Drawing causal inferences in epidemiologic studies of early life influences

Maternal pre-pregnancy BMI and offspring hyperactivity-inattention trajectories from 3 to 8 years in the EDEN birth cohort study

Evidence suggests obesity during pregnancy is associated with offspring attention-deficit hyperactivity disorder. However, studies have been limited to evaluating the association at a single age with inadequate data on important maternal lifestyle confounders and unmeasured familial confounding. The objective of this study was to examine the association between maternal pre-pregnancy body mass index (BMI) and child hyperactivity-inattention symptoms (HIS) at 3, 5 and 8 years. Data came from the EDEN mother-child cohort. Maternal pre-pregnancy BMI status (kg/m2) was calculated using pre-pregnancy weight and height (self-reported by mothers or measured by midwives). HIS were assessed by parental-report on the Strengths and Difficulties Questionnaire at 3, 5 and 8 years of age and used to derive developmental trajectories of HIS (n = 1428). Multivariate models were adjusted for confounders including socioeconomic status, maternal lifestyle behaviours (exercise, diet, smoking, alcohol), childcare and a stimulating home environment. Paternal BMI was used as a negative control. Compared to a normal pre-pregnancy BMI, pre-pregnancy maternal obesity was positively associated with increased odds of a high HIS trajectory between 3 and 8 years old in both unadjusted and adjusted logistic regression (adjusted odds ratio [aOR] 1.87 [95% CI 1.12, 3.12]). Pre-pregnancy overweight was not significantly associated after adjustment for confounders (aOR 1.32 [0.87, 2.01]). Maternal pre-pregnancy obesity, but not overweight, was associated with increased likelihood of a high HIS trajectory in children from 3 to 8 years old. This association persisted despite controlling for many important maternal lifestyle factors and paternal BMI. Further research is warranted to identify possible mediators involved.

Trajectories of child free sugars intake and dental caries - a population-based birth cohort study

OBJECTIVES

To investigate the association between trajectories of free sugars intake during the first five years of life and dental caries experience at five years.

METHODS

Data from the SMILE population-based prospective birth cohort study, collected at one, two and five years old, were used. A 3-days dietary diary and food frequency questionnaire were used to estimate free sugars intake (FSI) in grams. The primary outcomes were dental caries prevalence and experience (dmfs). The Group-Based Trajectory Modelling method was used to characterize three FSI trajectories ('Low and increasing'; 'Moderate and increasing'; and 'High and increasing'), which were the main exposures. Multivariable regression models were generated to compute adjusted prevalence ratios (APR) and rate ratios (ARR) for the exposure, controlling for socioeconomic factors.

RESULTS

The prevalence of caries was 23.3%, with a mean dmfs of 1.4, and a median of 3.0 among those who had caries. There were clear gradients of caries prevalence and experience by the FSI trajectories. The 'High and increasing' had an APR of 2.13 (95%CI 1.23-3.70) and ARR of 2.77 (95%CI 1.45-5.32) against the 'Low and increasing'. The 'Moderate and increasing' group had intermediate estimates. A quarter of the caries cases could have been prevented if the whole study sample had been in the 'Low and increasing' FSI trajectory.

CONCLUSION

A sustained, high trajectory of FSI from a young age was positively associated with child dental caries. Measures to minimise consumption of free sugars must commence early in life.

CLINICAL SIGNIFICANCE

The study has provided high level evidence to inform clinicians' decisions in promoting a healthy dietary pattern for young children.

Parental Weight Status and Offspring Behavioral Problems and Psychiatric Symptoms

OBJECTIVES

To assess relations of prepregnancy maternal and paternal obesity with offspring behavioral problems and psychiatric symptoms at 7-8 years in the Upstate KIDS study, a prospective cohort study.

STUDY DESIGN

Maternal body mass index (BMI) was calculated from prepregnancy height and weight provided in vital records or self-report at 4 months postpartum. Mothers reported paternal height and weight. At 7-8 years, mothers indicated if their children had been diagnosed with ADHD or anxiety (n = 1915). Additionally, children's behavior was measured with the Strengths and Difficulties Questionnaire at 7 years of age (n = 1386) and the Vanderbilt ADHD Diagnostic Parent Rating Scale at 8 years of age (n = 1484). Based on Strengths and Difficulties Questionnaire scores, we identified children with borderline behavioral problems. Adjusted risk ratios (aRR) and 95% CIs were estimated with robust multivariable Poisson regression.

RESULTS

Compared with children of mothers with a BMI of <25, children whose mothers had BMI 25-30, 30-35, and ≥35 kg/m2 had higher risks of reported ADHD (aRR, 1.14, 95% CI, 0.78-1.69; aRR, 1.96, 95% CI, 1.29-2.98; and aRR, 1.82, 95% CI,1.21-2.74, respectively). Risks of hyperactivity problems identified by the Strengths and Difficulties Questionnaire and a positive screen for inattentive or hyperactive/impulsive behavior with the Vanderbilt ADHD Diagnostic Parent Rating Scale were also higher with increasing maternal prepregnancy BMI. Paternal BMI was not associated with child outcomes.

CONCLUSIONS

Our findings suggest that maternal, rather than paternal, obesity is associated with maternal report of child ADHD diagnosis and inattentive or hyperactivity problems. Further research is needed to understand how maternal obesity might influence these behavioral changes during or after pregnancy.

Factors Influencing Early Feeding of Foods and Drinks Containing Free Sugars-A Birth Cohort Study

Early feeding of free sugars to young children can increase the preference for sweetness and the risk of consuming a cariogenic diet high in free sugars later in life. This study aimed to investigate early life factors influencing early introduction of foods/drinks containing free sugars. Data from an ongoing population-based birth cohort study in Australia were used. Mothers of newborn children completed questionnaires at birth and subsequently at ages 3, 6, 12, and 24 months. The outcome was reported feeding (Yes/No) at age 6–9 months of common foods/drinks sources of free sugars (hereafter referred as foods/drinks with free sugars). Household income quartiles, mother’s sugar-sweetened beverage (SSB) consumption, and other maternal factors were exposure variables. Analysis was conducted progressively from bivariate to multivariable log-binomial regression with robust standard error estimation to calculate prevalence ratios (PR) of being fed foods/drinks with free sugars at an early age (by 6–9 months). Models for both complete cases and with multiple imputations (MI) for missing data were generated. Of 1479 mother/child dyads, 21% of children had been fed foods/drinks with free sugars. There was a strong income gradient and a significant positive association with maternal SSB consumption. In the complete-case model, income Q1 and Q2 had PRs of 1.9 (1.2–3.1) and 1.8 (1.2–2.6) against Q4, respectively. The PR for mothers ingesting SSB everyday was 1.6 (1.2–2.3). The PR for children who had been breastfed to at least three months was 0.6 (0.5–0.8). Similar findings were observed in the MI model. Household income at birth and maternal behaviours were significant determinants of early feeding of foods/drinks with free sugars.

Parental Obesity and Early Childhood Development

BACKGROUND

Previous studies identified associations between maternal obesity and childhood neurodevelopment, but few examined paternal obesity despite potentially distinct genetic/epigenetic effects related to developmental programming.

METHODS

Upstate KIDS (2008-2010) recruited mothers from New York State (excluding New York City) at ∼4 months postpartum. Parents completed the Ages and Stages Questionnaire (ASQ) when their children were 4, 8, 12, 18, 24, 30, and 36 months of age corrected for gestation. The ASQ is validated to screen for delays in 5 developmental domains (ie, fine motor, gross motor, communication, personal-social functioning, and problem-solving ability). Analyses included 3759 singletons and 1062 nonrelated twins with ≥1 ASQs returned. Adjusted odds ratios (aORs) and 95% confidence intervals were estimated by using generalized linear mixed models accounting for maternal covariates (ie, age, race, education, insurance, marital status, parity, and pregnancy smoking).

RESULTS

Compared with normal/underweight mothers (BMI <25), children of obese mothers (26% with BMI ≥30) had increased odds of failing the fine motor domain (aOR 1.67; confidence interval 1.12-2.47). The association remained after additional adjustment for paternal BMI (1.67; 1.11-2.52). Paternal obesity (29%) was associated with increased risk of failing the personal-social domain (1.75; 1.13-2.71), albeit attenuated after adjustment for maternal obesity (aOR 1.71; 1.08-2.70). Children whose parents both had BMI ≥35 were likely to additionally fail the problem-solving domain (2.93; 1.09-7.85).

CONCLUSIONS

Findings suggest that maternal and paternal obesity are each associated with specific delays in early childhood development, emphasizing the importance of family information when screening child development.

Impact of parental obesity on neonatal markers of inflammation and immune response

Background/Objectives

Maternal obesity may influence neonatal and childhood morbidities through increased inflammation and/or altered immune response. Less is known about paternal obesity. We hypothesized that excessive parental weight contributes to elevated inflammation and altered immunoglobulin (Ig) profiles in neonates.

Subjects/Methods

In the Upstate KIDS Study maternal pre-pregnancy body mass index (BMI) was obtained from vital records and paternal BMI from maternal report. Biomarkers were measured from newborn dried blood spots (DBS) among neonates whose parents provided consent. Inflammatory scores were calculated by assigning one point for each of 5 pro-inflammatory biomarkers above the median and one point for an anti-inflammatory cytokine below the median. Linear regression models and generalized estimating equations were used to estimate mean differences (β) and 95% confidence intervals (CI) in the inflammatory score and Ig levels by parental overweight/obesity status compared to normal weight.

Results

Among 2974 pregnancies, 51% were complicated by excessive maternal weight (BMI>25), 73% by excessive paternal weight, and 28% by excessive gestational weight gain. Maternal BMI categories of overweight (BMI 25.0-29.9) and obese class II/III (BMI≥35) were associated with increased neonatal inflammation scores (β=0.12, 95% CI: 0.02, 0.21; p=0.02, and β=0.13, CI: −0.002, 0.26; p=0.05, respectively) but no increase was observed in the obese class I group (BMI 30-34.9). Mothers with class I and class II/III obesity had newborns with increased IgM levels (β=0.11, CI: 0.04, 0.17; p=0.001 and β=0.12, CI: 0.05, 0.19); p<0.001, respectively). Paternal groups of overweight, obese class I and obese class II/III had decreased neonatal IgM levels (β=−0.08, CI: −0.13,-0.03, p=0.001; β=−0.07, CI: −0.13, −0.01, p=0.029 and β=−0.11, CI:−0.19,-0.04, p=0.003, respectively).

Conclusions

Excessive maternal weight was generally associated with increased inflammation and IgM supporting previous observations of maternal obesity and immune dysregulation in offspring. The role of paternal obesity requires further study.

Determining the consequences of maternal obesity for offspring health

NEW FINDINGS

What is the topic of this review? Observational studies have highlighted the association of increasing maternal body mass index with offspring adiposity and the subsequent risk of cardiometabolic disorders in adulthood. The in utero environment has become a target for intervention in order to reduce the burden of obesity, despite the mechanistic pathways of this association remaining unclear. What advances does it highlight? This short review provides a critical appraisal of the recent literature, including biological pathways and strategies to address causal relationships. The global obesity epidemic has been causally linked to changes in diet and lifestyle. Observational data and animal studies have now highlighted associations between in utero environmental exposures and increased susceptibility to obesity and related cardiometabolic disorders in later life. Maternal body mass index has been reported to show an independent association with offspring adiposity from an early age and to play an important role in the predisposition to obesity and metabolic disease in later life. Thus, the in utero environment has been the focus of recent targeted interventions to improve public health. In this review, we summarize recent progress in this field, including the use of animal models to investigate mechanistic links between maternal obesity and offspring metabolic risk. We then assess the level of evidence and challenges in establishing causal inferences from present birth cohorts.

Fetal exposure to parental smoking and the risk of type 2 diabetes: Are lifestyle-related factors more important?

Current evidence suggests that although fetal exposure to smoking might increase the risk of type 2 diabetes in later life, lifestyle factors either during pregnancy or in adulthood may play a more important role. A ‘family‐based association study’ design might be helpful to detangle the relation among intrauterine exposure, lifestyle factors, and risk of diabetes in later life.

Modifiable early-life risk factors for childhood adiposity and overweight: an analysis of their combined impact and potential for prevention

BACKGROUND

Early life may be a "critical period" when appetite and regulation of energy balance are programmed, with lifelong consequences for obesity risk. Insight into the potential impact of modifying early-life risk factors on later obesity can be gained by evaluating their combined effects.

OBJECTIVE

The objective was to examine the relation between the number of early-life risk factors and obesity outcomes among children in a prospective birth cohort (Southampton Women's Survey).

DESIGN

Five risk factors were defined: maternal obesity [prepregnant body mass index (BMI; in kg/m(2)) >30], excess gestational weight gain (Institute of Medicine, 2009), smoking during pregnancy, low maternal vitamin D status (<64 nmol/L), and short duration of breastfeeding (none or <1 mo). Obesity outcomes examined when the children were aged 4 and 6 y were BMI, dual-energy X-ray absorptiometry-assessed fat mass, overweight, or obesity (International Obesity Task Force). Data were available for 991 mother-child pairs, with children born between 1998 and 2003.

RESULTS

Of the children, 148 (15%) had no early-life risk factors, 330 (33%) had 1, 296 (30%) had 2, 160 (16%) had 3, and 57 (6%) had 4 or 5. At both 4 and 6 y, there were positive graded associations between number of early-life risk factors and each obesity outcome (all P < 0.001). After taking account of confounders, the relative risk of being overweight or obese for children who had 4 or 5 risk factors was 3.99 (95% CI: 1.83, 8.67) at 4 y and 4.65 (95% CI: 2.29, 9.43) at 6 y compared with children who had none (both P < 0.001).

CONCLUSIONS

Having a greater number of early-life risk factors was associated with large differences in adiposity and risk of overweight and obesity in later childhood. These findings suggest that early intervention to change these modifiable risk factors could make a significant contribution to the prevention of childhood obesity.

Associations between infant feeding and the size, tempo and velocity of infant weight gain: SITAR analysis of the Gemini twin birth cohort

Objective:

Infant growth trajectories, in terms of size, tempo and velocity, may programme lifelong obesity risk. Timing of breastfeeding cessation and weaning are both implicated in rapid infant growth; we examined the association of both simultaneously with a range of growth parameters.

Design:

Longitudinal population-based twin birth cohort.

Subjects:

The Gemini cohort provided data on 4680 UK infants with a median of 10 (interquartile range=8–15) weight measurements between birth and a median of 6.5 months. Age at breastfeeding cessation and weaning were reported by parents at mean age 8.2 months (s.d.=2.2, range=4–20). Growth trajectories were modelled using SuperImposition by Translation And Rotation (SITAR) to generate three descriptors of individual growth relative to the average trajectory: size (grams), tempo (weeks, indicating the timing of the peak growth rate) and velocity (% difference from average, reflecting mean growth rate). Complex-samples general linear models adjusting for family clustering and confounders examined associations between infant feeding and SITAR parameters.

Results:

Longer breastfeeding (>4 months vs never) was independently associated with lower growth velocity by 6.8% (s.e.=1.3%) and delayed growth tempo by 1.0 (s.e.=0.2 weeks), but not with smaller size. Later weaning (⩾6 months vs <4 months) was independently associated with lower growth velocity by 4.9% (s.e.=1.1%) and smaller size by 102 g (s.e.=25 g).

Conclusions:

Infants breastfed for longer grew slower for longer after birth (later peak growth rate) but were no different in size, while infants weaned later grew slower overall and were smaller but the timing of peak growth did not differ. Slower trajectories with a delayed peak in growth may have beneficial implications for programming later obesity risk. Replication in cohorts with longer follow-up, alternative confounding structures or randomised controlled trials are required to confirm the long-term effects and directionality, and to rule out residual confounding.