An Overweight Program for Prepubertal Children, Assessed By an Accurate Early Indicator of Adiposity Reduction

Introduction: The early effectiveness of a weight treatment program, combining diet with exercise, was assessed in compliant obese prepubertal children in a clinic-setting. Methods: Body composition was assessed monthly via air displacement plethysmography in a convenience cohort of consecutively referred prepubertal obese children (body mass index > P95), aged 3 to 9 years, to a hospital outpatient overweight treatment program. This program consisted of a dietary intervention based on macronutrients in proportions consistent with the recommendations for age combined with exercise counseling of more than one hour per day at least three times per week. A reduction of > 5% in fat mass index (as surrogate of adiposity), was the primary outcome; and the time to achieve it was the secondary outcome. Per-protocol analysis was performed. Results: Sixty children fully compliant with the program were included. The median (min-max) time of follow-up was 105 (35 337) days. The early success rate was 70.0% (95% CI 57.5 80.1). Forty-two children achieved the outcome at a median 71 (35 316) days. No difference for sex was found. The lower the FMI at referral, the higher the probability of success. Conclusion: The short-term effectiveness of this weight management program for compliant prepubertal children seems satisfactory. Early positive reinforcement based on a reliable indicator may contribute to the success of weight management programs.


Introduction
Early successful treatment of overweight and obesity contributes to prevent programming of obesity and associated comorbidities into adolescence and adulthood [1].
consecutively referred prepubertal children with confirmed obesity was recruited during one year.
Obesity was confirmed in the first visit, defined as BMI over the 95 th centile, adjusted for age and sex [8,9].Height and weight were used for calculation of BMI and BMI z-scores, using WHO AnthroPlus (http://whoanthroplus.software.informer.com/ ).Height was measured using the Seca 240 Wall-Mounted Stadiometer (3M, A&D Medical, ACC, ADC) and body weight was measured using the Seca 763 scale (Seca Medical Measuring Systems and Scale, UK), according to the recommended technique [10]; the average of three measurements was recorded for analysis.
The overweight treatment program was prescribed by the same paediatrician (CD) and the same dietician (AA), complying with the outpatient clinic treatment protocol.Puberty was excluded based on the Tanner stages [11].A customized weight management program was applied to all patients, including verbal and written prescriptions for: 1) a daily eating plan with a normocaloric diet composed of balanced macronutrients, in proportions consistent with Dietary Reference Intake recommendations for age.Special attention was given to recommend frequent small meals with foods with low simple sugar content, low energy density, and high fiber or water content [12,13]; and 2) exercising minimum of one hour per day at least three times per week [14].
The protocol includes scheduled assessments of weight, height and body composition measurements in the Nutrition Lab, at admission and monthly thereafter for follow-up.Observers performing the measurements were not aware of the weight management prescriptions.Individuals not complying with this scheduled were not included in the analysis (per-protocol analysis).No extra assessments were undertaken purposefully for this study.
Body composition was measured via Air Displacement Plethysmography (ADP), using the Bod Pod equipment (Life Measurements, Concorde, CA, USA), operated by the same observer (MPGD).According to the manufacturer's instructions, measurements were obtained with the subjects wearing tightfitting swimsuit and swim cap only.This method measures body mass, Fat Mass (FM), and Fat-Free Mass (FFM) expressed in Kg, with precision of 0.1 Kg.The FM is calculated by the equipment, assuming the density of fat to be 0.9007, and pre-determined age-and gender-specific densities of FFM [15].The FMI was computed as FM (kg) divided by height squared (m 2 ).
For this study, the end of follow-up was either the time when the primary outcome was achieved, or the day of the last body composition measurement up to twelve months, whichever occurred first.
The primary outcome was the variation of FMI; the achievement of a 5% reduction in FMI was arbitrarily chosen as the threshold for early success.The normal distribution of the variables at baseline was verified using the Kolmogorov-Smirnov test.As the distribution of FMI was asymmetrical, the Wilcoxon test for paired samples was used for comparisons over time.The incidence of early success is given with 95% Confidence Interval (CI), calculated using OpenEpi (CDC, Atlanta, GA, USA), and incidence rates were compared using Qui-square or Fisher exact test as adequate.Kaplan-Meyer survival curves were used to describe time (days) to detect a 5% reduction in FMI in the whole sample.For those cases that achieved a 5% reduction in FMI, the time required to achieve detection was described using median, quartiles and extremes.An exploratory analysis (Mann-Whitney test, Cox regression and Kaplan-Meyer survival curves) was performed to identify factors associated to the effectiveness of the overweight treatment program.A level of significance α = 0.05 was considered.Statistics were performed using SPSS 22.0 (IBM Corp, Armonk, NY, USA).

Results
Ninety-four children were included in the overweight treatment program and 34 (36.2%) who did not comply with the scheduled assessments were excluded from the analysis.In the studied cohort of 60 children, 34 were females; the median (min.-max.)age at recruitment was 7.6 (3.0 -8.7) years, without statistical difference between sexes.The median (min.-max.)time of follow-up was 105 (35 -337) days.At admission, no significant differences were found between included and excluded children, in relation to the median age, sex distribution, and adiposity (median FMI).
The BMI and FMI assessed before the intervention and at the end of the follow-up are presented in Table 1.The BMI z-scores were significantly higher in males than females before the intervention and at the end of the follow-up, but no significant differences between sexes were found for FMI at either timeframe.
The FMI at the end of the follow-up was significantly lower than before the intervention (p < 0.001), with a median difference of 1 kg FM/ m 2 (Table 1).A 5% reduction in FMI was detected in 42 children (26 females), at a median (min.-max.)time of 71 (35 -238) days, therefore the success rate was 70.0% (95% CI 57.5 -80.1).Considering the whole cohort, the median time for detection of 5% reduction in FMI was achieved at 98 days (95% CI 70.0 -126.0)(Figure 1).

Discussion
This study was designed to assess compliant families through a per-protocol analysis.Therefore, these results concern the most motivated children and families, assuming that early reinforcement to already motivated individuals may further contribute to the treatment adherence.In this clinicsetting program combining dietary intervention with exercise counseling, 70% (95% CI 57.5 -80.1) of the obese prepubertal children who complied achieved a significant early (median 10 weeks) reduction in adiposity (> 5% assessed by the FMI).Less obese prepubertal children were found to be more prone to early success.
Comprehensive interventions for childhood obesity treatment including diet and exercise as prescribed to our patients, have shown sustainable results in clinic-setting programs, as evidenced by anthropometric indices [4,16].Dietary intervention based on low calorie-low fat diets have been largely used for treating obesity in children, but many experts instead recommend a normocaloric diet with macronutrients in proportions consistent with the recommendations for age [4,17], as used in the assessed program.Low calorie-low fat diet for childhood obesity management was found to have no further advantages compared with other calorically restricted diets and may have side-effects such as binge eating [18,19].In the assessed program, children were counseled on physical activity and encouraged to exercise more than one hour per day at least three times per week.A comprehensive review of evidence on physical activity for school-aged youths reported that programs of 30-to 60-minute duration, performed 3 to 7 days per week, resulted in reductions in total body fat and visceral adiposity among obese children and adolescents [14].In other systematic reviews on the efficacy of interventions for pediatric obesity, exercise was only found to have a moderate effect on adiposity [16,20].
The early effectiveness rate found in this observational study is difficult to compare with other studies, due to great variation in intervention programs, target age groups, outcome measurements, and defined thresholds for success [2].Nevertheless, it has been reported in systematic reviews that most studies demonstrate beneficial effects of intervention programs on child adiposity from baseline to the end of intervention or follow-up [2,4].Particularly, a significant short-term benefit has been reported [21].
Adherence to the programs remains an important issue, as it has been shown that rate of dropout from overweight treatment programs increases if family confidence is low [22].Hence, early positive reinforcement may contribute to increase adherence [6].
It has been recommended that clinic-based childhood obesity interventions should be guided by established strategies and protocols in order to measure improvement over time [23].While the definition of obesity is based on excessive adiposity, the best measurement for degree of body fatness remains controversial [24].BMI has been recommended as the routine screening approach in clinical practice for children older than 6 years, but there is insufficient evidence to provide a similar recommendation in younger children [25,26].BMI was found to have high specificity but low sensitivity for detection of excess adiposity in children [27].In particular, BMI may be biased as a proxy for longitudinal adiposity assessment in children, since strong correlations exist between BMI and the components of weight other than body fat mass, such as lean mass and bone mass [28,29].A systematic review [20] of trials measuring the effect of physical activity on adiposity (i.e., %FM) found a moderate treatment effect, whereas trials measuring the effect on BMI found no significant effect; moreover, combined lifestyle interventions trials led to small changes in BMI.Therefore, it was recommended that adiposity rather than BMI-derived outcomes be considered in order to better reflect the efficacy of interventions [20].In prepubertal obese children, FMI was found to have higher detection rate of adiposity reduction than BMI and similar but earlier detection rate than the %FM [7].Thus, we used FMI instead of BMI to monitor weight management.
The best suited expression to describe adiposity change as an outcome of weight management in growing children, either as raw units, percentages, z-scores, or centiles, is not known [30].To the best of our knowledge, no z-scores or centiles are available for ADP-measured FM in prepubertal children.Therefore, an arbitrary 5% reduction of FMI was considered the cut-off for short-term indicator of success, as it allows each child outcome to be adjusted to its own initial adiposity and focuses comparisons on the rate of successful cases rather than the linear distribution of FMI values.To assure some direct external comparability, the effectiveness was also assessed by comparing the distribution of FMI at referral and at the end of follow-up.The significant difference found is in line with other studies that used adiposity measurements as outcome indicators [2,4,21].A recent Cochrane review [2] highlights that the questions on what interventions are most effective at different levels of obesity severity, at different ages and developmental stages, remain largely unanswered.We found that the lower the FMI at referral, the higher the probability to achieve significant early reduction in FMI.No associations were found between either sex or age at referral and the incidence of the success or time to achieve it.This finding is in contrast with Danielsson, et al. [31] who reported that the effect of behavioral treatment initiated between 6 and 9 years of age appeared to be greater among severely obese children than among moderately obese ones.This difference is probably due to the way the indicators were computed, as Danielsson, et al. [31] analyzed the absolute variation of BMI-SD score whereas we compared the relative variation of FMI adjusted to the FMI at referral.Nevertheless, we cannot dismiss the possibility that having analyzed the data as per-protocol might have biased the association between the severity of obesity referral and the odds for effectiveness.
Effective treatment of obesity in prepubertal children is crucial; as the greater the BMI-SD score at earlier ages, the higher the odds for developing severe obesity in adolescence [31].On the other hand, lifestyle interventions with parental involvement have evidenced a large effect in children aged 8 year or less [20].Our data and data from the literature contribute to underline the importance of early detection and treatment of childhood obesity and the difficulty to treat the more severe cases.
Limitations of this study should be acknowledged.Eligibility for recruitment was based on the clinical diagnosis of obesity based on BMI and not on accurate indicators of excessive adiposity, such as indices providing reliable FM measurements by dual energy X-ray absorptiometry [27].ADP has been validated in children aged 7 to 10 years [32], but not yet in younger children, making it unreliable for ascertainment of the recruitment of obese children.Therefore, BMI was used for eligibility assessment, once it is the most widely used clinical criterion to screen obesity in outpatient children [8,27].Conceptually, this might have introduced a selection bias for including children with less adiposity, but the high BMI z-scores found in the baseline assessment (Table 1) makes this bias unlikely.The higher BMI z-scores found in males did not reflect on a significant difference on FMI, which is the outcome index used.A convenience sample was used to assess the effectiveness of the ongoing weight treatment program for childhood obesity.In spite of the convenience nature of the sample, the consecutive recruitment of patients and its number allowed the identification of statistical and clinical significant associations.Despite the absence of differences in age, sex, and adiposity at admission, between included and excluded individuals, the exclusion of non-compliant individuals from the per-protocol analysis causes a "loss for follow-up" bias, precluding any analysis of the early compliance to the intervention, and overestimates the treatment effect.The adherence to the prescribed dietary plan was not periodically verified in a structured manner, despite it was qualitatively enquired in the follow-up visits.A non-interventional external control group was not used because it would be unethical, considering that the overweight treatment protocol adopted has been considered good practice in the studied condition; hence, each child was its own historical control.Finally, this short-term study focused on early success of the overweight treatment program, thus sustainability of the effect is not presented.
To summarize, the short-term effectiveness of a clinic-setting overweight treatment program for compliant prepubertal children was assessed using FMI as early indicator of adiposity reduction.This program, consisting of a dietary intervention based on macronutrients in proportions recommended for age combined with exercise counseling of more than one hour per day at least three times per week seems satisfactory, since a significant adiposity reduction was achieved in 70% of the children who complied at a median 10 weeks.

Table 1 :
BMI and FMI assessments before the weight management intervention and at the end of follow-up, in the whole cohort, males and females BMI: Body Mass Index; FMI: fat mass index.Mann-Whitney test between sexes: a p < 0.05, b p > 0.05.c Wilcoxon test for paired samples: p < 0.001.