Prenatal Diet and Child Growth at 18 Months
Abstract
OBJECTIVE: Our objective was to evaluate the effect of an antenatal dietary and lifestyle intervention in pregnant women who are overweight or obese on child outcomes at age 18 months.
METHODS: We conducted a follow-up study of children at 18 months of age who were born to women who participated in the Limiting Weight Gain in Overweight and Obese Women during Pregnancy to Improve Health Outcomes randomized trial. The primary follow-up study outcome was prevalence of child BMI z scores >85th percentile. Secondary study outcomes included a range of anthropometric measures, neurodevelopment, general health, and child feeding. Intention to treat principles were used in analyses, according to the treatment group allocated at randomization.
RESULTS: A total of 1602 children were assessed at age 18 months (lifestyle advice, n = 816; standard care, n = 786), representing 75.0% of the eligible sample (n = 2136). There were no statistically significant differences in the prevalence of child BMI z scores >85th percentile for children born to women in the lifestyle advice group, compared with the standard care group (lifestyle advice, 505 [47.11%] versus standard care, 483 [45.36%]; adjusted relative risk: 1.04; 95% confidence interval: 0.94 to 1.16; P = .45). There was no evidence of effects on child growth, adiposity, neurodevelopment, or dietary and physical activity patterns.
CONCLUSIONS: There is no evidence that providing pregnant women who were overweight or obese with an antenatal dietary and lifestyle intervention altered 18-month child growth and adiposity.
- aMD —
- adjusted mean difference
- aRR —
- adjusted relative risk
- ASQ —
- Ages and Stages Questionnaire
- CI —
- confidence interval
- GWG —
- gestational weight gain
- IRSD —
- Index of Relative Socio-Economic Disadvantage
- SEIFA —
- Socioeconomic Indexes for Areas
- SFTM —
- skinfold thickness measure
- WHO —
- World Health Organization
What’s Known on This Subject:
Maternal obesity is associated with an increased risk of pre-school obesity in the child. In tackling childhood obesity, the World Health Organization has recognized the need to integrate recommendations and health care interventions for women before conception and during pregnancy.
What This Study Adds:
There was no effect of an antenatal dietary intervention on child growth or adiposity at 18 months of age. We identified a significant rate of overweight and obesity in early childhood and high prevalence of at-risk obesogenic behaviors.
The World Health Organization (WHO) estimates 42 million children younger than the age of 5 years are overweight or obese globally,1 representing a significant burden of disease and associated health care costs.2 Early-life exposures contribute to an individual child’s risk of obesity, including parental obesity, infant birth weight and early growth, sleep duration, feeding practices, and sedentary activity or screen time.3 Maternal obesity is associated with an increased risk of pre-school obesity, with the risk ranging from a 1.6-fold4 to a more than sixfold increase5 compared with offspring of women of normal BMI, potentially creating a vicious cycle impacting successive generations.6
In tackling childhood obesity, WHO acknowledges the need for complementary strategies, including integration of health care interventions for women during pregnancy.7 Much research has been focused on antenatal interventions to limit gestational weight gain (GWG),8 with findings from an individual participant data meta-analysis involving more than 12 000 women across 36 randomized studies indicating a modest 0.7 kg reduction in GWG.8 However, most researchers have reported outcomes to birth only,8 with little attention to ongoing follow-up and impact on longer-term outcomes, including childhood obesity.
We have reported previously the effect of an antenatal intervention in women who are overweight or obese in significantly improving diet and physical activity9 and reducing the risk of infant birth weight >4 and 4.5 kg.10,11 In this article, we report the effect of the LIMIT antenatal dietary and lifestyle intervention on 18-month child outcomes.
Methods
The LIMIT (Limiting Weight Gain in Overweight and Obese Women During Pregnancy to Improve Health Outcomes) randomized trial was registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12607000161426), and the methods and clinical findings have been reported previously9–11 and are detailed in the Supplemental Information. Women with a singleton pregnancy, between 10 and 20 weeks’ gestation, and a BMI ≥25.0 were recruited and randomly selected to receive an antenatal dietary and lifestyle intervention (lifestyle advice) or standard antenatal care (standard care).10
Women randomly assigned to lifestyle advice received an intervention across pregnancy involving a combination of dietary, exercise, and behavioral strategies and goal setting, provided by a research dietician and trained research assistants.9,10 Dietary and physical activity information was consistent with Australian standards.12,13
We conducted follow-ups of children born to women who participated in the LIMIT randomized trial 18 months after birth (Supplemental Information). After ethics approval and after obtaining informed parental consent, a trained research assistant conducted each child assessment while blinded to the treatment group allocated at trial entry. The primary outcome was child BMI z score >85th percentile for age and sex.14
A range of secondary outcomes included the following:
Child anthropometry: by using an established and validated protocol,15,16 child length, weight, weight-for-length, and anthropometric measures (arm, thigh, waist and hip circumferences; biceps, triceps, subscapular, abdominal, suprailiac, and thigh skinfold thickness measures [SFTMs]) were obtained. Measurement of child skinfold thickness17,18 is considered a reliable and relatively noninvasive method of assessing fat distribution, having been correlated with more invasive measures.17,19–21 Furthermore, percentage fat as determined by SFTMs has been validated against dual-energy x-ray absorptiometry calculations of fat mass in both infants and children.17,22 Weight, length, BMI, weight-for-length, and head circumference measures were converted to z scores for age and sex by using WHO standards14,23;
Child neurodevelopment: the Ages and Stages Questionnaire (ASQ) completed by the child’s primary caregiver was used to assess neurodevelopment across 5 domains, reflecting communication, gross motor, fine motor, problem solving, and personal-social skills.24 Children with a score in any domain >2 SD below the mean were referred for medical assessment;
Child dietary intake, physical activity and sedentary behavior, and sleep patterns were assessed via questionnaire on the basis of Growing Up in Australia: the Longitudinal Study of Australian Children25 and included duration of breast or formula feeding; introduction of solid foods; the number of servings of fruits, vegetables, and milk consumed daily; the consumption of red meat and processed meats per week; and consumption of noncore foods, including salty snacks, fried potatoes, takeaway foods, soft drinks, and other “extra” foods.25 Caregivers were asked to estimate the hours per week spent playing outdoors, screen time and other sedentary behaviors, duration of overnight sleeping, and daytime sleeping25; and
Family food behaviors included whether family meals were eaten together, preparation of different meals for children, use of food to encourage behaviors, and use of a bottle at bedtime.25
Sample Size
The available sample size at 18 months was predetermined by the LIMIT trial, with 2212 women randomly selected.10 We anticipated an incidence in the standard care group of BMI z scores >85th percentile of 25%.26 A sample of 1350 children (75% follow-up) provided 80% power to detect a difference in incidence of BMI z scores >85th percentile of 6.4% and small but relevant differences of 0.15 to 0.16 SD in continuous outcomes.
Statistical Analysis
Data were analyzed by using intention to treat principles according to the treatment group the woman was randomly assigned to in pregnancy. Missing data were imputed by using the fully conditional specification (chained equations) method to create 100 complete data sets under the assumption that the data were missing at random. To make the missing-at-random assumption more plausible, a range of auxiliary variables, including maternal baseline and infant birth and 6-month follow-up measures (Supplemental Information, Supplemental Tables 7–10), were included in the imputation model. The results of imputed analyses were compared with those from complete-case analyses, and further sensitivity analyses were conducted for the primary outcome on the assumption that data were missing not at random, assuming both higher and lower incidence of z scores >85th percentile compared with observed data, in both or only 1 treatment group. Imputed outcomes included anthropometric measures, SFTMs, and Ages and Stages scores. No data were available to enable meaningful imputation of missing dietary and physical activity values.
Adjusted and unadjusted analyses were performed, with adjustment for stratification variables (maternal early-pregnancy BMI [25.0–29.9 vs ≥30.0], parity [0 vs ≥1], and center of birth), maternal socioeconomic status (Socioeconomic Indexes for Areas [SEIFA] Index of Relative Socio-Economic Disadvantage [IRSD] quintile27), age, and smoking status. Child outcomes other than z scores were additionally adjusted for sex and actual age at assessment. For binary outcomes, relative risks with 95% confidence intervals (CIs) were estimated by using log binomial regression. For continuous outcomes, differences in means with 95% CI were estimated by using linear regression. Secondary analyses were performed to test for effect modification by maternal BMI category measured at the women’s first prenatal visit. Statistical significance was set at P < .05 (2-sided) with no adjustment for multiple comparisons. All analyses followed a prespecified plan.
Results
There were 2136 live-born infants from the LIMIT randomized trial contributing data to the imputed analyses (Fig 1). In total, 1602 children (lifestyle advice, n = 816; standard care, n = 786) (representing 75.0% of the eligible sample) were assessed at 18 months. Baseline characteristics of eligible maternal participants (Table 1) were similar to the entire randomly assigned cohort and similar between treatment groups.10
Flow of participants.
Baseline Maternal Characteristics From Children Assessed at 18 Months of Age
The prevalence of child BMI z scores >85th percentile was not significantly different between treatment groups (lifestyle advice, 505 [47.11%] versus standard care, 483 [45.36%]; adjusted relative risk [aRR]: 1.04; 95% CI: 0.94 to 1.16; P = .45) and similarly for child BMI z scores >90th percentile (lifestyle advice, 407 [37.97%] versus standard care, 394 [37.01%]; aRR: 1.02; 95% CI: 0.90 to 1.15; P = .78) (Table 2). There were no statistically significant differences in child weight (lifestyle advice, 11.90 kg [±1.55] versus standard care 11.93 kg [±1.55]; adjusted mean difference [aMD]: −0.00; 95% CI: −0.14 to 0.13; P = .96), weight z score (lifestyle advice, 0.73 [±0.98] versus standard care, 0.74 [±0.99]; aMD: −0.02; 95% CI: −0.11 to 0.08; P = .73), length (lifestyle advice, 82.90 cm [±4.38] versus standard care, 83.06 cm [±4.25]; aMD: −0.05; 95% CI: −0.40 to 0.30; P = .79), length z score (lifestyle advice, 0.10 [±1.26] versus standard care, 0.16 [±1.20]; aMD: −0.05; 95% CI: −0.17 to 0.07; P = .43), weight-for-length ratio (lifestyle advice, 0.14 [±0.01] versus standard care, 0.14 [±0.01]; aMD: −0.00; 95% CI: −0.00 to 0.00; P = .66), or weight-for-length ratio z score (lifestyle advice, 0.92 [±1.13] versus standard care, 0.89 [±1.04]; aMD: 0.02; 95% CI: −0.09 to 0.13; P = .74). Abdominal circumference was statistically significantly greater among children born to women who received lifestyle advice (lifestyle advice, 48.92 cm [±3.81] versus standard care, 48.43 cm [±3.46]; aMD: 0.49; 95% CI: 0.12 to 0.85; P = .009), although the remaining body circumference measures did not differ.
Eighteen-Month Child Weight and Anthropometric Outcomes by Treatment Group
Subscapular SFTM (lifestyle advice, 6.77 mm [±1.86] versus standard care, 6.50 mm [±1.67]; aMD: 0.24; 95% CI: 0.03 to 0.46; P = .025) and sum of SFTMs (lifestyle advice, 55.96 mm [±12.65] versus standard care, 54.11 mm [±11.77]; aMD: 1.68; 95% CI: 0.12 to 3.24; P = .035) were statistically significantly greater among children born to women who received lifestyle advice (Table 3), although the remaining SFTMs did not differ.
Eighteen-Month Child SFTMs by Treatment Group
Child Ages and Stages total scores were not statistically significantly different between the treatment groups (lifestyle advice, 250.48 [±36.41] versus standard care, 247.64 [±37.60]; aMD: 2.70; 95% CI: −1.26 to 6.66; P = .18) (Table 4). Individual components of the Ages and Stages score did not differ significantly between the 2 groups (Table 4).
Eighteen-Month Child Neurodevelopmental Outcomes by Treatment Group Assessed by Using the ASQ
At 18 months, dietary patterns were similar, with ∼90% of children having ever been breastfed (lifestyle advice, 627 [90.48%] versus standard care, 591 [90.64%]; aRR: 1.01; 95% CI: 0.97 to 1.05; P = .69) and with the mean duration of breastfeeding being ∼8 months (lifestyle advice, 8.28 [±6.01] versus standard care, 8.69 [±6.10]; aMD: −0.26; 95% CI: −0.92 to 0.40; P = .44) (Table 5). There were no statistically significant differences in the number of servings per day of fruit (lifestyle advice, 1.94 [±0.99] versus standard care, 1.95 [±0.95]; aMD: −0.01; 95% CI: −0.11 to 0.10; P = .89), vegetables (lifestyle advice, 1.72 [±0.95] versus standard care, 1.66 [± 0.94]; aMD: 0.07; 95% CI: −0.03 to 0.17; P = .18), or milk (lifestyle advice, 2.18 [± 1.10] versus standard care, 2.14 [± 1.17]; aMD: 0.06; 95% CI: −0.06 to 0.18; P = .35). There were no statistically significant differences between the groups in consumption of red (lifestyle advice, 3.62 [± 3.26] versus standard care, 3.53 [± 3.18]; aMD: 0.09; 95% CI: −0.25 to 0.43; P = .59) or processed meat (lifestyle advice, 2.37 [± 2.65] versus standard care, 2.40 [± 2.87]; aMD: −0.06; 95% CI: −0.35 to 0.24; P = .71). Similarly, consumption of extras, salty snacks, and soft drinks did not differ significantly between the groups.
Eighteen-Month Child Dietary Patterns
Reported time per week engaged in physical activity outdoors (lifestyle advice, 16.24 hours [±10.16] versus standard care, 15.73 hours [± 9.78]; aMD: 0.50; 95% CI: −0.58 to 1.57; P = .37) and screen time (lifestyle advice, 15.29 hours [±13.40] versus standard care, 15.12 hours [± 13.81]; aMD: −0.04; 95% CI: −1.50 to 1.43; P = .96) did not significantly differ between the 2 groups (Table 6). The mean duration of sleep overnight was ∼11 hours (lifestyle advice, 10.98 hours [± 1.03] versus standard care, 10.98 hours [± 1.00]; aMD: 0.00; 95% CI: −0.11 to 0.11; P = .93), with no differences identified in the proportion of children having >1 daytime sleep (lifestyle advice, 115 [16.59%] versus standard care, 110 [16.92%]; aRR: 0.99; 95% CI: 0.78 to 1.25; P = .91).
Eighteen-Month Child Physical Activity and Sleep Patterns
Women and children from the lifestyle advice group were significantly less likely than those in the standard care group to have family meals together most of the time (lifestyle advice, 533 [77.02%] versus standard care, 534 [82.03%]; aRR: 0.93; 95% CI: 0.88 to 0.98; P = .01) or for all members of the family to eat the same food on most occasions (lifestyle advice, 517 [74.60%] versus standard care, 520 [79.75%]; aRR: 0.94; 95% CI: 0.89 to 0.99; P = .03). There were no significant differences in using food to encourage behavior (lifestyle advice, 204 [29.78%] versus standard care, 184 [28.31%]; aRR: 1.05; 95% CI: 0.90 to 1.24; P = .52). Few parents of children with BMIs >85th percentile perceived their child to be overweight (lifestyle advice, 7 [2.30%] versus standard care, 8 [2.95%]).
There was no evidence that the effect of the antenatal intervention was modified by maternal early-pregnancy BMI category for any of the reported outcomes (data not shown).
Discussion
In our findings, it was indicated that providing an antenatal dietary and lifestyle intervention for women who were overweight or obese was not associated with effects on child growth, adiposity, neurodevelopment, or diet and related behaviors at 18 months. The statistically significant differences in abdominal circumference, subscapular SFTM, and the sum of SFTMs were likely due to chance, particularly as the actual differences were small and unlikely to be clinically relevant.
We have conducted the largest randomized trial to date in which an antenatal dietary and lifestyle intervention in women who were overweight and obese was evaluated, the findings representing the largest, most extensive and complete follow-up of children to 18 months. Our methodology was robust, with accurate measurement of early pregnancy weight, height, and BMI; detailed maternal dietary and physical activity history; and consistent provision of the intervention to participants. In our follow-up, we adhered to a research-quality protocol with standardized assessment of anthropometric measures and consistent evaluation of dietary and physical activity, sedentary behavior, and sleep patterns, all of which are well-recognized early-life factors contributing to child overweight and obesity.3
Although only 75% of the available cohort was assessed and contributed data at 18 months, the baseline and clinical characteristics of women and children for whom data were available and who participated in the follow-up study were similar between the 2 randomly assigned treatment groups and were similar to the full randomly assigned cohort.10 Furthermore, in our analyses, we included all children eligible for 18-month follow-up (96.6% of those randomly selected) through multiple imputation to address missing data. Additionally, in our sensitivity analyses, we used data imputed under a range of missing-not-at-random scenarios and demonstrated that results remained consistent under various plausible assumptions about the magnitude and direction of the difference between missing and observed data. We therefore consider the risk of bias and any impact on the validity of our findings to be low.
We found no evidence of effect from a comprehensive dietary and lifestyle intervention provided during pregnancy for women who are overweight or obese on measures of child growth, adiposity, and neurodevelopment at 18 months. This is in contrast to our primary trial reports in which significant, albeit modest, changes in maternal diet and physical activity during pregnancy9 were associated with an 18% and 41% relative reduction in risk of infant birth weight >410 and 4.5 kg, respectively.11 Although this may indeed be reflective of a true lack of any persisting effect from modifications in maternal diet and birth weight on weight at 18 months, it may also indicate that other factors are of relatively greater importance in driving early ex-utero growth and development, particularly in a high-risk cohort of children in which the prevalence of BMI z scores >85th percentile was >45%.
We are aware of 2 randomized trials in which an antenatal dietary and lifestyle intervention has been provided28,29 and subsequent follow-up of children at ∼2 years of age has occurred.30–34 Tanvig et al30,31 conducted follow-up of children whose mothers participated in the Danish Life in Pregnancy randomized trial,28 reporting data for 150 children aged between 2.5 and 3.2 years, representing 42% of the randomly assigned cohort. There were no significant differences identified between intervention and control group children with regards to BMI,30,31 metabolic risk factors,30 or child anthropometric measures, including body circumferences and SFTMs as well as body composition determined by dual energy radiograph.31 These findings are broadly consistent with ours, although the incidence of overweight and obesity in this cohort of children was ∼9%,30,31 far lower than the 45% observed in our trial population at 18 months.
In the ROLO (randomized controlled trial of low glycemic index diet in pregnancy to prevent macrosomia) trial, researchers evaluated the effect of a low glycemic index diet in women at risk for infant macrosomia, recruiting women who had previously given birth to an infant with weight >4 kg, although not all women were overweight or obese (mean BMI, 26.8).29 Follow-up of children at 2 years of age was conducted in 35% of the eligible cohort.32–34 Although no specific data relating to between-group comparisons are presented, child anthropometric measures did not differ between the intervention and control groups.33
Our cohort of 18-month-old children is one at considerable risk, with high rates of overweight and obesity already evident, in addition to many high-risk obesogenic behaviors. Many participants did not meet the daily dietary recommendations for a child aged 0 to 2 years,35 consuming more than the recommended 1 serving of fruit (93%) and less than the recommended 2.5 servings of vegetables (48.8%) per day. Furthermore, children consumed both red and processed meats and extra foods ∼6 times per week, with additional high consumption of salty snacks and soft drinks. These dietary intakes far exceed the recommendations to limit consumption of such discretionary foods to less than once per week.35 In available survey data from the Australian Bureau of Statistics, it is suggested that children aged 2 to 18 years have similarly poor consumption of fruits and vegetables, with <1% meeting the daily recommended vegetable intake, although 31% met the recommended intake of fruit.36 Low consumption of fruit and vegetables37 and high consumption of discretionary foods,37 including soft drinks,38 from an early age contribute to the establishment of suboptimal eating habits that persist into adolescence.37
Although physical activity in young children can be difficult to quantify, current recommendations advise a minimum of 3 hours of physical activity each day for children aged 1 to 3 years, with no screen time for children <2 years.39 Participants in our study did not meet recommendations for outdoor play, spending on average 2.3 hours per day, but well exceeded screen time recommendations with an average of 2.2 hours per day. Similar findings have been reported by others,40 with evidence that maternal behaviors are important in both encouraging physical activity and reducing screen time.41
Overweight and obesity are highly prevalent in the general community, often resulting in a perception of “normality.” Despite >45% of children in this follow-up study being overweight or obese, only 2.6% of parents correctly identified this. This is not uncommon, with correct recognition of child overweight and obesity reported to vary between 7.5%42,43 and 50%.44 Contributing factors include the child’s age and the parent’s own weight.44
Conclusions
Providing an antenatal dietary and lifestyle intervention for pregnant women who were overweight and obese did not alter child growth and adiposity at 18 months. Ongoing follow-up of this cohort of children is warranted in view of the well-recognized association between high infant birth weight and subsequent obesity45 and the high rates of overweight, obesity, and obesogenic behaviors evident at such an early age.
Acknowledgments
We thank the 2212 women who participated in the LIMIT randomized trial and the 1602 parents and infants who contributed to the 18-month outcome data. The following persons in Adelaide, South Australia, participated in the 18-month follow-up of the LIMIT Trial: the coordinating team: J.M. Dodd, A. Deussen, J. Louise, A. Newman, L. Kannieappan, M. Kelsey, C. Sheppard, Z. Sui, N. Salehi, J. Koch, S. Hendrijanto, D. Post, M. Cooney, A. Webber, R. Bartley, C. Holst, K. Robinson, S. Zhang, and V. Ball; for statistical analyses: J. Louise; and the writing group: J.M. Dodd, J. Louise, A. McPhee, A. Deussen, J.A. Owens, and J.S. Robinson.
Footnotes
- Accepted June 14, 2018.
- Address correspondence to Jodie M. Dodd, MD, PhD, The University of Adelaide, Women’s and Children’s Hospital, 72 King William Rd, North Adelaide, SA 5006, Australia. E-mail: jodie.dodd{at}adelaide.edu.au
This trial has been registered with the Australian and New Zealand Clinical Trials Registry (http://anzctr.org.au) (identifier ACTRN12607000161426).
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by intramural funding from the University of Adelaide Discipline of Obstetrics and Gynaecology and the Robinson Research Institute.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
References
- Copyright © 2018 by the American Academy of Pediatrics
