- Split View
-
Views
-
Cite
Cite
Susanne Hvolgaard Mikkelsen, Lena Hohwü, Jørn Olsen, Bodil Hammer Bech, Zeyan Liew, Carsten Obel, Parental Body Mass Index and Behavioral Problems in Their Offspring: A Danish National Birth Cohort Study, American Journal of Epidemiology, Volume 186, Issue 5, 1 September 2017, Pages 593–602, https://doi.org/10.1093/aje/kwx063
- Share Icon Share
Abstract
Maternal obesity has been associated with increased risk of offspring behavioral problems. We examined whether this association could be explained by familial factors by comparing associations for maternal body mass index (BMI) with associations for paternal BMI. We studied 38,314 children born to mothers enrolled in the Danish National Birth Cohort during 1996–2002. Data on maternal BMI was collected at 15 weeks of gestation, and paternal BMI was assessed when the child was 18 months old. When the child was 7 years old, the Strengths and Difficulties Questionnaire was completed by the parents. We estimated odds ratios for behavioral problems in offspring born to overweight/obese parents, and we found that maternal BMI was associated with offspring behavioral problems. Maternal BMI of 25.0–29.9 was associated with a 33% (odds ratio = 1.33, 95% confidence interval: 1.13, 1.57) higher risk of total difficulties in offspring, and maternal BMI of ≥30.0 was associated with an 83% (odds ratio = 1.83, 95% confidence interval: 1.49, 2.25) higher risk. Paternal obesity was also associated with higher risk of offspring behavioral problems, but stronger associations were observed with maternal prepregnancy obesity. Our results suggest that part of the association between maternal BMI and behavioral problems can be accounted for by genetic and social factors, but environmental risk factors may also contribute to the etiology of behavioral problems.
Obesity is increasing worldwide, including in pregnant women, with the highest prevalence reported in US cohorts (1–10). In the United States, 60% women of reproductive age are overweight, and more than one-third are obese (11, 12). High maternal body mass index (BMI) is of particular concern, because it is associated with adverse health outcomes for both mothers and children. Obesity during pregnancy has been linked to adverse pregnancy outcomes such as neural tube defects (8, 10, 13, 14), as well as stillbirth and macrosomia (8, 9, 15, 16). Also, some studies suggest that maternal obesity during pregnancy predisposes offspring to obesity (8–10, 17), although the contribution of genetic and social factors might explain part of this linkage. During the last few decades, the recorded prevalence of common childhood mental disorders has increased (18–20). It has been suggested that this increase may be related to childhood obesity (21, 22) as obese children have a higher prevalence of mental problems (23, 24). However, a study from Finland found that childhood attention-deficit/hyperactivity disorder (ADHD) symptoms predicted adolescent obesity, rather than the reverse (25).
Furthermore, it has been suggested that the increased prevalence of childhood mental disorders may be a result of the increased prevalence of obesity in pregnant women, and maternal obesity has been associated with adverse neurodevelopmental outcomes in offspring, such as ADHD (1, 6, 26, 27), autism (27, 28), and emotional symptoms (14, 16, 17, 27). A fetal programming effect caused by maternal obesity has been suggested. The intrauterine environment accompanying obesity might be affected by suboptimal maternal diet, diabetes mellitus, inflammation, and obstetric complications (2, 8–10, 14, 29), which may alter fetal brain development (3, 10–14, 30, 31).
In contrast, some findings indicate that the association between maternal obesity and adverse offspring neurodevelopmental outcomes is a result of confounding by genetics or socioeconomic factors (10, 15, 28, 32, 33). Given the global epidemic of obesity, it is of public health interest to disentangle the biological associations of intrauterine exposure to maternal obesity from spurious associations due to confounding. In this study, we aimed to compare associations with maternal BMI and associations with paternal BMI. The rationale for comparing parental BMI associations is based on the assumption that the associations between maternal obesity and offspring behavioral problems are not caused by intrauterine programming but rather by genetic and socioeconomic factors. We expected to find similar associations for paternal BMI and maternal BMI. Additionally, we estimated the combined associations of parental BMI. If the association between maternal BMI and offspring behavioral problems is a result of confounding by genetic and socioeconomic factors, we expected to find the strongest associations when both parents were obese.
METHODS
Study population
We used data from the Danish National Birth Cohort (DNBC) (34). Women were recruited at their first pregnancy visit to the general practitioner, during the period of 1996–2002. Once enrolled, the women were offered 4 interviews; twice during pregnancy (at approximately 15 weeks and 30 weeks of gestation) and twice after delivery (when the child was approximately 6 months old and 18 months old). Self-reported data on maternal height and weight was collected at the first pregnancy interview, and data on paternal height and weight was collected at the interview when the child was 18 months old.
A total of 101,033 pregnancies were recruited to participate in the DNBC. We restricted the study to women who completed the first pregnancy interview and who had singleton pregnancies (n = 88,745). We excluded pregnancies with missing data on maternal height and weight (n = 1,444) and paternal height and weight (n = 27,077), resulting in a sample of 60,224 children.
When the child was 7 years of age, a follow-up questionnaire, including the Strength and Difficulties Questionnaire (SDQ) (35), was completed by the primary caregiver, usually the mother (>90%). We excluded those with incomplete/missing answers (n = 318), thus leaving us with SDQ information on 54,332 children. After linking with the 7-year follow-up data, the final sample size was 38,314 children with complete data on both parental height and weight and SDQ.
Exposure
Maternal BMI was calculated on the basis of self-reported information on prepregnancy weight and height from the first pregnancy interview. At the interview 18 months after birth, paternal weight and height were reported, often by the mother, and used to calculate paternal BMI. Parental BMIs were classified into 3 groups: underweight/normal weight (BMI <25.0), overweight (BMI 25.0–29.9), and obese (BMI ≥30.0), according to the WHO classification (36).
Outcome
The SDQ is a measure of emotional, behavioral, and social functioning in children and adolescents (35). The questionnaire comprises 25 questions generating 5 subscale scores. The calculation of each subscale is based on 5 questions. Answer categories were 0 (not true), 1 (somewhat true), or 2 (certainly true), and each subscale was scored from 0–10, with higher scores indicating worse behavior (except for prosocial behavior).
The present study was based on the parent-reported SDQ assessed for the child's behavior during the previous 6 months. We applied the recommended cutoff points for SDQ to classify children with abnormal scores indicating hyperactivity (girls: 6–10 and boys: 7–10), emotional symptoms (5–10), peer problems (3–10), conduct problems (4–10), and total difficulties score (girls: 12–40 and boys 14–40) (37).
In the DNBC, the SDQ included an impact supplement, which comprises 5 questions assessing whether the difficulties have an impact on the child's daily life (36). In this study we focused on the problem subscales: hyperactivity, emotional symptoms, peer problems, conduct problems, and the total difficulties score, with and without the impact supplement. The scores were dichotomized into “normal/borderline” and “abnormal” scores, using the above cutoff points.
Confounders and mediators
Potential confounders and mediators were chosen a priori based on previous studies. Measured covariates included parental age at birth (<25 years, 25–29 years, 30–34 years, >34 years), BMI of the opposite parent (underweight/normal (<25.0), overweight (25.0–29.9), obese (≥30)), parity (0, ≥1), maternal smoking during pregnancy (cigarettes per day: 0, 1–9, ≥10), parental socioeconomic status at first interview (high, medium, low; defined by job categories or education), maternal marital status at first interview (cohabiting/married, single), offspring birth year (1997–2003), sex (girl, boy), history of parental hyperactivity (yes/no), gestational weight gain (<10 kg, 10–20 kg, >20 kg), BMI category of the child at 7 years of age (underweight/normal, overweight, obese), gestational age (<32 weeks, 32–36 weeks, 37–41 weeks, >41 weeks), birth weight (<2,000 g, 2,000–3,000 g, 3,001–4,000 g, >4,000 g), and breastfeeding status (no/<6 months, ≥6 months). Maternal weight gain during pregnancy was adjusted for gestational age at the last available measure of weight during pregnancy and defined according to the 2009 Institute of Medicine Guidelines (38). BMI of the child was defined according to international sex and age cutoff points (39). Because gestational weight gain, gestational age, birth weight, breastfeeding, and childhood obesity may be mediators on the pathway between maternal BMI and offspring behavioral problems, adjustments for these covariates were not included in the final model. Nevertheless, we examined whether adjustment for these variables significantly changed the adjusted estimates. In addition we examined interaction by maternal smoking and socioeconomic status.
Statistical analysis
Parental BMI expressed categorically
We estimated odds ratios with corresponding 95% confidence intervals for the associations between parental BMI and offspring SDQ outcomes using multiple logistic regression analyses. Exposures were analyzed in models with and without adjustments, using parental BMI <25.0 as reference category. We also studied the association between combined parental BMI and total difficulties (maternal BMI <25.0 and paternal BMI <25.0; maternal BMI <25.0 and paternal BMI ≥25.0; maternal BMI ≥25.0 and paternal BMI <25.0; maternal BMI ≥25.0 and paternal BMI ≥25.0), with the combination of maternal BMI <25.0 and paternal BMI <25.0 as reference category.
We conducted sensitivity analyses to assess whether excluding children of underweight (BMI <18.5) mothers or fathers in the reference group changed the results. We examined whether missing data on paternal BMI was associated with maternal BMI, and whether excluding children with missing data on paternal BMI may have caused selection of the association between maternal BMI and offspring SDQ outcomes. In addition, among the 60,224 singletons with data on both maternal and paternal BMI, we examined whether parental BMI was associated with whether the parents answered the SDQ.
Because some of the covariates contained missing values, we performed subanalyses comparing unadjusted association estimates among the total study population with unadjusted association estimates among children with no missing covariates.
Parental BMI expressed continuously
To examine the full BMI range, parental BMI was analyzed continuously, expressed as cubic splines with 5 knots, which are the default knot values based on Harrell's recommended percentiles, and with BMI at 24.0 as the reference (40). Cubic spline models fit exposure and outcome relationships that are not necessarily linear.
Because few women (less than 3%) contributed 2 children to the cohort, we used robust variance estimators to calculate the 95% confidence intervals. P < 0.05 was considered statistically significant. All statistical analyses were conducted with Stata, version 12 (StataCorp LP, College Station, Texas).
RESULTS
A total of 19% of the children were born to overweight mothers, and 7% were born to obese mothers (Table 1). We found a correlation coefficient of 0.2 between maternal and paternal BMI. The percentages of children with abnormal scores were: 7.3% for emotional symptoms, 5.1% for conduct problems, 4.9% for hyperactivity, 4.2% for peer problems, 1.8% for total difficulties with impact, and 2.8% for total difficulties without impact.
Characteristic . | No. of Subjects . | % . | Maternal BMIa . | Paternal BMIa . |
---|---|---|---|---|
Mean (SD) . | Mean (SD) . | |||
All | 38,314 | 23.42 (4.07) | 25.09 (3.11) | |
Sex | ||||
Girls | 18,747 | 48.93 | 23.43 (4.06) | 25.09 (3.10) |
Boys | 19,567 | 51.07 | 23.41 (4.07) | 25.08 (3.13) |
Parity | ||||
0 | 17,650 | 46.07 | 23.33 (4.01) | 24.98 (3.13) |
≥1 | 20,580 | 53.71 | 23.50 (4.12) | 25.18 (3.10) |
Missing | 84 | 0.22 | 23.35 (3.90) | 25.26 (2.84) |
Maternal smoking during pregnancy (cigarettes/day) | ||||
0 | 29,594 | 77.24 | 23.44 (4.05) | 25.04 (3.06) |
1–9 | 3,687 | 9.62 | 23.18 (3.98) | 25.18 (3.17) |
≥10 | 4,960 | 12.95 | 23.48 (4.22) | 25.33 (3.36) |
Missing | 73 | 0.19 | 22.86 (3.86) | 24.89 (2.76) |
Socioeconomic statusb | ||||
High | 21,511 | 56.14 | 22.97 (3.64) | 24.84 (2.95) |
Middle | 13,771 | 35.94 | 23.92 (4.39) | 25.36 (3.20) |
Low | 2,878 | 7.51 | 24.31 (4.88) | 25.63 (3.63) |
Missing | 154 | 0.40 | 23.75 (4.97) | 25.17 (3.13) |
History of parental hyperactivity | ||||
No | 29,009 | 75.71 | 23.41 (4.05) | 25.05 (3.08) |
Yes | 3,281 | 8.56 | 23.74 (4.41) | 25.50 (3.37) |
Missing | 6,024 | 15.72 | 23.26 (3.92) | 25.05 (3.12) |
Marital status | ||||
Cohabiting/married | 37,689 | 98.37 | 23.42 (4.06) | 25.09 (3.11) |
Single | 542 | 1.41 | 23.23 (4.64) | 25.28 (3.55) |
Missing | 83 | 0.22 | 22.93 (3.82) | 24.95 (3.08) |
Maternal age at delivery, years | ||||
<25 | 2,728 | 7.13 | 23.67 (4.42) | 25.09 (3.42) |
25–29 | 14,748 | 38.55 | 23.46 (4.13) | 24.99 (3.12) |
30–34 | 14,765 | 38.60 | 23.39 (4.02) | 25.13 (3.04) |
>34 | 6,011 | 15.71 | 23.29 (3.85) | 25.23 (3.11) |
Paternal age at delivery, years | ||||
<25 | 1,171 | 3.09 | 23.59 (4.42) | 24.90 (3.47) |
25–29 | 10,002 | 26.39 | 23.41 (3.99) | 24.92 (3.13) |
30–34 | 15,278 | 40.31 | 23.46 (4.15) | 25.08 (3.09) |
>34 | 11,452 | 30.21 | 23.35 (3.97) | 25.26 (3.08) |
Offspring birth year | ||||
1997–1999 | 12,254 | 31.98 | 23.27 (3.97) | 25.02 (3.05) |
2000–2001 | 17,128 | 44.70 | 23.46 (4.09) | 25.09 (3.13) |
2002–2003 | 8,932 | 23.31 | 23.54 (4.14) | 25.17 (3.17) |
Offspring BMIc at age 7 years | ||||
Underweight/normal weight | 32,934 | 85.96 | 23.21 (3.87) | 24.95 (3.01) |
Overweight | 2,495 | 6.51 | 25.27 (4.84) | 26.31 (3.52) |
Obese | 141 | 0.37 | 27.35 (6.30) | 27.07 (3.58) |
Missing | 2,744 | 7.16 | 24.08 (4.77) | 25.56 (3.54) |
Maternal weight gain during pregnancy, kg | ||||
<10 | 4,588 | 11.97 | 25.97 (5.66) | 25.55 (3.49) |
10–20 | 22,484 | 58.68 | 22.97 (3.62) | 24.99 (3.02) |
>20 | 4,930 | 12.87 | 23.29 (3.42) | 25.17 (3.14) |
Missing | 6,312 | 16.47 | 23.27 (3.96) | 25.03 (3.12) |
Paternal BMI | ||||
<25.0 | 21,091 | 55.05 | 22.91 (3.72) | |
25.0–29.9 | 14,748 | 38.49 | 23.77 (4.14) | |
≥30.0 | 2,475 | 6.46 | 25.65 (5.28) | |
Maternal BMI | ||||
<25.0 | 28,333 | 73.9 | 24.80 (2.90) | |
25.0–29.9 | 7,190 | 18.8 | 25.67 (3.32) | |
≥30.0 | 2,791 | 7.3 | 26.52 (3.98) |
Characteristic . | No. of Subjects . | % . | Maternal BMIa . | Paternal BMIa . |
---|---|---|---|---|
Mean (SD) . | Mean (SD) . | |||
All | 38,314 | 23.42 (4.07) | 25.09 (3.11) | |
Sex | ||||
Girls | 18,747 | 48.93 | 23.43 (4.06) | 25.09 (3.10) |
Boys | 19,567 | 51.07 | 23.41 (4.07) | 25.08 (3.13) |
Parity | ||||
0 | 17,650 | 46.07 | 23.33 (4.01) | 24.98 (3.13) |
≥1 | 20,580 | 53.71 | 23.50 (4.12) | 25.18 (3.10) |
Missing | 84 | 0.22 | 23.35 (3.90) | 25.26 (2.84) |
Maternal smoking during pregnancy (cigarettes/day) | ||||
0 | 29,594 | 77.24 | 23.44 (4.05) | 25.04 (3.06) |
1–9 | 3,687 | 9.62 | 23.18 (3.98) | 25.18 (3.17) |
≥10 | 4,960 | 12.95 | 23.48 (4.22) | 25.33 (3.36) |
Missing | 73 | 0.19 | 22.86 (3.86) | 24.89 (2.76) |
Socioeconomic statusb | ||||
High | 21,511 | 56.14 | 22.97 (3.64) | 24.84 (2.95) |
Middle | 13,771 | 35.94 | 23.92 (4.39) | 25.36 (3.20) |
Low | 2,878 | 7.51 | 24.31 (4.88) | 25.63 (3.63) |
Missing | 154 | 0.40 | 23.75 (4.97) | 25.17 (3.13) |
History of parental hyperactivity | ||||
No | 29,009 | 75.71 | 23.41 (4.05) | 25.05 (3.08) |
Yes | 3,281 | 8.56 | 23.74 (4.41) | 25.50 (3.37) |
Missing | 6,024 | 15.72 | 23.26 (3.92) | 25.05 (3.12) |
Marital status | ||||
Cohabiting/married | 37,689 | 98.37 | 23.42 (4.06) | 25.09 (3.11) |
Single | 542 | 1.41 | 23.23 (4.64) | 25.28 (3.55) |
Missing | 83 | 0.22 | 22.93 (3.82) | 24.95 (3.08) |
Maternal age at delivery, years | ||||
<25 | 2,728 | 7.13 | 23.67 (4.42) | 25.09 (3.42) |
25–29 | 14,748 | 38.55 | 23.46 (4.13) | 24.99 (3.12) |
30–34 | 14,765 | 38.60 | 23.39 (4.02) | 25.13 (3.04) |
>34 | 6,011 | 15.71 | 23.29 (3.85) | 25.23 (3.11) |
Paternal age at delivery, years | ||||
<25 | 1,171 | 3.09 | 23.59 (4.42) | 24.90 (3.47) |
25–29 | 10,002 | 26.39 | 23.41 (3.99) | 24.92 (3.13) |
30–34 | 15,278 | 40.31 | 23.46 (4.15) | 25.08 (3.09) |
>34 | 11,452 | 30.21 | 23.35 (3.97) | 25.26 (3.08) |
Offspring birth year | ||||
1997–1999 | 12,254 | 31.98 | 23.27 (3.97) | 25.02 (3.05) |
2000–2001 | 17,128 | 44.70 | 23.46 (4.09) | 25.09 (3.13) |
2002–2003 | 8,932 | 23.31 | 23.54 (4.14) | 25.17 (3.17) |
Offspring BMIc at age 7 years | ||||
Underweight/normal weight | 32,934 | 85.96 | 23.21 (3.87) | 24.95 (3.01) |
Overweight | 2,495 | 6.51 | 25.27 (4.84) | 26.31 (3.52) |
Obese | 141 | 0.37 | 27.35 (6.30) | 27.07 (3.58) |
Missing | 2,744 | 7.16 | 24.08 (4.77) | 25.56 (3.54) |
Maternal weight gain during pregnancy, kg | ||||
<10 | 4,588 | 11.97 | 25.97 (5.66) | 25.55 (3.49) |
10–20 | 22,484 | 58.68 | 22.97 (3.62) | 24.99 (3.02) |
>20 | 4,930 | 12.87 | 23.29 (3.42) | 25.17 (3.14) |
Missing | 6,312 | 16.47 | 23.27 (3.96) | 25.03 (3.12) |
Paternal BMI | ||||
<25.0 | 21,091 | 55.05 | 22.91 (3.72) | |
25.0–29.9 | 14,748 | 38.49 | 23.77 (4.14) | |
≥30.0 | 2,475 | 6.46 | 25.65 (5.28) | |
Maternal BMI | ||||
<25.0 | 28,333 | 73.9 | 24.80 (2.90) | |
25.0–29.9 | 7,190 | 18.8 | 25.67 (3.32) | |
≥30.0 | 2,791 | 7.3 | 26.52 (3.98) |
Abbreviations: BMI, body mass index; SD, standard deviation.
a BMI calculated as weight (kg)/height (m)2.
b Socioeconomic information was derived from national registers at Statistics Denmark and based on the current or most recent job within 6 months or the type of education. High: management or jobs requiring higher education; middle: office workers, service workers, skilled manual workers, and working in the military; and low: unskilled workers and the unemployed. Women who could not be classified in this way (4.1%) were categorized according to their husband's socioeconomic status.
c BMI of the child was defined according to international sex and age cutoff points (39).
Characteristic . | No. of Subjects . | % . | Maternal BMIa . | Paternal BMIa . |
---|---|---|---|---|
Mean (SD) . | Mean (SD) . | |||
All | 38,314 | 23.42 (4.07) | 25.09 (3.11) | |
Sex | ||||
Girls | 18,747 | 48.93 | 23.43 (4.06) | 25.09 (3.10) |
Boys | 19,567 | 51.07 | 23.41 (4.07) | 25.08 (3.13) |
Parity | ||||
0 | 17,650 | 46.07 | 23.33 (4.01) | 24.98 (3.13) |
≥1 | 20,580 | 53.71 | 23.50 (4.12) | 25.18 (3.10) |
Missing | 84 | 0.22 | 23.35 (3.90) | 25.26 (2.84) |
Maternal smoking during pregnancy (cigarettes/day) | ||||
0 | 29,594 | 77.24 | 23.44 (4.05) | 25.04 (3.06) |
1–9 | 3,687 | 9.62 | 23.18 (3.98) | 25.18 (3.17) |
≥10 | 4,960 | 12.95 | 23.48 (4.22) | 25.33 (3.36) |
Missing | 73 | 0.19 | 22.86 (3.86) | 24.89 (2.76) |
Socioeconomic statusb | ||||
High | 21,511 | 56.14 | 22.97 (3.64) | 24.84 (2.95) |
Middle | 13,771 | 35.94 | 23.92 (4.39) | 25.36 (3.20) |
Low | 2,878 | 7.51 | 24.31 (4.88) | 25.63 (3.63) |
Missing | 154 | 0.40 | 23.75 (4.97) | 25.17 (3.13) |
History of parental hyperactivity | ||||
No | 29,009 | 75.71 | 23.41 (4.05) | 25.05 (3.08) |
Yes | 3,281 | 8.56 | 23.74 (4.41) | 25.50 (3.37) |
Missing | 6,024 | 15.72 | 23.26 (3.92) | 25.05 (3.12) |
Marital status | ||||
Cohabiting/married | 37,689 | 98.37 | 23.42 (4.06) | 25.09 (3.11) |
Single | 542 | 1.41 | 23.23 (4.64) | 25.28 (3.55) |
Missing | 83 | 0.22 | 22.93 (3.82) | 24.95 (3.08) |
Maternal age at delivery, years | ||||
<25 | 2,728 | 7.13 | 23.67 (4.42) | 25.09 (3.42) |
25–29 | 14,748 | 38.55 | 23.46 (4.13) | 24.99 (3.12) |
30–34 | 14,765 | 38.60 | 23.39 (4.02) | 25.13 (3.04) |
>34 | 6,011 | 15.71 | 23.29 (3.85) | 25.23 (3.11) |
Paternal age at delivery, years | ||||
<25 | 1,171 | 3.09 | 23.59 (4.42) | 24.90 (3.47) |
25–29 | 10,002 | 26.39 | 23.41 (3.99) | 24.92 (3.13) |
30–34 | 15,278 | 40.31 | 23.46 (4.15) | 25.08 (3.09) |
>34 | 11,452 | 30.21 | 23.35 (3.97) | 25.26 (3.08) |
Offspring birth year | ||||
1997–1999 | 12,254 | 31.98 | 23.27 (3.97) | 25.02 (3.05) |
2000–2001 | 17,128 | 44.70 | 23.46 (4.09) | 25.09 (3.13) |
2002–2003 | 8,932 | 23.31 | 23.54 (4.14) | 25.17 (3.17) |
Offspring BMIc at age 7 years | ||||
Underweight/normal weight | 32,934 | 85.96 | 23.21 (3.87) | 24.95 (3.01) |
Overweight | 2,495 | 6.51 | 25.27 (4.84) | 26.31 (3.52) |
Obese | 141 | 0.37 | 27.35 (6.30) | 27.07 (3.58) |
Missing | 2,744 | 7.16 | 24.08 (4.77) | 25.56 (3.54) |
Maternal weight gain during pregnancy, kg | ||||
<10 | 4,588 | 11.97 | 25.97 (5.66) | 25.55 (3.49) |
10–20 | 22,484 | 58.68 | 22.97 (3.62) | 24.99 (3.02) |
>20 | 4,930 | 12.87 | 23.29 (3.42) | 25.17 (3.14) |
Missing | 6,312 | 16.47 | 23.27 (3.96) | 25.03 (3.12) |
Paternal BMI | ||||
<25.0 | 21,091 | 55.05 | 22.91 (3.72) | |
25.0–29.9 | 14,748 | 38.49 | 23.77 (4.14) | |
≥30.0 | 2,475 | 6.46 | 25.65 (5.28) | |
Maternal BMI | ||||
<25.0 | 28,333 | 73.9 | 24.80 (2.90) | |
25.0–29.9 | 7,190 | 18.8 | 25.67 (3.32) | |
≥30.0 | 2,791 | 7.3 | 26.52 (3.98) |
Characteristic . | No. of Subjects . | % . | Maternal BMIa . | Paternal BMIa . |
---|---|---|---|---|
Mean (SD) . | Mean (SD) . | |||
All | 38,314 | 23.42 (4.07) | 25.09 (3.11) | |
Sex | ||||
Girls | 18,747 | 48.93 | 23.43 (4.06) | 25.09 (3.10) |
Boys | 19,567 | 51.07 | 23.41 (4.07) | 25.08 (3.13) |
Parity | ||||
0 | 17,650 | 46.07 | 23.33 (4.01) | 24.98 (3.13) |
≥1 | 20,580 | 53.71 | 23.50 (4.12) | 25.18 (3.10) |
Missing | 84 | 0.22 | 23.35 (3.90) | 25.26 (2.84) |
Maternal smoking during pregnancy (cigarettes/day) | ||||
0 | 29,594 | 77.24 | 23.44 (4.05) | 25.04 (3.06) |
1–9 | 3,687 | 9.62 | 23.18 (3.98) | 25.18 (3.17) |
≥10 | 4,960 | 12.95 | 23.48 (4.22) | 25.33 (3.36) |
Missing | 73 | 0.19 | 22.86 (3.86) | 24.89 (2.76) |
Socioeconomic statusb | ||||
High | 21,511 | 56.14 | 22.97 (3.64) | 24.84 (2.95) |
Middle | 13,771 | 35.94 | 23.92 (4.39) | 25.36 (3.20) |
Low | 2,878 | 7.51 | 24.31 (4.88) | 25.63 (3.63) |
Missing | 154 | 0.40 | 23.75 (4.97) | 25.17 (3.13) |
History of parental hyperactivity | ||||
No | 29,009 | 75.71 | 23.41 (4.05) | 25.05 (3.08) |
Yes | 3,281 | 8.56 | 23.74 (4.41) | 25.50 (3.37) |
Missing | 6,024 | 15.72 | 23.26 (3.92) | 25.05 (3.12) |
Marital status | ||||
Cohabiting/married | 37,689 | 98.37 | 23.42 (4.06) | 25.09 (3.11) |
Single | 542 | 1.41 | 23.23 (4.64) | 25.28 (3.55) |
Missing | 83 | 0.22 | 22.93 (3.82) | 24.95 (3.08) |
Maternal age at delivery, years | ||||
<25 | 2,728 | 7.13 | 23.67 (4.42) | 25.09 (3.42) |
25–29 | 14,748 | 38.55 | 23.46 (4.13) | 24.99 (3.12) |
30–34 | 14,765 | 38.60 | 23.39 (4.02) | 25.13 (3.04) |
>34 | 6,011 | 15.71 | 23.29 (3.85) | 25.23 (3.11) |
Paternal age at delivery, years | ||||
<25 | 1,171 | 3.09 | 23.59 (4.42) | 24.90 (3.47) |
25–29 | 10,002 | 26.39 | 23.41 (3.99) | 24.92 (3.13) |
30–34 | 15,278 | 40.31 | 23.46 (4.15) | 25.08 (3.09) |
>34 | 11,452 | 30.21 | 23.35 (3.97) | 25.26 (3.08) |
Offspring birth year | ||||
1997–1999 | 12,254 | 31.98 | 23.27 (3.97) | 25.02 (3.05) |
2000–2001 | 17,128 | 44.70 | 23.46 (4.09) | 25.09 (3.13) |
2002–2003 | 8,932 | 23.31 | 23.54 (4.14) | 25.17 (3.17) |
Offspring BMIc at age 7 years | ||||
Underweight/normal weight | 32,934 | 85.96 | 23.21 (3.87) | 24.95 (3.01) |
Overweight | 2,495 | 6.51 | 25.27 (4.84) | 26.31 (3.52) |
Obese | 141 | 0.37 | 27.35 (6.30) | 27.07 (3.58) |
Missing | 2,744 | 7.16 | 24.08 (4.77) | 25.56 (3.54) |
Maternal weight gain during pregnancy, kg | ||||
<10 | 4,588 | 11.97 | 25.97 (5.66) | 25.55 (3.49) |
10–20 | 22,484 | 58.68 | 22.97 (3.62) | 24.99 (3.02) |
>20 | 4,930 | 12.87 | 23.29 (3.42) | 25.17 (3.14) |
Missing | 6,312 | 16.47 | 23.27 (3.96) | 25.03 (3.12) |
Paternal BMI | ||||
<25.0 | 21,091 | 55.05 | 22.91 (3.72) | |
25.0–29.9 | 14,748 | 38.49 | 23.77 (4.14) | |
≥30.0 | 2,475 | 6.46 | 25.65 (5.28) | |
Maternal BMI | ||||
<25.0 | 28,333 | 73.9 | 24.80 (2.90) | |
25.0–29.9 | 7,190 | 18.8 | 25.67 (3.32) | |
≥30.0 | 2,791 | 7.3 | 26.52 (3.98) |
Abbreviations: BMI, body mass index; SD, standard deviation.
a BMI calculated as weight (kg)/height (m)2.
b Socioeconomic information was derived from national registers at Statistics Denmark and based on the current or most recent job within 6 months or the type of education. High: management or jobs requiring higher education; middle: office workers, service workers, skilled manual workers, and working in the military; and low: unskilled workers and the unemployed. Women who could not be classified in this way (4.1%) were categorized according to their husband's socioeconomic status.
c BMI of the child was defined according to international sex and age cutoff points (39).
Maternal prepregnancy BMI
We found a dose-response relationship between maternal BMI and offspring emotional problems, conduct problems, hyperactivity, peer problems, total difficulties, and total difficulties with impact (Table 2). Including gestational weight gain, gestational age, birth weight, breastfeeding, or offspring BMI in the model adjustments changed the estimates only slightly (Web Table 1, available at https://academic.oup.com/aje).
Behavioral Problems . | Overweightb . | Obeseb . | ||||||
---|---|---|---|---|---|---|---|---|
UOR . | 95% CI . | AORc . | 95% CI . | UOR . | 95% CI . | AORc . | 95% CI . | |
Maternal Prepregnancy BMIa | ||||||||
Emotional problems | 1.19 | 1.08, 1.32 | 1.15 | 1.03, 1.28 | 1.60 | 1.40, 1.82 | 1.46 | 1.26, 1.69 |
Conduct problems | 1.26 | 1.12, 1.41 | 1.18 | 1.04, 1.34 | 1.79 | 1.55, 2.08 | 1.55 | 1.31, 1.83 |
Hyperactivity | 1.33 | 1.20, 1.46 | 1.25 | 1.10, 1.42 | 1.71 | 1.47, 2.00 | 1.45 | 1.23, 1.73 |
Peer problems | 1.38 | 1.22, 1.56 | 1.31 | 1.14, 1.50 | 2.03 | 1.73, 2.37 | 1.73 | 1.45, 2.07 |
Total difficulties | 1.48 | 1.28, 1.72 | 1.33 | 1.13, 1.57 | 2.30 | 1.92, 2.76 | 1.83 | 1.49, 2.25 |
Total difficulties with impact | 1.43 | 1.19, 1.73 | 1.28 | 1.04, 1.57 | 2.47 | 1.98, 3.07 | 1.91 | 1.49, 2.45 |
Paternal BMIa | ||||||||
Emotional problems | 1.02 | 0.94, 1.10 | 1.03 | 0.94, 1.13 | 1.38 | 1.19, 1.59 | 1.27 | 1.08, 1.49 |
Conduct problems | 1.07 | 0.97, 1.18 | 1.00 | 0.90, 1.12 | 1.59 | 1.35, 1.88 | 1.18 | 0.97, 1.42 |
Hyperactivity | 1.11 | 1.01, 1.22 | 1.03 | 0.93, 1.15 | 1.73 | 1.47, 2.04 | 1.37 | 1.14, 1.65 |
Peer problems | 1.16 | 1.04, 1.29 | 1.11 | 0.99, 1.25 | 2.10 | 1.78, 2.48 | 1.72 | 1.43, 2.08 |
Total difficulties | 1.29 | 1.13, 1.46 | 1.18 | 1.03, 1.37 | 2.50 | 2.06, 3.03 | 1.77 | 1.41, 2.22 |
Total difficulties with impact | 1.39 | 1.18, 1.63 | 1.28 | 1.08, 1.53 | 2.55 | 2.00, 3.25 | 1.84 | 1.39, 2.44 |
Behavioral Problems . | Overweightb . | Obeseb . | ||||||
---|---|---|---|---|---|---|---|---|
UOR . | 95% CI . | AORc . | 95% CI . | UOR . | 95% CI . | AORc . | 95% CI . | |
Maternal Prepregnancy BMIa | ||||||||
Emotional problems | 1.19 | 1.08, 1.32 | 1.15 | 1.03, 1.28 | 1.60 | 1.40, 1.82 | 1.46 | 1.26, 1.69 |
Conduct problems | 1.26 | 1.12, 1.41 | 1.18 | 1.04, 1.34 | 1.79 | 1.55, 2.08 | 1.55 | 1.31, 1.83 |
Hyperactivity | 1.33 | 1.20, 1.46 | 1.25 | 1.10, 1.42 | 1.71 | 1.47, 2.00 | 1.45 | 1.23, 1.73 |
Peer problems | 1.38 | 1.22, 1.56 | 1.31 | 1.14, 1.50 | 2.03 | 1.73, 2.37 | 1.73 | 1.45, 2.07 |
Total difficulties | 1.48 | 1.28, 1.72 | 1.33 | 1.13, 1.57 | 2.30 | 1.92, 2.76 | 1.83 | 1.49, 2.25 |
Total difficulties with impact | 1.43 | 1.19, 1.73 | 1.28 | 1.04, 1.57 | 2.47 | 1.98, 3.07 | 1.91 | 1.49, 2.45 |
Paternal BMIa | ||||||||
Emotional problems | 1.02 | 0.94, 1.10 | 1.03 | 0.94, 1.13 | 1.38 | 1.19, 1.59 | 1.27 | 1.08, 1.49 |
Conduct problems | 1.07 | 0.97, 1.18 | 1.00 | 0.90, 1.12 | 1.59 | 1.35, 1.88 | 1.18 | 0.97, 1.42 |
Hyperactivity | 1.11 | 1.01, 1.22 | 1.03 | 0.93, 1.15 | 1.73 | 1.47, 2.04 | 1.37 | 1.14, 1.65 |
Peer problems | 1.16 | 1.04, 1.29 | 1.11 | 0.99, 1.25 | 2.10 | 1.78, 2.48 | 1.72 | 1.43, 2.08 |
Total difficulties | 1.29 | 1.13, 1.46 | 1.18 | 1.03, 1.37 | 2.50 | 2.06, 3.03 | 1.77 | 1.41, 2.22 |
Total difficulties with impact | 1.39 | 1.18, 1.63 | 1.28 | 1.08, 1.53 | 2.55 | 2.00, 3.25 | 1.84 | 1.39, 2.44 |
Abbreviations: AOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; UOR, unadjusted odds ratio.
a BMI was calculated as weight (kg)/height (m)2. Obese: BMI of ≥30.0; overweight: BMI of 25.0–29.9
b Reference group was mother or fathers with BMIs of <25.0.
c Adjustments for parental age, marital status, socioeconomic status, smoking, parity, birth year, sex, parental BMI, and parental hyperactivity.
Behavioral Problems . | Overweightb . | Obeseb . | ||||||
---|---|---|---|---|---|---|---|---|
UOR . | 95% CI . | AORc . | 95% CI . | UOR . | 95% CI . | AORc . | 95% CI . | |
Maternal Prepregnancy BMIa | ||||||||
Emotional problems | 1.19 | 1.08, 1.32 | 1.15 | 1.03, 1.28 | 1.60 | 1.40, 1.82 | 1.46 | 1.26, 1.69 |
Conduct problems | 1.26 | 1.12, 1.41 | 1.18 | 1.04, 1.34 | 1.79 | 1.55, 2.08 | 1.55 | 1.31, 1.83 |
Hyperactivity | 1.33 | 1.20, 1.46 | 1.25 | 1.10, 1.42 | 1.71 | 1.47, 2.00 | 1.45 | 1.23, 1.73 |
Peer problems | 1.38 | 1.22, 1.56 | 1.31 | 1.14, 1.50 | 2.03 | 1.73, 2.37 | 1.73 | 1.45, 2.07 |
Total difficulties | 1.48 | 1.28, 1.72 | 1.33 | 1.13, 1.57 | 2.30 | 1.92, 2.76 | 1.83 | 1.49, 2.25 |
Total difficulties with impact | 1.43 | 1.19, 1.73 | 1.28 | 1.04, 1.57 | 2.47 | 1.98, 3.07 | 1.91 | 1.49, 2.45 |
Paternal BMIa | ||||||||
Emotional problems | 1.02 | 0.94, 1.10 | 1.03 | 0.94, 1.13 | 1.38 | 1.19, 1.59 | 1.27 | 1.08, 1.49 |
Conduct problems | 1.07 | 0.97, 1.18 | 1.00 | 0.90, 1.12 | 1.59 | 1.35, 1.88 | 1.18 | 0.97, 1.42 |
Hyperactivity | 1.11 | 1.01, 1.22 | 1.03 | 0.93, 1.15 | 1.73 | 1.47, 2.04 | 1.37 | 1.14, 1.65 |
Peer problems | 1.16 | 1.04, 1.29 | 1.11 | 0.99, 1.25 | 2.10 | 1.78, 2.48 | 1.72 | 1.43, 2.08 |
Total difficulties | 1.29 | 1.13, 1.46 | 1.18 | 1.03, 1.37 | 2.50 | 2.06, 3.03 | 1.77 | 1.41, 2.22 |
Total difficulties with impact | 1.39 | 1.18, 1.63 | 1.28 | 1.08, 1.53 | 2.55 | 2.00, 3.25 | 1.84 | 1.39, 2.44 |
Behavioral Problems . | Overweightb . | Obeseb . | ||||||
---|---|---|---|---|---|---|---|---|
UOR . | 95% CI . | AORc . | 95% CI . | UOR . | 95% CI . | AORc . | 95% CI . | |
Maternal Prepregnancy BMIa | ||||||||
Emotional problems | 1.19 | 1.08, 1.32 | 1.15 | 1.03, 1.28 | 1.60 | 1.40, 1.82 | 1.46 | 1.26, 1.69 |
Conduct problems | 1.26 | 1.12, 1.41 | 1.18 | 1.04, 1.34 | 1.79 | 1.55, 2.08 | 1.55 | 1.31, 1.83 |
Hyperactivity | 1.33 | 1.20, 1.46 | 1.25 | 1.10, 1.42 | 1.71 | 1.47, 2.00 | 1.45 | 1.23, 1.73 |
Peer problems | 1.38 | 1.22, 1.56 | 1.31 | 1.14, 1.50 | 2.03 | 1.73, 2.37 | 1.73 | 1.45, 2.07 |
Total difficulties | 1.48 | 1.28, 1.72 | 1.33 | 1.13, 1.57 | 2.30 | 1.92, 2.76 | 1.83 | 1.49, 2.25 |
Total difficulties with impact | 1.43 | 1.19, 1.73 | 1.28 | 1.04, 1.57 | 2.47 | 1.98, 3.07 | 1.91 | 1.49, 2.45 |
Paternal BMIa | ||||||||
Emotional problems | 1.02 | 0.94, 1.10 | 1.03 | 0.94, 1.13 | 1.38 | 1.19, 1.59 | 1.27 | 1.08, 1.49 |
Conduct problems | 1.07 | 0.97, 1.18 | 1.00 | 0.90, 1.12 | 1.59 | 1.35, 1.88 | 1.18 | 0.97, 1.42 |
Hyperactivity | 1.11 | 1.01, 1.22 | 1.03 | 0.93, 1.15 | 1.73 | 1.47, 2.04 | 1.37 | 1.14, 1.65 |
Peer problems | 1.16 | 1.04, 1.29 | 1.11 | 0.99, 1.25 | 2.10 | 1.78, 2.48 | 1.72 | 1.43, 2.08 |
Total difficulties | 1.29 | 1.13, 1.46 | 1.18 | 1.03, 1.37 | 2.50 | 2.06, 3.03 | 1.77 | 1.41, 2.22 |
Total difficulties with impact | 1.39 | 1.18, 1.63 | 1.28 | 1.08, 1.53 | 2.55 | 2.00, 3.25 | 1.84 | 1.39, 2.44 |
Abbreviations: AOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; UOR, unadjusted odds ratio.
a BMI was calculated as weight (kg)/height (m)2. Obese: BMI of ≥30.0; overweight: BMI of 25.0–29.9
b Reference group was mother or fathers with BMIs of <25.0.
c Adjustments for parental age, marital status, socioeconomic status, smoking, parity, birth year, sex, parental BMI, and parental hyperactivity.
Children born to mothers who were overweight had a 25% (odds ratio (OR) = 1.25, 95% confidence interval (CI): 1.10, 1.42) higher risk of abnormal hyperactivity score when compared with maternal BMI <25.0, and children whose mothers were obese had a 45% (OR = 1.45, 95% CI: 1.23, 1.73) higher risk (Table 2). We found associations with maternal BMI in offspring emotional and conduct problems to be similar to those for hyperactivity (Table 2). Offspring of overweight mothers had a 31% (OR = 1.31, 95% CI: 1.14, 1.50) higher risk of peer problems, and offspring of obese mothers had a 73% (OR = 1.73, 95% CI: 1.45, 2.07) higher risk (Table 2).
Additionally, offspring of obese mothers had an 83% (OR = 1.83, 95% CI: 1.49, 2.25) higher risk of having an abnormal total difficulties score and almost twice the risk (OR = 1.91, 95% CI: 1.49, 2.45) of abnormal total difficulties with impact, when compared with offspring of mothers with a BMI <25.0 (Table 2).
We did not find any strong evidence for interaction by maternal smoking and socioeconomic status (Web Table 2).
Examining maternal BMI as a continuous variable suggested that increased maternal BMI was associated with higher risks of offspring emotional symptoms, conduct problems, peer problems, total difficulties, and total difficulties with impact (Figure 1). A small, but statistically significant, higher risk of offspring hyperactivity was observed with maternal BMI in the range approximately 25.0–35.0 (Figure 1).
Paternal BMI
We found paternal obesity to be associated with offspring emotional problems, hyperactivity, and peer problems (Table 2). Children with an obese father had a 37% (OR = 1.37, 95% CI: 1.14, 1.65) higher risk of abnormal hyperactivity score compared with those with paternal BMI of <25.0 (Table 2). Paternal overweight was not associated with offspring emotional problems, hyperactivity, or peer problems (Table 2). We did not observe any statistically significant association between paternal BMI and offspring conduct problems.
Compared with paternal BMI of <25.0, paternal overweight and obesity were associated with a 28% (OR = 1.28, 95% CI: 1.08, 1.53) and 84% (OR = 1.84, 95% CI: 1.39, 2.44) higher risk, respectively, of abnormal total difficulties with impairment (Table 2).
The results from examining paternal BMI continuously indicated that increasing paternal BMI was associated with higher risks of offspring behavioral problems (Figure 1).
Comparisons between maternal BMI and paternal BMI
Both maternal and paternal obesity were associated with higher risk of offspring behavioral problems, but stronger associations were observed for maternal obesity (Table 2). Moreover, maternal overweight was associated with offspring behavioral problems. In contrast, paternal overweight was not statistically significantly associated with the subscales: emotional problems, conduct problems, hyperactivity, or peer problems (Table 2).
Comparing results of maternal BMI and paternal BMI expressed continuously suggested similar estimates to those obtained when studying the risk of peer problems (Figure 1G and 1H) and total difficulties (Figure 1I and 1J). In contrast, the results indicated a stronger association of maternal BMI with emotional problems (Figure 1A and 1B), conduct problems (Figure 1C and 1D), and total difficulties with impact (Figure 1K and 1L), compared with paternal BMI. However, we found a tendency for paternal BMI to be more strongly associated with hyperactivity than was maternal BMI (Figure 1E and 1F).
Investigating combined parental BMI, we found the highest risk among offspring whose parents were both overweight or obese. Compared with offspring with the combination of maternal and paternal BMI <25.0, children whose parents both had BMIs of ≥25.0 had a 2-fold increased risk (OR = 2.06, 95% CI: 1.63, 2.61) of total difficulties with impact (Table 3).
Maternal BMIa . | Paternal BMIa and Difficulties Category . | |||||||
---|---|---|---|---|---|---|---|---|
Total Difficulties . | Total Difficulties With Impact . | |||||||
<25.0 . | ≥25.0 . | <25.0 . | ≥25.0 . | |||||
OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | |
Unadjusted Odds Ratios | ||||||||
<25.0 | 1.00 | Referent | 1.42 | 1.22, 1.66 | 1.00 | Referent | 1.57 | 1.30, 1.90 |
≥25.0 | 1.73 | 1.43, 2.09 | 2.24 | 1.89, 2.64 | 1.83 | 1.44, 2.33 | 2.37 | 1.92, 2.92 |
Adjusted Odds Ratiosb | ||||||||
<25.0 | 1.00 | Referent | 1.32 | 1.12, 1.57 | 1.00 | Referent | 1.49 | 1.21, 1.84 |
≥25.0 | 1.58 | 1.28, 1.95 | 1.92 | 1.60, 2.32 | 1.70 | 1.31, 2.20 | 2.06 | 1.63, 2.61 |
Maternal BMIa . | Paternal BMIa and Difficulties Category . | |||||||
---|---|---|---|---|---|---|---|---|
Total Difficulties . | Total Difficulties With Impact . | |||||||
<25.0 . | ≥25.0 . | <25.0 . | ≥25.0 . | |||||
OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | |
Unadjusted Odds Ratios | ||||||||
<25.0 | 1.00 | Referent | 1.42 | 1.22, 1.66 | 1.00 | Referent | 1.57 | 1.30, 1.90 |
≥25.0 | 1.73 | 1.43, 2.09 | 2.24 | 1.89, 2.64 | 1.83 | 1.44, 2.33 | 2.37 | 1.92, 2.92 |
Adjusted Odds Ratiosb | ||||||||
<25.0 | 1.00 | Referent | 1.32 | 1.12, 1.57 | 1.00 | Referent | 1.49 | 1.21, 1.84 |
≥25.0 | 1.58 | 1.28, 1.95 | 1.92 | 1.60, 2.32 | 1.70 | 1.31, 2.20 | 2.06 | 1.63, 2.61 |
Abbreviations: AOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; UOR, unadjusted odds ratio.
a BMI was calculated as weight (kg)/height (m)2.
b Adjustments for parental age, marital status, socioeconomic status, smoking, parity, birth year, sex, and parental hyperactivity.
Maternal BMIa . | Paternal BMIa and Difficulties Category . | |||||||
---|---|---|---|---|---|---|---|---|
Total Difficulties . | Total Difficulties With Impact . | |||||||
<25.0 . | ≥25.0 . | <25.0 . | ≥25.0 . | |||||
OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | |
Unadjusted Odds Ratios | ||||||||
<25.0 | 1.00 | Referent | 1.42 | 1.22, 1.66 | 1.00 | Referent | 1.57 | 1.30, 1.90 |
≥25.0 | 1.73 | 1.43, 2.09 | 2.24 | 1.89, 2.64 | 1.83 | 1.44, 2.33 | 2.37 | 1.92, 2.92 |
Adjusted Odds Ratiosb | ||||||||
<25.0 | 1.00 | Referent | 1.32 | 1.12, 1.57 | 1.00 | Referent | 1.49 | 1.21, 1.84 |
≥25.0 | 1.58 | 1.28, 1.95 | 1.92 | 1.60, 2.32 | 1.70 | 1.31, 2.20 | 2.06 | 1.63, 2.61 |
Maternal BMIa . | Paternal BMIa and Difficulties Category . | |||||||
---|---|---|---|---|---|---|---|---|
Total Difficulties . | Total Difficulties With Impact . | |||||||
<25.0 . | ≥25.0 . | <25.0 . | ≥25.0 . | |||||
OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | OR . | 95% CI . | |
Unadjusted Odds Ratios | ||||||||
<25.0 | 1.00 | Referent | 1.42 | 1.22, 1.66 | 1.00 | Referent | 1.57 | 1.30, 1.90 |
≥25.0 | 1.73 | 1.43, 2.09 | 2.24 | 1.89, 2.64 | 1.83 | 1.44, 2.33 | 2.37 | 1.92, 2.92 |
Adjusted Odds Ratiosb | ||||||||
<25.0 | 1.00 | Referent | 1.32 | 1.12, 1.57 | 1.00 | Referent | 1.49 | 1.21, 1.84 |
≥25.0 | 1.58 | 1.28, 1.95 | 1.92 | 1.60, 2.32 | 1.70 | 1.31, 2.20 | 2.06 | 1.63, 2.61 |
Abbreviations: AOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; UOR, unadjusted odds ratio.
a BMI was calculated as weight (kg)/height (m)2.
b Adjustments for parental age, marital status, socioeconomic status, smoking, parity, birth year, sex, and parental hyperactivity.
In addition, our results indicated a tendency toward higher risk with maternal BMI ≥25.0 and paternal BMI <25.0 compared with the opposite (maternal BMI <25.0 and paternal BMI ≥25.0) (Table 3), suggesting that maternal BMI ≥25.0 is a stronger predictor of offspring behavioral problems than is paternal BMI ≥25.0.
The results from the subanalysis examining whether missing values of covariates might have caused selection in the adjusted estimates indicated no statistically significant differences between the unadjusted estimates among the study population (n = 38,314) and the unadjusted estimates among children with no missing covariates (n = 32,163) (Web Table 3).
DISCUSSION
Main results
In this large national birth cohort with prospectively collected data, we found maternal prepregnancy overweight and obesity to be associated with higher risk of offspring behavioral problems. Because we also observed an association between paternal obesity and behavioral problems in the offspring, part of the association between maternal prepregnancy BMI and offspring behavioral problems may be attributed to familial factors rather than the intrauterine environment. Maternal prepregnancy BMI may be a proxy for other risk factors, such as genetics and socioeconomic factors.
Studies have found bipolar disorder (41), ADHD symptoms (42), and low intellectual ability (43) to be associated with increased risk of obesity. Psychiatric symptoms and obesity may share common causes and be linked genetically. In light of high heritability, the risk of psychiatric disorders may be transmitted from parents to offspring, contributing to parental obesity as well as offspring behavioral problems. Otherwise, compared with parents without behavioral problems, parents with behavioral problems may have different eating and/or exercise habits, with an increased likelihood of obesity. When we examined the BMI range we studied as continuous variables, our findings supported the hypothesis that part of the association between maternal BMI and offspring behavioral problems may be a result of unmeasured familial confounding. Very similar magnitudes of associations for maternal and paternal BMI with offspring behavioral problems suggest that the associations are driven by genetic and other shared family factors that are just as likely to be passed from the father to the offspring as from the mother to the offspring.
On the other hand, we found a trend of stronger association for maternal BMI than for paternal BMI, suggesting that maternal overweight/obesity may play a role in the risk of offspring behavioral problems beyond the family risk. This hypothesis is supported by our results from investigating combined parental BMI, suggesting maternal BMI to be a stronger predictor of offspring behavioral problems than paternal BMI, and we cannot reject an intrauterine programming effect on the developing fetus. Genetic factors causing a genetic predisposition in offspring might exert indirect risk through interplay with intrauterine exposure factors linked to maternal obesity, with gene-environment interaction contributing to the etiology of behavioral problems.
Comparison with other studies
Few published studies have explored the reasons for the association between maternal BMI and offspring behavioral problems (10, 15, 28, 32, 33). In a Swedish sibling study, investigators suggested that the association between maternal obesity and ADHD could be ascribed to time-stable unmeasured familial confounding, possibly of genetic origin (15), and to some extent our results support this hypothesis. The sibling analysis can be a powerful tool for accounting for confounding of unmeasured, stable, and within-family factors, but the design does not implicitly control for time-varying confounders (44–46). The underlying causes of maternal BMI changes between pregnancies may affect the outcome differently between siblings. Weight gain/loss between pregnancies may represent different biological processes, including stress and depression (47, 48). Factors that may affect the exposure and outcome differently between siblings must be carefully considered, but these factors may be difficult to control for. Therefore, the incorporation of maternal-paternal comparisons in our study contributes to a deeper understanding of the association between maternal BMI and offspring behavioral problems by adding information to the adjustments for family factors.
Strengths and limitations
The present study was based upon prospective data, and a major strength of our study was this availability to determine the plausibility of intrauterine influence by comparing associations for maternal and paternal BMI. Missing data on parental BMI may correlate with parental behavioral problems and obesity. In light of high heritability of psychiatric disorders, missing data may further correlate with offspring behavioral problems, and selection bias cannot be ruled out. We might have underestimated the association between parental BMI and offspring behavioral problems.
Among children with missing data on paternal BMI, and therefore not included in the analyses, maternal BMI was similar to maternal BMI among children included in the analyses (difference = 0.07, 95% CI: 0.00, 0.15). Also, in a subanalysis we examined whether excluding children with missing data on paternal BMI could have biased the associations for maternal BMI, but no statistically significant differences were observed between children with data on maternal BMI and children with data on both maternal and paternal BMI (Web Table 4).
Data on maternal BMI relied on self-reported information. In another study, based on data from a subcohort in the DNBC of 5,033 women, this self-reported anthropometric information was validated relative to weights and heights observed in antenatal care at approximately the ninth gestational week; BMI was underestimated with a mean of 0.66, with a slightly increased underreporting with increasing BMI, but the BMI categories agreed in 91.4% of the cases (49, 50). Because women with a higher BMI tend to report a lower weight compared with normal-weight women, misclassification affecting the results with bias towards the null cannot be ruled out (51, 52), but overall the reporting bias due to self-reported BMI is assumed to be minor in the DNBC (50), and with an average underestimation of 0.66 most women are expected to be represented in the true BMI category.
Additionally, bias between self-reported and measured anthropometrics has been widely described in the literature, arguing in favor of high validity of self-reported anthropometric data compared with measured values (53–56). Information on paternal weight and height was available in the DNBC only as reported by the mothers, and we have to take into account that mother-reported paternal anthropometrics might have caused nondifferential misclassification biasing the association towards the null. A study on the accuracy of informant reports on estimates of the weight and height of family members found a correlation coefficient of 0.94 between mother-reported paternal BMI and measured paternal BMI, concluding that mothers’ reports of paternal BMI were highly predictive of the actual BMI (57). Also the observed maternal-paternal correlation for BMI in this study was similar to that reported for spouses in studies with measured height and weight (58). Further we found the maternal-paternal correlation for BMI to be independent of maternal BMI group, which argues against a systematic under- or overestimation of paternal BMI related to maternal BMI.
Paternal BMI was measured when the child was 18 months old, but we did not expect the BMI for fathers to change significantly in the time period from the start of the pregnancy to 18 months after birth. If it changed, we assumed it affected neither the paternal BMI group nor SDQ outcome.
When comparing maternal and paternal BMI, we have to consider the level of paternal discrepancy. Rates of paternal discrepancy have been found to vary between studies from 0.8% to 30%, but in a general population the rate has been suggested to be 10% (59). First, in the DNBC the mothers were asked to report the weight and height of the biological fathers, and second, women who doubted the child's paternity were expected to be less likely to participate. We cannot reject information bias, but the bias of our estimates might be marginal.
Studies have shown that the psychometric properties of the SDQ, as well as its validity and reliability, are satisfactory (52, 60–65). We found that missing data on SDQ was associated with a higher parental BMI compared with complete SDQ data. Because we cannot reject the possibility that missing data on SDQ is associated with offspring behavioral problems, selection bias resulting in an underestimation of the association between parental BMI and offspring behavioral problems might be present (66).
Conclusion
We found both maternal prepregnancy BMI and paternal BMI to be associated with offspring behavioral problems. If the increased risk of offspring behavioral problems associated with maternal obesity is due to a fetal programming effect, it is expected to be marginal, and focusing only on weight loss among women before pregnancy—without ameliorating the background risks experienced by obese parents and their children—may not reduce offspring behavioral problems.
ACKNOWLEDGMENTS
Author affiliations: Section for General Medical Practice, Department of Public Health, Aarhus University, Aarhus, Denmark (Susanne Hvolgaard Mikkelsen, Lena Hohwü, Carsten Obel); Department of Clinical Epidemiology, Aarhus University, Aarhus, Denmark (Jørn Olsen); Department of Epidemiology, Fielding School of Public Health, University of California, Los Angeles, Los Angeles, California (Jørn Olsen, Zeyan Liew); and Section for Epidemiology, Department of Public Health, Aarhus University, Aarhus, Denmark (Jørn Olsen, Bodil Hammer Bech).
Conflict of interest: none declared.
REFERENCES
Author notes
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; BMI, body mass index; CI, confidence interval; DNBC, Danish National Birth Cohort; SDQ, Strengths and Difficulties Questionnaire.