and cross-sectional studies of the National Institute of Health (http://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort). Eight questions out of 14 were appropriate for this analysis (Table S1). We excluded questions regarding sample size/power estimate, sufficient timeframe to observe effect, different levels for exposure, quality of exposure measure, several measures of exposure and adjustment for confounding variables, as the answers were obvious, or they were already considered in the inclusion criteria. Quality assessment was conducted independently by two investigators (RvK and LA) with each study rated as poor, fair, or good by mutual agreement.

Statistical methods

The primary outcome variables were overweight (including obesity) or obesity only (defined according to the International Obesity Task Force (IOTF) [39]) and were analyzed in two separate models. If data on BMI measurements at different ages were available, the measurement at the oldest available age was used in the analysis, since tracking of BMI increases by age [40–42].

The main explanatory variable was the number of cigarettes smoked by the mother during pregnancy of the child, who was included in the analysis. If the study provided multiple measures at different stages of pregnancy, we used the maximum number of cigarettes at any time point. In studies where the number of cigarettes was observed only in categories (e.g., none, 1–10, 11–20, >20 cigarettes per day), the actual numbers of cigarettes smoked during pregnancy were generated by randomly imputing a number from an assumed uniform distribution in the respective category for each mother. For the last, open categories (i.e., >20 cigarettes per day), numbers were imputed from an exponential distribution where the parameters of this distribution were estimated from the observations from all remaining studies using the actual observations above the lower category bound.

Potential confounders considered in the analysis were identified using a directed acyclic graph (Fig. 2). The number of potential confounders included in the models was driven by their availability in the studies included in the meta-analysis. In the main analysis, we considered (a) maternal weight status (underweight (BMI < 18 kg/m2), overweight (25 kg/m2 ≤ BMI < 30 kg/m2), obese (BMI ≥ 30 kg/m2), or normal weight (18 kg/m2 ≤ BMI < 25 kg/m2; which was used as reference)) (if available pre-pregnancy weight was used; if not available, then maternal weight at assessment of child’s BMI was used); (b) breastfeeding (for at least 1 month if available, else ever breastfeeding) (yes vs. no); (c) maternal education (at least high school completed or 10 years of school education vs. no high school completed or <10 years of school education).

Fig. 2.

Fig. 2

Open in a new tab

Directed acyclic graph on potential confounders

We also considered size at birth including small for gestational age (SGA; weight <10th percentile) or large for gestational age (LGA; weight >90th percentile) with reference to appropriate for gestational age (AGA; weight for gestational age between 10th and 90th percentile) as defined in the original studies or applying country-specific percentiles if not reported, and preterm delivery (<37 weeks of gestation) to be of substantial interest. First, effect modification was examined by stratifying for SGA, AGA, and LGA. Then, models with adjustment for SGA, LGA, and preterm delivery were provided in a supplementary analysis. These models would give the direct effect of smoking on overweight/obesity (beyond the effects working through SGA, LGA, or preterm delivery), whereas the main analysis gives the best estimate from the data of the overall causal effect of maternal smoking, namely the effect of a hypothetical intervention reducing maternal smoking on off-spring overweight/obesity [43].

Missing values for the potential confounders/mediating variables were imputed by a model-based single imputation step (PROC MI, SAS, V.9.4), the imputation model included the exposure, the confounders, and a categorical study effect. As the percentage of missing values was small (<2.2% of the observations for maternal weight status, child’s birth weight for gestational age, preterm delivery, breastfeeding, maternal education) and the sample size large we did not correct the analysis results by applying Rubin’s rules [44].

In a first step, the dichotomized effect of maternal smoking (yes vs. no) during pregnancy on either offspring overweight including obese children, overweight excluding obese children, or obesity excluding overweight children was analyzed in logistic regression models with adjustment for potential confounders (maternal weight status, breast-feeding, maternal education) and stratification for infant sex. A random intercept term for the respective study was included to account for variation between and correlation within studies. Family variations could not be taken into account in these models, thus sibling/twin data were excluded.

To analyze the dose–response relationship of number of cigarettes smoked during pregnancy, a generalized additive mixed model was used as described by Lin and Zhang for binary outcomes [45]. Such models use additive non-parametric functions to model the effect of covariates, while they additionally account for correlation of children–mother pairs within studies by adding a random study effect to the predictor. We used P-splines (smoothed linear functionals) for the estimation of the non-linear effect, with data-driven estimation of the smoothness of the effect by restricted maximum likelihood. The analysis was performed separately for boys and girls since some previous studies reported gender-specific differences of the association between maternal smoking in pregnancy and overweight in the offspring [24, 46–49]. Furthermore age-stratified models for the age groups <3, ≥3 to <5 years, ≥5 to <8, and ≥8 years (chosen to achieve as similar as possible numbers per stratum) were estimated.

In sensitivity analyses, further potential confounders (with data not available in all studies) were considered: (A) paternal smoking (yes vs. no), (B) child TV watching/video games (high = ‘ ≥ 1 h per day’; moderate/low = ‘ < 1 h per day’) at obesity assessment, (C) child physical activity (sufficient = ‘ ≥ 1 h per day’, low = ‘ < 1 h per day’) at obesity assessment.

Two additional sensitivity analyses were performed: one in which observations with imputed data (number of cigarettes and potential confounders) were excluded and another which only included studies where the study quality was rated good.

Results

The results of the literature search are shown in Fig. 1 with 26 studies meeting the inclusion criteria. Their investigators were invited to participate in the present IPD meta-analysis and 16 provided data [19, 22–28, 46, 50–56]. Study characteristics are shown in Table 1: the included studies (13 prospective studies and 3 retrospective studies) were undertaken in eight different countries with the assessment of BMI carried out in children of age 5 or older in most studies. In two studies, younger children with mean ages of 4.7 and 3.8 years were included [23, 26]. Thirteen of the 16 studies provided information on the precise number of maternal cigarettes smoked. For the remaining studies with interval censored data (with assessments in 4–5 dose categories) [28, 46, 52] imputation was performed. Paternal smoking during pregnancy was assessed in eight studies. Different definitions for small (and large) for gestational age were used across studies. Most studies used country-specific percentiles; two Brazilian studies used the Williams per-centiles [57] to define small (large) for gestational age. Another study used population-specific percentiles (10th and 90th) defined as cut-off points [22], whereas two studies used a web-calculator [23, 25]. Children were assumed to be breastfed if the mother reported at least 1 month of breastfeeding, in one study this was at least 1.5 months [27], in another at least 3 months exclusive breastfeeding [51], and in four studies any breastfeeding ever was assessed at time or at interview [23, 25, 26, 56]. Maternal pre-pregnancy BMI was assessed in nine studies, at interviews after pregnancies ended in five studies [19, 23, 27, 50, 51] and imputed in two studies by using the conditional distributions of the complete datasets [25, 50]. High maternal education was defined as completed high school or ≥9–10 years of school except for one study where ≥12 years of schooling was assumed as high education, and one study where a combination of education and occupation was assessed [22, 26]. The study quality was rated good in 11 studies and fair in 5 studies (Table S2 of the supplemental material).

Table 1.

Study characteristics

Author [ref], country Study, study type N cases include in our IPD Children’s age in years (mean ± SD) Continuous assessment of number of cigarettes Assessment of paternal smoking Potential mediators Potential confounder variables assessed Study quality (assessed with NIH toola
SGA/LGA Preterm Breastfeeding Maternal BMI after pregnancy Maternal BMI before pregnancy Maternal education
Møller [22], Denmark Danish National Birth Cohort, prospective study 44,544 7.0 ± 0.3 Yes Yes Yes, defined in study population, as <10th, respectively, >90th percentile adjusted for gestational age and gender Yes Yes (≥1 month) Yes Yes Yes, combination of education and occupation (low, medium, high)b Fair
Bettiol [50], Brazil Ribeirao Preto birth cohort, prospective study 723 10.6 ± 0.3 Yes Yes Yes, based on the Williams curve (Williams et al. [57]) Yes Yes (≥1 month) No No Yes, at least 9–10 years school (assessed in categories) Good
da Silva [50], Brazil São Luís birth cohort 672 8.2 ± 0.3 Yes Yes Yes, based on the Williams curve (Williams et al. [57]) Yes Yes (≥1 month) Yes No Yes, at least 9–10 years school (assessed in categories) Good
Gilman [19], United States Collaborative Perinatal Project (CPP), prospective study 12,516 Ca. 7 Yes No Yes, based on United States percentiles (Talge [79]) Yes No Yes No Yes, at least 10 years school Good
Grzeskowiak [23], Australia Women’s and Children’s Health Network, prospective study 6877 4.7 ± 0.3 Yes No Yes, calculated with generic birth weight centile calculator from gestation.net Yes Yes (any breastfeeding yes/no) Yes No No Good
Howe [24], United Kingdom ALPAC, prospective study 9127 15 ± 3.6 Yes Yes Yes, based on British percentiles (Cole [80]) Yes Yes (≥1 month) Yes Yes Yes, at least A level Good
Boerschmann [51], Germany German GDM offspring study, prospective study 492 13.5 ± 4.6 Yes Yes Yes, based on German Yes Yes (fully breastfed ≥3 months) Yes No No Fair, because of limited
Jones [46], Australia “Live births in Tasmania”, prospective study 390 Ca. 8 No, categorical assessment “null” “1–10”, “11–20”, “21–40”, “>40” (cig. per day) No percentiles (Voigt [81]) Yes, based on Australian percentiles (Dobbins [82]) Yes Yes (≥1 month) Yes Yes Yes, completed high school external validity Fair, because of limited external validity
Koshy [25], United Kingdom “15 primary schools in Merseyside”, retrospective study 1829 7.9 ± 1.9 Yes Yes Only SGA, IUGR computed Yes Yes (any breastfeeding yes/no) No No Yes, secondary education and above Fair
Oken [52], United Statesc Project Viva, prospective study 970 7.9 ± 0.8 No, categorical assessment “never smoker”, “<1”, “1–4”, “5–14”, “15–24”, “≥25” (cig. per day) No Yes, based on US percentiles (Oken [83]) Yes Yes (≥1 month) Yes Yes Yes, completed high school Good
Syme [53], Canada Saguenay Youth Study (SYS), retrospective cohort study of prenatal exposure to maternal cigarette smoking 478 13.7 ± 1.2 Yes Yes Yes, based on US percentiles (Talge [79]) Yes Yes (total duration in months) Yes Yes Yes, completed high school Good
Sharma [26], United States Prevention’s pregnancy nutrition surveillance system (PNSS), prospective study 71,270 3.8 ± 0.5 Yes No Yes, based on United States percentiles (Talge [79]) Yes Yes (any breastfeeding yes/no) Yes Yes Yes, ≥12 years school Good
Thiering [55], Germany GINI LISA, prospective study 6323 13.0 ± 3.9 Yes No Yes, using German percentiles (Voigt [81]) Yes Yes (≥1 month exclusively breastfed) Yes Yes Yes, ≥10 years school Good
Prabhu [56], United Kingdom SEATON, prospective study 841 7.7 ± 2.7 Yes Yes Yes, using British percentiles (Cole [80]) Yes Yes, (breastfeeding at 4th month after birth) No Yes Yes (age at leaving education at least 16) Good
Widerøe [27], Norway Trondheim and Bergen (Norway), and Uppsala (Sweden), prospective study 515 5.3 ± 0.2 Yes No Yes, defined in study population, as <10th, respectively, >90th percentile adjusted for sex, parity Yes Yes (≥1.5 months) Yes No Yes, at least 9 years school + 1–2 years further education Good
von Kries [28], Germany “Six Bavarian communities”, retrospective study 5594 6.2 ± 0.4 No, categorical assessment “no cigarettes”, “1–10”, “11–20”, “>20” (cig. per day) No Yes, using German percentiles (Voigt [81]) Yes Yes (≥1 months fully breastfed) No Yes Yes, at least >9 years school Fair
Open in a new tab
a

Detailed quality assessment in online supplement Table S1

b

Socio-occupational status based on the current or most recent job within 6 months, or, if the woman was attending school, on the type of education. Women in training were categorized according to the type of education they headed for. The category “high” included women in management jobs or in jobs requiring higher education (generally >4 years beyond high school). Office workers, service workers, skilled manual workers, and women in the military constituted the “middle” category. The “low” category included unskilled workers and unemployed women. Women with no connection to the labor market (not in training, not disability-retired, not house wife, not on public support) were also categorized in the “low” category.

c

The most recent outcome data (mid-childhood) assessed in that study was used (not included in that publication)

In total, N = 422,064 BMI measurements (including multiple measurements per child) of children/adolescents years were available. After excluding twins and siblings (only first child was included), observations with missing data on maternal number of cigarettes, and observations where sex- and age-specific weight class according to the IOTF [39] could not be assigned (excluding children aged < 2 years with no such reference data, or children with missing data on gender) N = 238,340 mother–child pairs were available for analysis (boys N = 121,254, girls N = 117,086; Fig. 3).

Fig. 3.

Fig. 3

Open in a new tab

Flow chart on mother–child pairs included in our final study population

The prevalence of offspring overweight (including obesity) was 18.50% (N = 44,088), of which obesity counted for 5.07% (N = 12,081). In all, 21.77% (N = 51,887) of mothers reported to have smoked during pregnancy with a mean number of cigarettes per day of 11.06 (SD = 9.06). The overall ORs in offspring of mothers who smoked compared with offspring of mothers who did not smoke during pregnancy was 1.26 (95% confidence interval (CI) = [1.22–1.29]) for overweight (including obesity) (girls: 1.22 with 95% CI = [1.18–1.27]; boys: 1.30 with 95% CI = [1.25–1.35]) and 1.24 (95% CI = [1.18–1.29]) (girls: 1.25 with 95% CI = [1.17–1.37]; boys: 1.22 with 95% CI = [1.14–1.51]) for obesity in the adjusted (for maternal weight status, breastfeeding, maternal education) random effect model that included data for all 16 studies. For overweight excluding obesity, the corresponding OR was 1.26 (95% CI = [1.22–1.30]). In the sub-sample where paternal smoking was assessed (N = 58,812), the OR for the global association between maternal smoking and both overweight (including obesity) and obesity only without adjustment for paternal smoking was higher (overweight: 1.46, 95% CI = [1.39–1.55]; obesity: 1.54, 95% CI = [1.39–1.71]); after adjusting for paternal smoking OR were 1.37 (95% CI = [1.29–1.45]) for overweight (including obesity) and 1.40 (95% CI = [1.26–1.57]) for obesity only.

We analyzed the number of cigarettes on a continuous scale to assess a dose–response relationship for both overweight and obesity overall and stratified by sex. The odds of a child being overweight or obese increased linearly up to 10–15 cigarettes per day and levelled out for doses higher than 15 cigarettes per day (Fig. 4). For example for 12 cigarettes per day, ORs were 1.29 (95% CI = [1.25–1.33]) for overweight (including obesity) and 1.26 (95% CI = [1.20–1.33]) for obesity only, reflecting an OR per additional cigarette of 1.02 [1.02–1.02] for overweight (including obesity) and 1.02 [1.02–1.02]) for obesity only. The association for overweight appeared to be slightly more pronounced in boys than in girls but with widely overlapping 95% CIs (Fig. 4).

Fig. 4.

Fig. 4

Open in a new tab

Association of maternal number of cigarettes smoked per day and risk of offspring overweight (including obesity) and obesity only stratified by gender (____ = odds ratio (OR) for the association between maternal number of cigarettes and offspring overweight/obesity; _ _ _ = 95% CI of the OR; the vertical dashes above the x axis indicate the density of the observations underlying the model)

Stratified analysis by age at BMI assessment showed an increase of the effect size by age, with the largest ORs observed for those aged 5–8 years (Fig. 5).

Fig. 5.

Fig. 5

Open in a new tab

Association of maternal number of cigarettes smoked per day and risk of offspring overweight (including obesity) and obesity only stratified for age groups (two to younger than three years (N = 82,572/N = 70,054), 3 to younger than 5-year-old children (N = 85,019/N = 72,805), 5 to younger than 8-year-old children (N = 78,954/N = 71,997), over 8-year-old children (N = 17,936/N = 15,458) (____ = odds ratio (OR) for the association between maternal number of cigarettes and offspring overweight/obesity; _ _ _ = 95% CI of the OR; ……. = OR with 95% CI for the overall effect of the main model; the vertical dashes above the x axis indicate the density of the observations underlying the model)

For birth weight for gestational age, stratified analysis did not suggest effect modification (associations between maternal smoking and offspring overweight (including obesity) was OR = 1.26 with 95% CI = [1.17–1.36] in SGA children, OR = 1.33 with 95% CI = [1.29–1.37] in AGA children, and OR = 1.29 with 95% CI = [1.18–1.42] in LGA children). Models with adjustment for SGA (Figure S1) and LGA (Figure S2) both showed a general increase in effect compared with the main model. In the model with adjustment for preterm delivery, nearly no change in the association was seen (Figure S3).

Sensitivity analyses, adjusting for additional potential confounding variables–assessed only in some of the included studies–yielded very similar results compared with models without additional adjustment for these variables. With adjustment for paternal smoking (N = 58,812; eight studies) a similar pattern was observed compared with the model not adjusted for paternal smoking: for overweight (including obesity) the increasing risk per cigarette was OR = 1.02 (95% CI = [1.02–1.03]) compared with OR = 1.03 (95% CI = [1.02–1.03]) for the model not adjusted for paternal smoking; for obesity OR = 1.02 (95% CI = [1.02–1.03]) compared with OR = 1.03 (95% CI = [1.02–1.04]) (Figure S4). In the sample where child TV watching/video games was assessed (N = 18,850; six studies), additional adjustment did not change the results for the association with overweight (including obesity) (Figure S5). For obesity only in general, CIs were very wide precluding any conclusions. When adjusting the original model additionally for child physical activity (N = 12,338; eight studies) the magnitude of the dose–response effect for both overweight (including obesity) and obesity only for the main analysis was unchanged (Figure S6).

Restricting the analysis to the 11 studies with good quality (excluding also retrospective studies except one with validation of exposure in medical records), showed essentially no change in the association of the number of cigarettes smoked during pregnancy with offspring overweight (including obesity) and obesity only. Associations were of slightly smaller magnitude with a linear effect up to 20 cigarettes per day. However, confidence limits were widely overlapping (Figure S7). Including only completely assessed data without imputation (for the interval censured, maternal smoke dose exposures, or missing values for confounder variables) showed very similar dose–response effects for both overweight (including obesity) and obesity only compared with the main analysis (Figure S8).

Discussion

Our data show a linear increase in offspring risk for becoming overweight and obese by number of cigarettes smoked during pregnancy for up to 10–15 cigarettes per day. This relationship was most pronounced in children aged 5–8 years, which accords with previous evidence that the effect emerges in the preschool years [49]. Thus, even few maternal cigarettes smoked per day may confer risk for subsequent offspring overweight and obesity. With further increments in smoking frequency beyond 15 cigarettes per day, there was no apparent increased additional risk.

Most previous studies attempting to assess dose–response relationships for maternal smoking did not analyze the number of cigarettes smoked on a continuous scale, but compared categories using 5–10 cigarette groupings (reference none smoking) thus yielding imprecise estimates of the dose–response relationship [17, 23–29, 29–33, 58, 59]. Some of these studies did not detect a dose–response relationship [19, 34, 36, 60]. Only two studies assessed dose–response relationships by number of cigarettes on a continuous scale [22, 35] and these assumed a linear association over the whole range of frequency of cigarette use. In the present analysis, applying P-splines for the estimation of non-linear effects, with data-driven estimation of the smoothness of the effect by generalized cross-validation minimization, no fixed linear association was forced on the data. Indeed, a linear association was only observed for up to 10–15 cigarettes. The observation of flattening of the effect with very high number of cigarettes smoked by the mother might be due to reporting bias, which might arise if heavy smoking mothers lose awareness of the number of cigarettes smoked. Assuming selective under-reporting of excessive smoking, however, would rather account for an upward shift of the curve.

Implications of study findings

As cotinine concentrations in the offspring related to paternal cigarette smoke exposure alone [61] can be similar to concentrations when only a few cigarettes are smoked by the mother, the linear dose–response relationship up to 10–15 cigarettes may have implications for the under-standing of the role of paternal smoking for offspring overweight [2]. The paternal smoking effect might be a reflection of low doses by passive smoking; exposing the pregnant mother to environmental tobacco smoke (ETS) may have a genuine effect on the child’s risk for overweight. Cotinine values in urine of neonates from non-smoking mothers increase in relation to number of daily cigarettes smoked by the father during pregnancy [62]. Interestingly, two studies reported a dose–response relationship for the risk of overweight and obesity for paternal smoking during pregnancy [17, 25]. Whether this effect of paternal smoking is mediated by passive smoking of the mother during pregnancy, or is transmitted via the spermatozoal genome (meaning the preconceptional toxical exposure of the father) as explored in a recent methylation study [63] is unknown. A low exposure to maternal smoking, which appears to have an effect on offspring overweight/obesity, may be mimicked by ETS. Therefore, one implication of our findings is that any environmental smoke exposure during pregnancy might causally be related to overweight/obesity in offspring.

Mechanistic pathways linking prenatal exposure to cigarette smoking to obesity are not well understood. One potential pathway may involve exposure-related effects on the developing brain-reward system. The system processes hedonic properties of food (as well as drugs of abuse) and includes brain structures, such as the amygdala [64]. In a brain-imaging study of adolescents, prenatal exposure to maternal cigarette smoking was associated with higher adiposity and preference for fatty foods and lower volume of the amygdala; further, amygdala volume correlated inversely with fat intake [65]. Diets high in fats are considered rewarding [66] and obesogenic [67], as fats compared with other macronutrients (i.e., carbohydrates and proteins) are of higher energy density and efficiency [68]. The amygdala has been studied extensively in the context of both drug addiction and the regulation of fat preference. With respect to the former, lower amygdala volume has been observed in individuals with alcohol addiction in whom it was associated with greater alcohol craving and more likely relapse into alcohol consumption [69]. With respect to the regulation of fat preference, activation of the amygdala by intra-amygdala administrations of neuropeptide Y and enterostatin decreases dietary preference for fat in experimental animals [70, 71]. In human brain-imaging studies, the amygdala is activated by high-fat vs. low-fat food stimuli [72]. These observations are consistent with the possible role of the prenatal exposure-induced reduction of the amygdala size in increasing fat preference and, in turn, risk for obesity.

Strengths and limitations

The major strengths of this study are the large sample size and application of a dose–response model allowing assessment of dose–response in a uniform analysis by number of cigarettes smoked and confounding factors. In contrast to previous studies, this study did not restrict estimates to a linear association, but instead employed P-splines to examine possible non-linear effects.

The validity of the findings is supported by the robustness of these results confirmed by sensitivity analyses considering paternal smoking and other possible confounding variables.

The dose–response relationship observed in the main analysis might still reflect residual confounding due to imprecise measurement and limited information on potential confounders. However, the sensitivity analysis, based on studies, which provided more extensive information on confounders including paternal smoking, physical activity, and TV watching/video games, yielded very similar risk estimates and strengthens the main conclusion. Confounding by unknown risk factors, for example, nutrition and eating patterns [73] cannot be excluded.

Furthermore, we showed that size for gestational age is not an effect modifier for the association between maternal smoking during pregnancy and offspring overweight. Hence, it might act as mediator. Adjustment for size at birth and gestational age, (Fig. 2) yielded generally higher estimates with a similar pattern as the main analysis results. These estimates can be interpreted as the direct effect of smoking on overweight or obesity (independent of the effects working through SGA, LGA, or preterm delivery), whereas the models without adjustment for these potential mediating variables estimates the total effect of maternal smoking. These higher estimates might imply that there are two oppositely acting pathways from maternal smoking during pregnancy through offspring overweight and obesity: one reducing child adiposity by reducing birth weight and another increasing child adiposity through another pathway.

Selection bias due to non-participation of eligible studies, whose authors did not contribute data to the IPD analyses [17, 18, 29, 32, 33, 35, 36, 58–60], might be an issue. We summarized study characteristics and dose–response results for the number of cigarettes smoked during pregnancy or overall results for the association between smoking in pregnancy and offspring anthropometric outcome in studies not providing data for the IPD meta-analysis in Table S3 of the supplemental material. Unfortunately it was impossible to provide a summary estimate of the dose–response relationship reported in the studies, which had not provide data, because units, outcomes, statistics differed between studies. In studies reporting ORs for the association between overweight/obesity and maternal smoking, the strength of the effects were comparable with the main findings.

It would have been ideal to use also repeated BMI outcome measures of the same child for the analysis. Therefore, we tried to estimate such models with an additional random effect for the child’s identification number, but unfortunately these models did not converge irrespective of which statistical software was used (neither R nor SAS).

A concern for validity is that mothers may have underreported the number of cigarettes smoked during pregnancy due to negative social stigma associated with smoking in pregnancy. In cases where under-reporting was selective, meaning that only those reporting the lowest number of cigarettes were misreporting and those who reported smoking more cigarettes gave the true numbers, this could be an explanation for the flattening of the dose–response effect. However, there is no ideal biomarker for early pregnancy smoking exposure. Cotinine concentration in the newborn’s hair constitutes a very precise measure for the cumulative smoke exposure during pregnancy during the last 3 months of the pregnancy [74]. Such data have demonstrated a close association between the self-reported number of maternal cigarettes smoked and the measured newborn hair cotinine concentration [75]. However, maternal smoking in the third trimester might not be the best indicator for overall smoke exposure of the fetus [76]. Good markers for early pregnancy smoke exposure are required. End-tidal breath carbon monoxide levels and urine cotinine levels in the mother do provide more accurate measurements for recent nicotine and carbon monoxide exposure [77], but may indicate transient exposures rather than chronicity during pregnancy. Substantial within-person fluctuation may exist if women repeatedly try to quit or cutdown. This may explain why CIs widen at doses >15 cigarettes. Pickett et al. suggest that where timing, intensity, and duration of exposure are critical, self-reported history of cigarette consumption may be a better measure for fetal exposure [78]. Maternal smoking status at different stages of pregnancy was only reported in few studies, therefore in our study we could not assess whether the duration of smoking is also important for child overweight and obesity. If a longer duration is strongly associated with offspring overweight and obesity, as suggested by a large study from the United States [26], our current results would be an underestimate of the true association among continued smokers.

Conclusion

A linear dose–response relationship between maternal smoking during pregnancy and the child’s risk for overweight was observed for mothers who smoked 1–15 cigarettes per day. As these findings suggest that even very low doses of cigarette smoke exposure during pregnancy may increase the risk of offspring overweight and obesity, family smoking cessation programs and recommendations about avoiding passive smoke exposure are warranted.

Supplementary Material

Figure Legends Supplemental Figures
Supplemental Tables 1-3
S_Figure 1
S_Figure 2
S_Figure 3
S_Figure 4
S_Figure 5
S_Figure 6
S_Figure 7
S_Figure 8

Acknowledgements

We want to thank the funders of the individual studies: the UK Medical Research Council and the Wellcome Trust (Grant ref: 102215/2/13/2) and the University of Bristol, the Danish National Research Foundation, Pharmacy Foundation, the March of Dimes Birth Defects Foundation, the Augustinus Foundation, and the Health Foundation, the US NICHD (contracts no. 1-HD-4–2803 and no. 1-HD-1–3127, R01 HD HD034568), the NHMRC, the CNPq (Portuguese acronym for the National Research Council—grant 523474/96–2) and FAPESP (Portuguese acronym for the São Paulo State Research Council—grant 00/0908–7). We would like to thank the participating families of all studies for the use of data. For the ASPAC study, we want to thank the midwives for their help in recruiting families, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses. This work was supported by the Deutschen Forschungsgesellschaft (German Research Foundation, DFG) [KR 1926/9–1, KU1443/4-1]. Dr. Gilman’s contribution was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

Footnotes

Conflict of interest The authors declare that they have no conflict of interest.

Disclaimer

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.

Electronic supplementary material The online version of this article (https://doi.org/10.1038/s41366-018-0050-0) contains supplementary material, which is available to authorized users.

References

  • 1.Rayfield S, Plugge E. Systematic review and meta-analysis of the association between maternal smoking in pregnancy and childhood overweight and obesity. J Epidemiol Community Health. 2016;71:162–73. 10.1136/jech-2016-207376 [DOI] [PubMed] [Google Scholar]
  • 2.Riedel C, Schönberger K, Yang S, Koshy G, Chen Y-C, Gopinath B, et al. Parental smoking and childhood obesity: higher effect estimates for maternal smoking in pregnancy compared to paternal smoking - a meta-analysis. Int J Epidemiol. 2014;43:1593–606. [DOI] [PubMed] [Google Scholar]
  • 3.Ino T Maternal smoking during pregnancy and offspring obesity: meta-analysis. Pediatr Int. 2010;52:94–99. [DOI] [PubMed] [Google Scholar]
  • 4.Oken E, Levitan EB, Gillman MW. Maternal smoking during pregnancy and child overweight: systematic review and metaanalysis. Int J Obes. 2008;32:201–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Zinkhan EK, Lang BY, Yu B, Wang Y, Jiang C, Fitzhugh M, et al. Maternal tobacco smoke increased visceral adiposity and serum corticosterone levels in adult male rat offspring. Pediatr Res. 2014;76:17–23. [DOI] [PubMed] [Google Scholar]
  • 6.Gao Y-J, Holloway AC, Zeng Z, Lim GE, Petrik JJ, Foster WG, et al. Prenatal exposure to nicotine causes postnatal obesity and altered perivascular adipose tissue function. Obes Res. 2005;13:687–92. [DOI] [PubMed] [Google Scholar]
  • 7.Somm E, Schwitzgebel VM, Vauthay DM, Camm EJ, Chen CY, Giacobino J-P, et al. Prenatal nicotine exposure alters early pancreatic islet and adipose tissue development with consequences on the control of body weight and glucose metabolism later in life. Endocrinology. 2008;149:6289–99. [DOI] [PubMed] [Google Scholar]
  • 8.Oliveira E, Moura EG, Santos-Silva AP, Fagundes ATS, Rios AS, Abreu-Villaça Y, et al. Short- and long-term effects of maternal nicotine exposure during lactation on body adiposity, lipid profile, and thyroid function of rat offspring. J Endocrinol. 2009;202:397–405. [DOI] [PubMed] [Google Scholar]
  • 9.Holloway AC, Lim GE, Petrik JJ, Foster WG, Morrison KM, Gerstein HC. Fetal and neonatal exposure to nicotine in Wistar rats results in increased beta cell apoptosis at birth and postnatal endocrine and metabolic changes associated with type 2 diabetes. Diabetologia. 2005;48:2661–6. [DOI] [PubMed] [Google Scholar]
  • 10.Yousefi M, Karmaus W, Zhang H, Ewart S, Arshad H, Holloway JW. The methylation of the LEPR/LEPROT genotype at the promoter and body regions influence concentrations of leptin in girls and BMI at age 18 years if their mother smoked during pregnancy. Int J Mol Epidemiol Genet. 2013;4:86–100. [PMC free article] [PubMed] [Google Scholar]
  • 11.Rhee KE, Phelan S, McCaffery J. Early determinants of obesity: genetic, epigenetic, and in utero influences. Int J Pediatr. 2012;2012:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Joubert BR, Felix JF, Yousefi P, Bakulski KM, Just AC, Breton C, et al. DNA methylation in newborns and maternal smoking in pregnancy: genome-wide consortium meta-analysis. Am J Hum Genet. 2016;98:680–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Lee KWK, Richmond R, Hu P, French L, Shin J, Bourdon C, et al. Prenatal exposure to maternal cigarette smoking and DNA methylation: epigenome-wide association in a discovery sample of adolescents and replication in an independent cohort at birth through 17 years of age. Environ Health Perspect. 2015;123:193–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Leary SD, Smith GD, Rogers IS, Reilly JJ, Wells JC, Ness AR. Smoking during pregnancy and offspring fat and lean mass in childhood. Obes Silver Spring. 2006;14:2284–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Kleiser C, Schaffrath Rosario A, Mensink GB, Prinz-Langenohl R, Kurth BM. Potential determinants of obesity among children and adolescents in Germany: results from the cross-sectional KiGGS Study. BMC Public Health. 2009;9:46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Plachta-Danielzik S, Kehden B, Landsberg B, Schaffrath Rosario A, Kurth BM, Arnold C, et al. Attributable risks for childhood overweight: evidence for limited effectiveness of prevention. Pediatrics. 2012;130:e865–e871. [DOI] [PubMed] [Google Scholar]
  • 17.Harris HR, Willett WC, Michels KB. Parental smoking during pregnancy and risk of overweight and obesity in the daughter. Int J Obes. 2005;2013:1356–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Iliadou AN, Koupil I, Villamor E, Altman D, Hultman C, Långström N, et al. Familial factors confound the association between maternal smoking during pregnancy and young adult offspring overweight. Int J Epidemiol. 2010;39:1193–202. [DOI] [PubMed] [Google Scholar]
  • 19.Gilman SE, Gardener H, Buka SL. Maternal smoking during pregnancy and children’s cognitive and physical development: a causal risk factor? Am J Epidemiol. 2008;168:522–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Eliopoulos C, Klein J, Phan MK, Knie B, Greenwald M, Chitayat D, et al. Hair concentrations of nicotine and cotinine in women and their newborn infants. JAMA. 1994;271:621–3. [PubMed] [Google Scholar]
  • 21.Eliopoulos C, Klein J, Chitayat D, Greenwald M, Koren G. Nicotine and cotinine in maternal and neonatal hair as markers of gestational smoking. Clin Investig Med Médecine Clin Exp. 1996;19:231–42. [PubMed] [Google Scholar]
  • 22.Møller SE, Ajslev TA, Andersen CS, Dalgård C, Sørensen TIA. Risk of childhood overweight after exposure to tobacco smoking in prenatal and early postnatal life. PLoS ONE. 2014;9:e109184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Grzeskowiak LE, Hodyl NA, Stark MJ, Morrison JL, Clifton VL. Association of early and late maternal smoking during pregnancy with offspring body mass index at 4 to 5 years of age. J Dev Orig Health Dis. 2015;6:485–92. [DOI] [PubMed] [Google Scholar]
  • 24.Howe LD, Matijasevich A, Tilling K, Brion MJ, Leary SD, Smith GD, et al. Maternal smoking during pregnancy and offspring trajectories of height and adiposity: comparing maternal and paternal associations. Int J Epidemiol. 2012;41:722–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Koshy G, Delpisheh A, Brabin BJ. Dose response association of pregnancy cigarette smoke exposure, childhood stature, overweight and obesity. Eur J Public Health. 2011;21:286–91. [DOI] [PubMed] [Google Scholar]
  • 26.Sharma AJ, Cogswell ME, Li R. Dose-response associations between maternal smoking during pregnancy and subsequent childhood obesity: effect modification by maternal race/ethnicity in a low-income US cohort. Am J Epidemiol. 2008;168:995–1007. [DOI] [PubMed] [Google Scholar]
  • 27.Widerøe M, Vik T, Jacobsen G, Bakketeig LS. Does maternal smoking during pregnancy cause childhood overweight? Paediatr Perinat Epidemiol. 2003;17:171–9. [DOI] [PubMed] [Google Scholar]
  • 28.von Kries R, Bolte G, Baghi L, Toschke AM, Group GMES. Parental smoking and childhood obesity--is maternal smoking in pregnancy the critical exposure? Int J Epidemiol. 2008;37:210–6. [DOI] [PubMed] [Google Scholar]
  • 29.Hill SY, Shen S, Locke Wellman J, Rickin E, Lowers L. Offspring from families at high risk for alcohol dependence: increased body mass index in association with prenatal exposure to cigarettes but not alcohol. Psychiatry Res. 2005;135:203–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Chen A, Pennell ML, Klebanoff MA, Rogan WJ, Longnecker MP. Maternal smoking during pregnancy in relation to child overweight: follow-up to age 8 years. Int J Epidemiol. 2006;35:121–30. [DOI] [PubMed] [Google Scholar]
  • 31.Reilly JJ, Armstrong J, Dorosty AR, Emmett PM, Ness A, Rogers I, et al. Early life risk factors for obesity in childhood: cohort study. BMJ. 2005;330:1357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Power C, Jefferis BJ. Fetal environment and subsequent obesity: a study of maternal smoking. Int J Epidemiol. 2002;31:413–9. [PubMed] [Google Scholar]
  • 33.Dior UP, Lawrence GM, Sitlani C, Enquobahrie D, Manor O, Siscovick DS, et al. Parental smoking during pregnancy and off-spring cardio-metabolic risk factors at ages 17 and 32. Atherosclerosis. 2014;235:430–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Durmuş B, Ay L, Hokken-Koelega ACS, Raat H, Hofman A, Steegers EAP, et al. Maternal smoking during pregnancy and subcutaneous fat mass in early childhood. The Generation R Study. Eur J Epidemiol. 2011;26:295–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Gorog K, Pattenden S, Antova T, Niciu E, Rudnai P, Scholtens S, et al. Maternal smoking during pregnancy and childhood obesity: results from the CESAR Study. Matern Child Health J. 2011;15:985–92. [DOI] [PubMed] [Google Scholar]
  • 36.Huang DYC, Lanza HI, Anglin MD. Trajectory of adolescent obesity: exploring the impact of prenatal to childhood experiences. J Child Fam Stud. 2014;23:1090–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Dahlhoff M, Pfister S, Blutke A, Rozman J, Klingenspor M, Deutsch MJ, et al. Peri-conceptional obesogenic exposure induces sex-specific programming of disease susceptibilities in adult mouse offspring. Biochim Biophys Acta. 2014;1842:304–17. [DOI] [PubMed] [Google Scholar]
  • 38.Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:2008–12. [DOI] [PubMed] [Google Scholar]
  • 39.Cole TJ. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000;320:1240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Sachdev HS, Fall CHD, Osmond C, Lakshmy R, Dey Biswas SK, Leary SD, et al. Anthropometric indicators of body composition in young adults: relation to size at birth and serial measurements of body mass index in childhood in the New Delhi birth cohort. Am J Clin Nutr. 2005;82:456–66. [DOI] [PubMed] [Google Scholar]
  • 41.Ylihärsilä H, Kajantie E, Osmond C, Forsén T, Barker DJ, Eriksson JG. Body mass index during childhood and adult body composition in men and women aged 56–70 y. Am J Clin Nutr. 2008;87:1769–75. [DOI] [PubMed] [Google Scholar]
  • 42.Bayer O, Krüger H, von Kries R, Toschke AM. Factors associated with tracking of BMI: a meta-regression analysis on BMI tracking. Obes Silver Spring Md. 2011;19:1069–76. [DOI] [PubMed] [Google Scholar]
  • 43.VanderWeele TJ, Mumford SL, Schisterman EF. Conditioning on intermediates in perinatal epidemiology. Epidemiol Camb Mass. 2012;23:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Rubin DB. Multiple imputation for nonresponse in surveys. Hoboken, NJ: Wiley-Interscience; 2004. [Google Scholar]
  • 45.Lin X, Zhang D. Inference in generalized additive mixed mod-elsby using smoothing splines. J R Stat Soc Ser B Stat Methodol. 1999;61:381–400. [Google Scholar]
  • 46.Jones G, Riley M, Dwyer T. Maternal smoking during pregnancy, growth, and bone mass in prepubertal children. J Bone Miner Res J Am Soc Bone Miner Res. 1999;14:146–51. [DOI] [PubMed] [Google Scholar]
  • 47.Suzuki K, Kondo N, Sato M, Tanaka T, Ando D, Yamagata Z. Gender differences in the association between maternal smoking during pregnancy and childhood growth trajectories: multilevel analysis. Int J Obes. 2011;35:53–59. [DOI] [PubMed] [Google Scholar]
  • 48.Suzuki K, Kondo N, Sato M, Tanaka T, Ando D, Yamagata Z. Maternal smoking during pregnancy and childhood growth trajectory: a random effects regression analysis. J Epidemiol. 2012;22:175–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Riedel C, Fenske N, Muller MJ, Plachta-Danielzik S, Keil T, Grabenhenrich L, et al. Differences in BMI z-scores between offspring of smoking and nonsmoking mothers: a longitudinal study of German children from birth through 14 years of age. Env Health Perspect. 2014;122:761–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Silva AA, Barbieri MA, Cardoso VC, Batista RF, Simões VM, Vianna EO, et al. Prevalence of non-communicable diseases in Brazilian children: follow-up at school age of two Brazilian birth cohorts of the 1990’s. BMC Public Health. 2011;11:486. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Boerschmann H, Pflüger M, Henneberger L, Ziegler A-G, Hummel S. Prevalence and predictors of overweight and insulin resistance in offspring of mothers with gestational diabetes mellitus. Diabetes Care. 2010;33:1845–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Oken E, Huh SY, Taveras EM, Rich-Edwards JW, Gillman MW. Associations of maternal prenatal smoking with child adiposity and blood pressure. Obes Res. 2005;13:2021–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Syme C, Abrahamowicz M, Mahboubi A, Leonard GT, Perron M, Richer L, et al. Prenatal exposure to maternal cigarette smoking and accumulation of intra-abdominal fat during adolescence. Obes Silver Spring Md. 2010;18:1021–5. [DOI] [PubMed] [Google Scholar]
  • 54.Pausova Z, Paus T, Abrahamowicz M, Bernard M, Gaudet D, Leonard G, et al. Cohort profile: the Saguenay Youth Study (SYS). Int J Epidemiol. 2016;46:e19 10.1093/ije/dyw023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Thiering E, Brüske I, Kratzsch J, Thiery J, Sausenthaler S, Meisinger C, et al. Prenatal and postnatal tobacco smoke exposure and development of insulin resistance in 10 year old children. Int J Hyg Environ Health. 2011;214:361–8. [DOI] [PubMed] [Google Scholar]
  • 56.Prabhu N, Smith N, Campbell D, Craig LC, Seaton A, Helms PJ, et al. First trimester maternal tobacco smoking habits and fetal growth. Thorax. 2010;65:235–40. [DOI] [PubMed] [Google Scholar]
  • 57.Williams RL. Intrauterine growth curves: intra- and international comparisons with different ethnic groups in California. Prev Med. 1975;4:163–72. [DOI] [PubMed] [Google Scholar]
  • 58.Fried PA, Watkinson B, Gray R. Growth from birth to early adolescence in offspring prenatally exposed to cigarettes and marijuana. Neurotoxicol Teratol. 1999;21:513–25. [DOI] [PubMed] [Google Scholar]
  • 59.Kuhle S, Allen AC, Veugelers PJ. Prevention potential of risk factors for childhood overweight. Can J Public Health Rev Can St Publique. 2010;101:365–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Cavlek T, Cavlek M, Bozikov J, Sturz B, Grsic K. Influende of parental smoking on children’s growth and weight at birth and age 6. Paediatr Croat. 2010;54:117–23. [Google Scholar]
  • 61.Taylor AE, Davey Smith G, Bares CB, Edwards AC, Munafo MR. Partner smoking and maternal cotinine during pregnancy: implications for negative control methods. Drug Alcohol Depend. 2014;139:159–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Mansi G, Raimondi F, Pichini S, Capasso L, Sarno M, Zuccaro P, et al. Neonatal urinary cotinine correlates with behavioral alterations in newborns prenatally exposed to tobacco smoke. Pediatr Res. 2007;61:257–61. [DOI] [PubMed] [Google Scholar]
  • 63.Laubenthal J, Zlobinskaya O, Poterlowicz K, Baumgartner A, Gdula MR, Fthenou E, et al. Cigarette smoke-induced transgenerational alterations in genome stability in cord blood of human F1 offspring. FASEB J. 2012;26:3946–56. [DOI] [PubMed] [Google Scholar]
  • 64.Kenny PJ. Common cellular and molecular mechanisms in obesity and drug addiction. Nat Rev Neurosci. 2011;12:638–51. [DOI] [PubMed] [Google Scholar]
  • 65.Haghighi A, Schwartz DH, Abrahamowicz M, Leonard GT, Perron M, Richer L, et al. Prenatal exposure to maternal cigarette smoking, amygdala volume, and fat intake in adolescence. JAMA Psychiatry. 2013;70:98–105. [DOI] [PubMed] [Google Scholar]
  • 66.Rolls ET. Taste, olfactory and food texture reward processing in the brain and obesity. Int J Obes. 2005;2011:550–61. [DOI] [PubMed] [Google Scholar]
  • 67.Bray GA, Popkin BM. Dietary fat intake does affect obesity! Am J Clin Nutr. 1998;68:1157–73. [DOI] [PubMed] [Google Scholar]
  • 68.Tappy L Thermic effect of food and sympathetic nervous system activity in humans. Reprod Nutr Dev. 1996;36:391–7. [DOI] [PubMed] [Google Scholar]
  • 69.Wrase J, Makris N, Braus DF, Mann K, Smolka MN, Kennedy DN, et al. Amygdala volume associated with alcohol abuse relapse and craving. Am J Psychiatry. 2008;165:1179–84. [DOI] [PubMed] [Google Scholar]
  • 70.Primeaux SD, York DA, Bray GA. Neuropeptide Y administration into the amygdala alters high fat food intake. Peptides. 2006;27:1644–51. [DOI] [PubMed] [Google Scholar]
  • 71.Boghossian S, Park M, York DA. Melanocortin activity in the amygdala controls appetite for dietary fat. Am J Physiol Regul Integr Comp Physiol. 2010;298:R385–R393. [DOI] [PubMed] [Google Scholar]
  • 72.Grabenhorst F, Rolls ET, Parris BA, d’Souza AA. How the brain represents the reward value of fat in the mouth. Cereb Cortex. 2010;20:1082–91. (N Y N 1991). [DOI] [PubMed] [Google Scholar]
  • 73.Trygg K, Lund-Larsen K, Sandstad B, Hoffman HJ, Jacobsen G, Bakketeig LS. Do pregnant smokers eat differently from pregnant non-smokers? Paediatr Perinat Epidemiol. 1995;9:307–19. [DOI] [PubMed] [Google Scholar]
  • 74.Llaquet H, Pichini S, Joya X, Papaseit E, Vall O, Klein J, et al. Biological matrices for the evaluation of exposure to environmental tobacco smoke during prenatal life and childhood. Anal Bioanal Chem. 2010;396:379–99. [DOI] [PubMed] [Google Scholar]
  • 75.Jacqz-Aigrain E, Zhang D, Maillard G, Luton D, André J, Oury JF. Maternal smoking during pregnancy and nicotine and cotinine concentrations in maternal and neonatal hair. BJOG Int J Obstet Gynaecol. 2002;109:909–11. [DOI] [PubMed] [Google Scholar]
  • 76.Toschke AM, Montgomery SM, Pfeiffer U, von Kries R. Early intrauterine exposure to tobacco-inhaled products and obesity. Am J Epidemiol. 2003;158:1068–74. [DOI] [PubMed] [Google Scholar]
  • 77.Venditti CC, Smith GN. Self-reported cigarette smoking status imprecisely quantifies exposure in pregnancy. Open J Obstet Gynecol. 2012;02:56–61. [Google Scholar]
  • 78.Pickett KE, Rathouz PJ, Kasza K, Wakschlag LS, Wright R. Self- reported smoking, cotinine levels, and patterns of smoking in pregnancy. Paediatr Perinat Epidemiol. 2005;19:368–76. [DOI] [PubMed] [Google Scholar]
  • 79.Talge NM, Mudd LM, Sikorskii A, Basso O. United States birth weight reference corrected for implausible gestational age estimates. Pediatrics. 2014;133:844–53. [DOI] [PubMed] [Google Scholar]
  • 80.Cole TJ, Freeman JV, Preece MA. British 1990 growth reference centiles for weight, height, body mass index and head circumference fitted by maximum penalized likelihood. Stat Med. 28. 1998;17:407–29. [PubMed] [Google Scholar]
  • 81.Voigt M, Schneider K, Jährig K. Analyse des Geburtengutes des Jahrgangs 1992 der Bundesrepublik Deutschland. Geburtshilfe Frauenheilkd. 1996;56:550–8. [DOI] [PubMed] [Google Scholar]
  • 82.Dobbins TA, Sullivan EA, Roberts CL, Simpson JM. Australian national birthweight percentiles by sex and gestational age, 1998- 2007. Med J Aust. 3. 2012;197:291–4. [DOI] [PubMed] [Google Scholar]
  • 83.Oken E, Kleinman KP, Rich-Edwards J, Gillman MW. A nearly continuous measure of birth weight for gestational age using a United States national reference. BMC Pediatr. 2003;3:6. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure Legends Supplemental Figures
NIHMS983919-supplement-Figure_Legends_Supplemental_Figures.docx (12.6KB, docx)
Supplemental Tables 1-3
NIHMS983919-supplement-Supplemental_Tables_1-3.docx (49.6KB, docx)
S_Figure 1
NIHMS983919-supplement-S_Figure_1.tiff (928.4KB, tiff)
S_Figure 2
NIHMS983919-supplement-S_Figure_2.tiff (938.7KB, tiff)
S_Figure 3
NIHMS983919-supplement-S_Figure_3.tiff (881.7KB, tiff)
S_Figure 4
NIHMS983919-supplement-S_Figure_4.tiff (797.8KB, tiff)
S_Figure 5
NIHMS983919-supplement-S_Figure_5.tiff (850.2KB, tiff)
S_Figure 6
NIHMS983919-supplement-S_Figure_6.tiff (900.5KB, tiff)
S_Figure 7
NIHMS983919-supplement-S_Figure_7.tiff (825.8KB, tiff)
S_Figure 8
NIHMS983919-supplement-S_Figure_8.tiff (804.9KB, tiff)

RESOURCES