Weight loss and the Effect on Stature in Children During a Residential Intervention Program
Abstract
Weight loss is generally high in residential weight-loss programs but the effect of a large weight loss on linear growth is not known. We report the weight loss and the influence on linear growth in a large group of children during a residential weight-loss program focusing on nutrition and physical activity. In a longitudinal noncontrolled intervention study of 990 overweight children (540 girls) attending the weight reduction program from 1990 to 2001 for about 11 weeks (age: 10–14 years, mean BMI-standard deviation score (SDS) at enrollment: 2.83) weight and height were measured initially and after end of treatment. Weekly measurements of height and weight were performed on 138 children. The children lost on average 9.4 kg, reduced their BMI by 4.5 kg/m2 and BMI-SDS by 0.98. In a multiple regression analysis (P ≤ 0.001) weight loss was higher in boys than girls (1.7 kg), higher if the weight was higher at admission (−0.192 kg/kg at baseline) and was positively associated with duration of stay (−80 g/day). Initially the boys' BMI-SDS was higher than the girls' BMI-SDS (P ≤ 0.05) but after 8 weeks of treatment the boys had lower BMI-SDS than the girls. There was no negative effect on linear growth during the treatment; on the contrary, linear growth accelerated during the stay as the average increase in height was 2.38 cm corresponding to 11.4 cm/year. In conclusion the children lost close to 1 kg/week during the stay without any negative effect on linear growth. The cause of the linear growth acceleration needs further investigation.
Introduction
The epidemic increase in Western countries in childhood obesity in the past decade makes the need for effective treatment more and more urgent (1). In Denmark the proportion of overweight children aged 14–16 years according to the internationally based cutoff points (2) has increased from 5.1 to 15.5% for boys and from 6.2 to 15.6% for girls during a period of 25 years from 1972 to 1997 (3). The prevalence of obese 14–16-year-old children increased from 0.4 and 0.9% to 2 and 2.6% for boys and girls, respectively, during the same period. The same trend has been reported for other European countries (4,5), but in the United States the raise in childhood obesity is markedly higher (6). Childhood obesity is associated with increased risk of hypertension, hypercholesterolemia, insulin resistance, diabetes, and psychosocial problems (7,8). Furthermore, childhood obesity may track into adulthood as long-term follow-up studies indicate (9) and higher childhood BMI is associated increased risk of coronary heart disease in adulthood (10). Early prevention and treatment in childhood are therefore important.
Most treatment programs for overweight and obese children focus on nutrition and exercise (1,11), and are often conducted as a home-based family intervention with group-based sessions sometimes with individual dietary counseling (12,13,14,15,16,17,18,19,20) or school-based interventions (12,21) but residential intervention programs have also been used (22,23,24,25,26,27,28,29). The results in home-based programs are often disappointing indicating that lifestyle leading to overweight is not easily changed in the family setting. The short-term effect of the family-based or school-based treatments is often a modest decrease in BMI of 2–8% (12,13,14,15,16,17,18). Higher reductions in BMI ranging from 18 to 34% have been reported for residential programs (24,27,28,29). These reports included only a small number of subjects and lasted for a longer period. Residential weight-loss programs of shorter duration have resulted in a decrease in BMI of 6–12% (22,25,26).
Childhood is a vulnerable period and possible adverse effects of obesity treatment must be considered. The most discussed adverse effect of weight-loss programs is the effect on growth (30). The diet used in weight-loss programs must therefore contain sufficient nutrients to promote growth, while concurrently decreasing excess body weight. It has therefore been recommended to monitor linear growth of the children during treatment for overweight (7).
In Denmark, one of the available treatments for overweight children is a residential treatment program at the institutions called “The Danish Christmas Seal Houses” (DCSH). There are four such institutions in Denmark. The running of the DCSH is financed by the income from the sale of annual Christmas Seals, reprints of old Christmas Seal editions and contributions from associations, business and private individuals.
The aim of this study was to investigate the effect of the weight-loss program on body weight and linear growth. We present the changes in weight, height, and BMI in overweight Danish children who participated in the treatment program for a period between 1994 and 2001 at the institution in Skælskør. In addition, the effect of gender, age, and duration of stay on weight loss is examined.
Methods and Procedures
Subjects
Children participating in the treatment were referred by family practitioners or school doctors. Inclusion criteria was an overweight percent >20. Children were not eligible to participate if they had psychiatric problems or physical disabilities preventing them from participating in the physical activity program. The stay was free of charge for the families. The children attended school at the institution. They were allowed to return home for public holidays, and every second weekend. In this analysis, 10–14-year-old children who had stayed at the institution between 60 and 90 days from 1994 to 2001 were included in the dataset (core dataset). In a subgroup of children, who stayed at the institution from 1999 to 2001, weekly weight and height measurements were available and used to investigate anthropometric changes over time.
Treatment program
The treatment was composed of three types of intervention: diet, physical activity, and education. The dietary part of the program consisted of moderate dietary restriction (5600–5800 kJ/day to 1333–1380 kcal/day) based on the concept of the traffic light diet system (31,32,33) including three low-fat meals with whole grain bread, fish or lean meat and vegetables, and snack between meals with whole grain bread and/or fruits. The children should taste all food items to reduce fastidiousness. There were no formal nutritional education lessons; instead it was integrated into the daily life. Before each meal, the food, how it was prepared, and the recommended portion size were announced by the personnel. The meals were supervised to avoid excess food intake.
The exercise component consisted of compulsory and voluntary physical activity. The compulsory exercise in a week included 3 × 1.3 km running, 15–20 km walking and swimming or aerobics for 1 h and all three activities should be completed in a week. The mandatory activities were supervised to secure participation of all children. The children were also encouraged to participate in voluntary physical activity including running, cycling, or swimming and the children were awarded according to a point system for voluntary activities. Furthermore, the children were encouraged to play football, beach volley, and other physical games to reduce sedentary behavior. Adult support to behavior modification and increase of the children's self-esteem is part of the objective of the treatment programs; however, these results have not been formally evaluated in this study.
Measurements (height, weight, and BMI)
Stature was measured to the nearest 1 cm using a statiometer (Seca, Hamburg, Germany). Body mass was measured to the nearest 0.1 kg on a digital balance scale (Seca, delta Model 707) before breakfast with the subjects lightly dressed and without shoes. BMI was calculated as body weight in kilograms divided by the square of height in meters. Because normal values of BMI vary substantially with age and gender, individual BMI values were converted to age- and gender-specific standard deviation score (SDS) values. Calculation of weight–, height–, and BMI-SDS values was based on reference data for Danish children (34).
Statistical analysis
Results are expressed as mean ± s.d. All data were analyzed using SPSS software (version 13.0; SPSS, Chicago, IL). Statistical significance was taken as P < 0.05. Paired t-tests were used to assess change over the treatment period. The difference in variables (age, duration of stay, baseline- and end-values and change in anthropometric measures) between genders was examined by independent samples t-test. Bivariate correlations were examined by Pearson's correlation coefficient or by Spearman's rho.
Dependent variables, i.e., change in weight and BMI, were analyzed by multiple linear regression analysis using General Linear Models with gender as fixed factor and baseline values, duration of stay and age as covariates. Linear regression was used to fit data examining the change in BMI during the treatment period.
Results
From 1994 to 2001 there were 1,370 10–14-year-old children, who had stayed at the institution in Skælskør. Of these, 313 and 67 children were excluded from the analysis because the stay was <60 days or >90 days, respectively. The limits were set to obtain a large group as homogeneous as possible regarding duration of treatment.
A total of 990 children were included (540 girls). The mean age was 12.3 ± 1.1 years, and the mean stay was 76 days with no difference between the genders. Baseline values for weight and BMI were higher in girls than in boys, whereas BMI-SDS was higher in boys (Table 1). There were no significant differences between the sexes for height but boys had higher height-SDS.
 |
At the end of the stay the average body weight, body weight-SDS, BMI and BMI-SDS were reduced significantly (all P ≤ 0.001). All children succeeded in loosing weight during the treatment with a mean weight loss of 9.4 kg. The distribution of the weight loss showed that 15% of the children lost <6 kg, 63% of the children lost between 6 and 12 kg, while 22% lost >12 kg (1). The loss in weight, BMI, and BMI-SDS were significantly higher in boys than in girls (P ≤ 0.001).
Frequency distribution of the children's change in weight during the treatment program according to gender.
During the treatment there was a significant increase in stature, with boys gaining 2.6 cm and girls 2.2 cm over the treatment period. This height velocity was higher than the average linear growth velocity during this age which is reflected in the significant increase in height-SDS score of 0.16 during the period. Calculated as increment per year the distribution of growth velocity showed that 22% of the children had an increase in stature <8 cm/year of which 7% did not increase in stature, when measured with 1-cm increments. About half of the children (55%) had a growth velocity between 8 and 16 cm/year whereas 22% grew >16 cm/year. The mean growth rate for boys and girls was 12.5 and 10.4 cm/year, respectively. The significantly smaller increase in stature in girls can be explained by the decrease in linear growth velocity for girls during the past years of the 10–14-year age range included in the study. Divided into age groups, the increment for boys during the treatment period was 2.3, 2.4, 2.8, 2.6, and 2.7 cm for the 10-, 11-, 12-, 13-, and 14-year age groups, respectively. For girls the same values were 2.5, 2.5, 2.3, 1.8, and 1.8 cm.
There was no correlation between the change in height or change in height-SDS and change in weight (r = −0.053, P = 0.098; r = 0.041, P = 0.202, respectively).
The change in weight and BMI were examined for influence of baseline values, age, gender and duration of stay (days) by multiple linear regression analysis (Table 2). With these few variables it was possible to explain as much as 57 and 47% of the variation in change in weight and BMI, respectively. The boys lost 1.7 kg more than girls. The change in weight and BMI were higher in younger children as an increase in age of 1 year corresponded to 0.4 kg and 0.14 kg/m2 lower weight change and BMI change, respectively. Though the range for duration of stay was confined to 60–90 days, an effect was observed. Even after 60 day of treatment the mean weight loss during the last 30 days was 0.08 kg/day. The baseline value was the most powerful covariate for both BMI and weight. The most obese children lost most weight corresponding to one additional kilogram at baseline resulted in 0.192 kg higher weight loss.
 |
In a subpopulation of 138 children (71 girls) who stayed at the institution from 1999–2001, weekly weight and height measurements were available. The mean age at baseline was 12.6 ± 1.2 years and the mean duration of the stay was 77 ± 6 days. One-hundred and seventeen children had stayed for all 12 weeks. There was no difference between genders for age or duration of stay. The values for the weight measures were significantly different at start and at end of treatment (P ≤ 0.001).
The time course of the weight loss measured by BMI-SDS is depicted for each gender in 2. The change in BMI-SDS was significantly higher for the boys (P ≤ 0.05). Initially the boys showed the highest values, but after 8 weeks the curves cross, so the boys end up with lower BMI-SDS than the girls.
Time course of BMI-standard deviation score (SDS) measured each week during treatment in a subpopulation of 138 children. Data are presented as mean +s.d. (boys); mean –s.d. (girls). Data for regression line (not shown): of b = −0.077, R = 0.993 (boys); b = −0.045, R = 0.975 (girls).
In the 8 years from 1994 to 2001 there was an increase in baseline values for weight measures equal to an increase in BMI-SDS of 0.11. There was also a small but significant reduction in weight loss obtained during the treatment program (P ≤ 0.001), i.e., from 1994 to 2001 the weight loss measured as the mean change in weight and mean change in BMI-SDS was reduced by 0.172 kg/year, and 0.004/year, respectively. Hence, as expected, the mean values for change in BMI-SDS presented in 2 were numerically lower than the mean values for the whole period.
Discussion
The results showed that all the children aged 10–14 years lost weight during the residential treatment program consisting of moderate dietary restriction, physical activity, and psychological support for 60–90 days. Large reductions in weight measures were observed, i.e., the mean weight and BMI were reduced by 14 and 16%, respectively. As this age is also a period of rapid linear growth the weight loss is remarkably large. This indicates that the treatment is very effective and compared to most other treatment programs for children of approximately the same duration the weight loss was higher (13,14,15,16,17,18,22,25,26). Three of these were residential programs but of slightly shorter durations (4–8 weeks) (22,25,26). The residential programs for which higher reductions in BMI have been reported (24,27,28,29) lasted considerably longer, i.e., 6–10 months. The number of participants was much smaller and the subjects were older and more obese which makes the results difficult to compare directly. The results we report are from a very large group of children and the treatment program has run for many years indicating a very efficient and well-tried method.
The weight change and BMI change were dependent of the gender, age, duration of stay, and the baseline values of which the baseline values showed the largest impact.
Longer residence at the institution was associated with greater weight loss even after a stay of 60 days. This may indicate that attending the program for 60 days might be too short to obtain the optimal outcome of the residential treatment program. However, the standard duration of the treatment has been shortened over the years to allow more children to participate in the treatment program.
There were significant differences for boys and girls. Though BMI-SDS before treatment was higher for the boys, the boys had lower values at the end of treatment. The time course of the reduction in BMI showed that the boys caught up with the girls after about 8 weeks. The gender differences with boys loosing more weight was also observed initially by Deforche et al. (23) and by Lazzer et al. (27); however, in other studies no significant difference between genders was observed (14,16,22,28). The reason for the larger weight loss in boys was probably that the boys were more physically active than the girls (35). Also the dependency of age may be explained by the physical activity as it has been reported that physical activity decreases with age (36,37).
The growth acceleration during the stay with the large weight loss was not expected. Some, mainly older studies, have reported a reduction in height velocity for children on weight-reducing diets (38,39,40,41) while other results suggest that energy restrictions do not cause any negative effects on height as indicated in this study (19,27,29,30,42,43,44). A positive effect of weight reduction on height as observed in this study is rarely reported. To our knowledge only Suskind et al. (42) have reported height velocities greater than normal at week 26 after an initial temporary reduction in height velocity, in weight reduction programs lasting for 36 weeks. There is however no data presented in the paper. The overweight children were taller than their normal weight peers before start of treatment. This has been reported previously and it has also been reported that the pattern of growth of obese children is different from that of normal weight subjects. Obese children have an accelerated growth rate and advanced bone age in prepuberty; however during pubertal development the height velocity is reduced (19,22,30,45,46).
The observed increase in height-SDS during the treatment is likely to reflect a genuine increase during the period, despite the fact that we do not have a velocity value for the period before the treatment. The mean velocity in this group of children with a broad age range is at the level of peak height velocity, which only lasts for a limited period, less than a year, in the average child. There are several potential explanations for an increase in growth velocity during the treatment. We speculate that the most plausible explanations are an improved diet with a higher intake of protein and micronutrients, a stimulation of bone growth through the high level of physical activity or a combination of the two. We have no data on the dietary intake of the children before start of the treatment, but find it likely that the content of “empty calories”, fat and sugar was high and thereby the density and intake of micronutrients low. If milk intake was higher during the treatment, this could further stimulate linear growth (47). The milk intake was about 500 ml/day which probably is higher than before start of treatment as milk was served for breakfast and dinner and soft drinks was not available. It is well known that increased physical activity can stimulate growth of bone mass in children (48,49). Furthermore it has been shown that physical activity can increase linear growth in both animals and malnourished children (50).
Other causes could be an improved posture and decreased lordosis as a result of the large weight loss. However, there was no association between weight loss and increase in height. It has been suggested that a high linear growth velocity could be caused by precocity of obese children (46), but we do not find it likely as there was a broad age range in our study. Finally, it could be that height measurement at the end of treatment was biased upward, but BMI is not used routinely in monitoring of the children, and there is therefore no benefit to the outcome of the treatment by measuring an increased height.
The study has some limitations. The study was not designed as a randomized controlled trial, but aimed to evaluate the program at the DCSH. Information about the pubertal stage of the children is not available. In addition, the large core study presents the short-term effects of the program as follow-up examinations were not part of the program. Nevertheless, as untreated obese children usually continue to gain weight and increase BMI (16,17,22) the described results are very encouraging.
In summary, the residential treatment program at the DCSH, consisting of a dietary exercise-behavioral intervention for overweight children, resulted in very large weight reduction without any negative effects on growth. The cause of the linear growth velocity acceleration should be examined further. The boys obtained the largest reductions in weight parameters, probably due to a higher physical activity level. Obesity is recognized as a chronic disease and the challenge remains to improve treatment programs focusing on long-term weight maintenance. More research is needed to determine the long-term effects of the program at the DCSH and to investigate whether the treatment could be further improved by involving the families more. Recently the DCSH has implemented a new program with increased collaboration with local health workers to establish a follow-up of the children after discharge to maintain the weight loss.
Acknowledgment
We greatly acknowledge the manager of the DCSH in Skælskør, Karl Arvidson, and his personnel for their cooperation and hospitality during the data collection.
Disclosure
The authors declared no conflict of interest.
