Burrows, Tracy; Janet, Warren M.; Collins, Clare E.
Excessive energy intake leads to energy imbalance and is potentially a key modifiable component of successful obesity treatment (1). A solid evidence base for effective dietary intervention strategies to treat child obesity is lacking because studies describing the changes in dietary intake and food patterns secondary to interventions are scarce (2). In a systematic review of child obesity interventions that included a dietary component, most studies did not report changes in dietary intake secondary to the intervention (3). When child dietary intake changes were reported, it was usually limited to macronutrients (3) or changes in food group portions by categories based on energy and nutrient density, collapsed into a “traffic light” hierarchy in which red foods are energy-dense and nutrient-poor and green foods are low-energy and nutrient-dense (3). To date, specific food intake changes achieved in overweight children and adolescents who have successfully lost weight have not been described in detail. Assumptions must then be made about which specific food intake changes drive the reduction in total energy intake. Researchers have been encouraged to report dietary outcomes to facilitate the development of effective and practical food-based guidance for use within obesity prevention and treatment programs (4).
The primary outcomes of the Hunter Illawarra Kids Challenge Using Parent Support (HIKCUPS) trial have been reported. This was a multicenter randomized controlled trial with 1- and 2-year follow-up for the treatment of overweight and obese children in Australia (5,6).
The aim of this study was to describe a secondary outcome measure of HIKCUPS, which includes the changes in dietary intake of children participating in the HIKCUPS intervention from baseline to 2-year follow-up. This information will assist researchers and clinicians in the provision of practical dietary recommendations for treatment of overweight children.
SUBJECTS AND METHODS
The HIKCUPS trial methods have been published (7). Briefly, HIKCUPS recruited 165 prepubescent (5–9 years) overweight children defined by body mass index (BMI) z score (8) and their families between 2005 and 2006 from 2 regional areas of moderate socioeconomic status (Hunter and Illawara regions) of New South Wales, Australia, and randomly allocated them to one of the three 6-month intervention groups (10-week face-to-face group sessions, 3-monthly follow-up telephone calls) with follow-up extending to 2 years from baseline. Additional retention strategies included a postcard handwritten by the participants at their final group session, which was posted to them 3 months postintervention. Intervention groups were a parent-centered dietary-modification program (diet), a child-centered physical activity skill development program (activity), or both programs combined (activity + diet). Children's height, weight, and waist circumference were measured using standardized procedures (7) by 2 assessors; the z scores were calculated for BMI using the LMS method (LMS Growth program, NRC, Canberra, Australia).
Details of the dietary modification program have been reported (9). Briefly, it used food-based recommendations (10) to target a reduction in total daily energy intake by 10%, a reduction in fat intake to 30% of energy by reduced consumption of energy-dense, nutrient-poor foods, and increased consumption of vegetables, fruit, reduced-fat dairy products, and whole-grain breads and cereals. The program was conducted by accredited practising dietitians and was aimed at parents only. A widely available healthy eating brochure (11) was provided to the activity group at baseline to standardize their exposure to nutrition information.
Parents provided written informed consent and child assent was obtained before baseline assessments. Approval was obtained from the human research ethics committees of the University of Newcastle and University of Wollongong.
Measurement of Dietary Intake
Dietary intake was measured using the Australian Child and Adolescent Eating Survey (ACAES), a 135-item semiquantitative food-frequency questionnaire (FFQ), which assesses intake for the previous 6 months; the FFQ was completed by parental proxy. The FFQ has been validated for both child and adolescent self-report and parent report of child intake (12,13).
Questions on specific food items from the FFQ were aggregated categorically into food groups corresponding to the National Food Selection Guide in The Australian Guide to Healthy Eating(14). Each item was further recoded to subgroups to further characterize food selections. These categories were determined by research dietitians with expertise in pediatrics and in reporting dietary change. For example, energy-dense, nutrient-poor foods were broken down into sweetened drinks, packaged snacks, sweets, and carry-out foods.
The number of response categories for each food item in the FFQ varied from 4 to 8, depending on the likely usual consumption frequency of the item, and ranged from “never” to an appropriate number of times “per month,” “per week,” and/or “per day.” For this study, response categories were collapsed with “never” and “less than once per month” considered “rarely consumed,” responses ranging from “1 per month to 1 per week” considered “moderately consumed,” and “commonly consumed” for responses greater than or equal to twice per week. These frequencies were tallied as the percentage of total responses for each food at each time point (baseline 2 years). The foods presented are those that were targeted within the dietary intervention (10).
Twenty-one questions related directly to vegetables and 11 related to fruits. Daily portions of fruits and vegetables were calculated by summing the weight of FFQ food items and dividing by the portion size dictated in The Australian Guide to Healthy Eating (fruit 150 g and vegetables 75 g) (14). All of the other foods in ACAES were quantified using multiples of standardized portions from the 1995 Australian National Nutrition Survey of Children and Adolescents, which was the most recent survey at the time the ACAES was developed. Portions of other foods were calculated by summing the weight of food items based on their food group and dividing by the standard portion size.
Food items were categorized into 2 broad groups of low-energy, nutrient-dense (70 items) and energy-dense, nutrient-poor (45 items) (14). Low-energy, nutrient-dense items included breads and cereals, fruit, vegetables, dairy foods, and meat and/or alternatives. The energy-dense, nutrient-poor foods were high in fat and/or sugar and/or salt. They are not significant micronutrient sources and are considered “extras” in the diet; these extras included baked snacks (cakes, biscuits), confectioneries, and sweetened drinks. Additionally, 11 supplementary questions were asked about food behaviors, such as the frequency of consumption of breakfast, carry-out food items, and vegetables with the evening meal and eating while watching television.
Descriptive statistics were calculated and normality was checked for selected FFQ items at each time point. Linear mixed models were generated to determine the main effects by time and group by time, with age and sex as covariates. Post hoc analysis was conducted using the least-significant difference method, with no imputation of missing values for subjects. The effects were time, baseline, and 2 years. Chi-square tests were used to determine the differences in the frequency response categories during follow-up. Statistical significance was set at P < 0.05. Statistical analysis was completed using SPSS version 18 (SPSS Inc, Chicago, IL).
Complete FFQs were available for 160 children at baseline and 87 children (54%) at 2 years. Baseline characteristics of participants 58% girls, mean (range) age 8.0 years (5–9 years), BMI z score 2.89 (0.78–4.9) (15) have been reported (9). There were no statistically significant differences in retention rates in each intervention group: diet (n = 27), activity (n = 27), and diet + activity (n = 32). Dropouts at 24 months had higher BMI z scores at baseline (P < 0.001) (6).
Changes in nutrient intakes are reported in Table 1. Total energy intakes (kilocalorie and kilocalorie per kilogram) decreased significantly (P < 0.001) for all intervention groups from baseline to 2-year follow-up, whereas there were no statistically significant differences between groups; the diet-only group reported the largest reductions. Reported consumption of food groups (portions per day) are shown in Table 2. The greatest decreases from baseline were for portions of breads and cereals and dairy for the diet-only group, mean (95% confidence interval) change −1.1 (−2.2 to −0.8) and −0.5 (−0.7 to −0.5) portions per day, respectively; however, when examining food types within food groups, there was a significant increase at 2 years for all of the intervention groups in the proportion “commonly consuming” whole-grain breads from 32% to 43% (P < 0.05) and a concurrent decrease in white bread consumption from 53% to 41%. In addition, there was a significant (P < 0.01) increase from baseline to 2 years across all of the groups in the proportion of children commonly consuming reduced-fat milk, (38%–63%), whereas full-fat milk consumption decreased from 38% to 20%; this was not statistically significant. For all of the intervention groups, FFQ responses indicated that children consumed a wide variety of fruits and vegetables. The most popular vegetables, based on highest percentage commonly consumed at baseline, in descending order were potato, carrot, broccoli, pea, and corn, whereas the most commonly consumed fruits were apple/pear, banana, and orange/mandarin. Potato was the only vegetable to show a decreased reporting frequency from baseline to 2 years, which was a desirable change because potatoes are commonly served fried or with added fat and a reduction was targeted in the dietary intervention (9,10). The percentage of total energy attributable to breads and cereals, fruit, vegetables, meat and protein alternatives, and dairy, as a percentage of total energy, significantly increased (P < 0.001) from baseline (mean ± standard error of the mean) 57% ± 0.9% to 65% ± 1.1% at 2 years. The percentage of total energy from energy-dense, nutrient-poor foods at baseline decreased from 42% ± 1.0% to 35% ± 1.0% at 2 years (P < 0.001).
In the supplementary questions relating to food behaviors at baseline and 2 years, >95% of participants reported consuming breakfast at home. At baseline, 45% of families reported consuming vegetables with dinner >5 times per week, and this increased to 55% at 2 years. There was a significant decrease (P < 0.05) from baseline to 2 years (21% to 7%) in the proportion of families that reported consuming their evening meal in front of the television. At baseline, 45% of parents reported that carry-out foods were consumed “never” or “less than once per week,” and this increased to 56% at 2 years.
To our knowledge, this is the only one of a few studies to report dietary data as both nutrient and food group intakes of overweight children who participated in a well-designed and controlled treatment program with long-term follow-up. The present study provides practical information for use in future diet-related interventions and for clinicians, and it may help inform future food-based guidelines that facilitate improvement in eating habits and reductions in total energy intake in overweight children.
The reductions in energy intake suggest that parents seeking treatment for their overweight children can effectively facilitate changes in their child's dietary intake—in particular, a reduction in total energy intake. Those families allocated to the activity-only group that targeted the child only and only received brief standardized advice did achieve some positive dietary changes. For the primary outcome of weight change, however, targeting parents as the agents of change within the study led to a significantly greater reduction in BMI z score as compared with targeting the child only at the 2-year follow-up (6). Consumption of fruits and vegetables at both time points for all overweight children was higher than expected and this could be caused by a reporting bias or seasonal variation not captured by the FFQ, or that fruit and vegetable intakes are higher in Australian children compared with those in the United States (16–18).
Results from the 2001 to 2002 National Health and Nutrition Examination Survey (NHANES) reported food group level intakes for people ages 2 years or older (19) using multiple-pass 24-hour recalls. They identified the vegetables most commonly consumed by children, in descending order, as potato, tomato, broccoli, spinach, and carrot (19). This is similar to the present study, except for the inclusion of dark green vegetables in the NHANES population, which were reported as “consumed rarely.” Green vegetables could be a vegetable subgroup disliked by both younger children and those who are overweight in Australia. The most popular fruits in the present study also were similar to those in the NHANES (19), in which apple, pear, and banana were the 3 most commonly consumed fruits.
The significant increase in the percentage of energy intake from core foods postintervention and the decrease in energy-dense, nutrient-poor foods suggest that families adhered to the food-based guidelines. Overweight children in the HIKCUPS study were able to decrease their regular consumption of sweetened drinks (soda and fruit juice), potato crisps (chips), chocolate, and sweets. Fruit juice and soda intake decreased, particularly in children whose parents had reported them as “commonly consuming” these beverages at baseline, with reductions from 73% to 60% and 34% to 16%, respectively, at 2 years. It was noted that calcium intakes (milligram per day) appeared to decrease from baseline to 2-year follow-up; therefore, when implementing food-based guidelines, as in the present study, focus should be placed on swapping full-fat with reduced-fat dairy and not simply reducing the intake of dairy products, to protect bone health in the long term. Reductions in sweetened drinks may require more focus during the obesity treatment period and/or during follow-up to achieve desirable long-term changes in consumption.
The reported consumption of energy-dense, nutrient-poor foods in the present study decreased significantly from baseline (42% ± 1.0% of energy, or 1161 kcal) at the 2-year follow-up (35% ± 1.0% of energy, or 818 kcal). Although this exceeds the current national recommendations (14), this reduction is clinically important and is equivalent to approximately 2 servings of energy-dense, nutrient-poor foods per day. Intakes at baseline of energy-dense, nutrient-poor foods exceed 2 previous national Australian population-based surveys of children, measured by 24-hour recalls (16,20). This is likely to be the result of some classification differences at the specific item level for noncore foods and because the present subjects were overweight volunteers whose parents were seeking treatment. In 1995, in a similar age group of children, noncore foods contribute 34% of total energy (20), whereas more recently in 2007, this was reported to be 35% (17). In the United States, energy-dense, nutrient-poor foods contributed to approximately one-third of children's total energy intakes (21).
The changes in portions of noncore foods reported here are similar to those in a previous family-based intervention that used the traffic light diet resulted in a decrease in the number of servings of RED (energy dense) (22). That study limited the reporting to change in total portions of RED foods per day, whereas the present study has been able to articulate which components within the RED food group were amenable to change.
With regard to food behaviors, breakfast consumption, which is often a focus of obesity intervention programs (23), did not appear to be an issue in this sample of young overweight children. Decreases were seen for consumption of the evening meal while sitting in front of the television. This was interpreted as a positive outcome because it has been suggested that television viewing while eating reduces attention to food consumption and may influence the pace of eating, leading to an increase in total energy consumed. Additionally, increased exposure to food advertisements also may adversely influence food choices (24,25).
Limitations of the present study include the use of an FFQ, which is prone to overreporting (26), and a possible training effect on completion of the FFQs. Parental recall of child intake for those younger than 12 years has been shown to be relatively reliable (27), and we have demonstrated that parental report of child fruit and vegetable intake in the present study was positively correlated with fasting plasma carotenoid concentrations (13,25). Despite the limitations, it is believed that the changes in the reported frequency of consumption of food categories are real because of sustained improvements in weight change (6,9). Despite efforts to maximize study retention with flexible assessment times, including offering home visits, follow-up rates at 2 years are low to moderate, which introduce bias to the results. The retention rates are similar to those in other studies (28). The reported results may overestimate dietary change because those children who were followed up had lower BMIs than those who were not. The results may not be generalizable to an ethnically diverse population and/or across a wider socioeconomic gradient.
This study addresses an omission in the present literature reporting food intake changes in overweight and obese primary school-age children in response to a well-designed and controlled intervention. It is apparent that food habits are amenable to desirable changes, particularly in terms of reducing the consumption of energy-dense, nutrient-poor items, including sweetened drinks, packaged snacks, sweets, and carry-out foods while increasing the consumption of low-fat dairy and whole-grain breads. This provides evidence that the recommendation of food-based guidelines within treatment programs is feasible and effective in facilitating a reduction in total energy in overweight children.
1. Food, Nutrition, Physical Activity and Prevention of Cancer: A Global Perspective. Washington, DC: World Cancer Research Fund/American Institute for Cancer Research; 2007.
2. Collins C, Warren J, Neve M, et al. Measuring effectiveness of dietetic interventions in child obesity. Arch Pediatr Adolesc Med 2006; 160:906–922.
3. Gibson L, Peto J, Warren J, et al. Lack of evidence of diets for obesity for children: a systematic review. Int J Epidemiol 2006; 35:1544–1552.
4. Magarey A, Watson J, Golley R, et al. Assessing dietary intake in children and adolescents: considerations and recommendations for obesity research. Int J Pediatr Obes 2010 [Epub ahead of print].
5. Okely A, Collins C, Morgan P, et al. Multi-site randomized controlled trial of a child-centred physical activity program, a parent-centred dietary modification program, or both in obese children: The HIKCUPS study. J Paediatr 2010; 157:388–394.
6. Collins C, Okely A, Morgan P, et al. Long-term outcomes of the HIKCUPS multi-site randomized trial: efficacy of a parent-centred dietary-modification program, child-centred physical activity program or both in overweight children. Pediatrics 2011; 127:619–627.
7. Jones R, Okely A, Collins C, et al. The HIKCUPS trial: a multi site randomised controlled trial of a combined physical activity skill development and dietary modification program in overweight and obese children. BMC Public Health 2007;7:15.
8. Cole T, Bellizzi M, Flegal M, et al. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000;320:1240–3.
9. Burrows T, Warren J, Baur L, et al. Impact of a child obesity intervention on dietary intake and behaviours. Int J Obes 2008; 32:1481–1488.
10. Gehling R, Magarey A, Daniels L. Food based recommendations to reduce fat intake: an evidence based approach to the development of a family focused child weight management program. J Paediatr Child Health 2005; 41:112–118.
11. Easy Family Eating for Healthy Kids. West Ryde: Pfizer Australia; 2005: 1–8.
12. Watson J, Collins C, Sibbritt D, et al. Reproducibility and comparative validity of a food frequency questionnaire for Australian children and adolescents. Int J Behav Nutr Phys Act 2009;6:62.
13. Burrows T, Warren J, Colyvas C, et al. Validation of overweight children's fruit and vegetable intake using plasma carotenoids. Obesity 2008; 17:162–168.
14. Smith A, Kellet E, Schmerlaib Y. The Australian Guide to Healthy Eating. Canberra: Commonwealth Department of Health and Family Services; 1998.
15. Cole T, Freeman J, Preece M. British 1990 growth reference centiles for weight, height body mass index and head circumference fitted by maximum penalized likelihood. Stat Med 1998; 17:407–429.
16. 2007 Australian National Children's Nutrition and Physical Activity Survey—Main Findings. Clayton, Australia: CSIRO; 2008.
17. Dennison B, Rockwell H, Baker S. Fruit and vegetable intake in young children. J Am Coll Nutr 1998; 17:371–378.
18. Hudson C, Stotts C, Pruet J, et al. Parents diet related attitudes and knowledge, family fast food dollars spent and the relation to BMI and fruit and vegetable intake of their pre school children. South Online J Nurs Res 2005; 6:1–23.
19. Bachman C, Reedy J, Subar A, et al. Sources of food group intakes among the US population. J Am Diet Assoc 2008; 108:804–814.
20. Bell A, Kremer P, Magarey A, et al. Contribution of ‘noncore’ foods and beverages to the energy intake and weight status of Australian children. Eur J Clin Nutr 2005; 59:639–645.
21. Kant A. Reported consumption of low nutrient density foods by American children and adolescents. Arch Pediatr Adolesc Med 2003; 157:789–796.
22. Epstein L, Paluch R, Beecher MD, et al. Increasing healthy eating vs. reducing high energy dense foods to treat paediatric obesity. Obesity 2008; 16:318–326.
23. Rampersaud G, Pereira M, Girard B, et al. Breakfast habits, nutritional status, body weight and academic performance in children and adolscents. J Am Diet Assoc 2005; 105:743–760.
24. Zuppa J, Morton H, Mehta KP. Television food advertising: counter productive to childrens health? A content analysis using The Australian Guide to Healthy Eating. Nutr Diet 2003; 60:78–84.
25. Collins C, Watson J, Burrows T. Measuring dietary intake in children and adolescents in the context of overweight and obesity. Int J Obes 2010; 34:1103–1115.
26. Kaskoun M, Johnson R, Goran M. Comparison of energy intake by semi-quantitative food frequency questionnaire with total energy expenditure by the doubly labelled water method in young children. Am J Clin Nutr 1994; 60:43–47.
27. Burrows T, Martin R, Collins C. A systematic review of the validity of dietary assessment methods in children when compared with the method of doubly labelled water. J Am Diet Assoc 2010; 110:1501–1510.
28. Savoye M, Shaw M, Dziura J, et al. Effects of a weight management program on body composition and metabolic parameters in overweight children. JAMA 2007; 297:2697–2704.