Comorbidities in Primary vs Secondary School Children With Obesity and Responsiveness to Lifestyle Intervention

Abstract

Context

Childhood obesity increases the risk of diseases as diabetes, cardiovascular disease, and nonalcoholic fatty liver disease.

Objective

To evaluate the prevalence of comorbidities in school-age children with obesity and to compare its prevalence and the effect of a lifestyle intervention between children in primary and secondary school and between boys and girls.

Design

Cross-sectional analysis and lifestyle intervention.

Setting

Centre for Overweight Adolescent and Children’s Healthcare.

Patients

Comorbidities were evaluated in 149 primary and 150 secondary school children with (morbid) obesity (162 girls). The effect of lifestyle intervention was studied in 82 primary and 75 secondary school children.

Intervention

One-year interdisciplinary lifestyle intervention.

Results

Insulin resistance (37%), impaired glucose tolerance (IGT) (3%), dyslipidemia (48%), hypertension (7%), and elevated liver transaminase levels (54%) were already common in primary school children. Glomerular hyperfiltration and insulin resistance were more prevalent in secondary school children. IGT was more prevalent in girls. The change in body mass index z score after intervention was greater in primary school children (primary vs secondary: −0.25 ± 0.32 vs −0.11 ± 0.47), even as the change in low-density lipoprotein cholesterol concentrations [primary vs secondary: −0.30 (interquartile range, −0.70 to 0.10) vs −0.10 (interquartile range, −0.40 to 0.30)] and systolic blood pressure z score (primary vs secondary: −0.32 ± 1.27 vs 0.24 ± 1.3). The change in body mass index z score, but not in comorbidities, was greater in boys (boys vs girls: −0.33 ± 0.45 vs −0.05 ± 0.31).

Conclusions

The presence of comorbidities is already evident in primary school children with obesity. The effect of a lifestyle intervention on these comorbidities is greater in primary compared with secondary school children, stressing the need for early interventions.

The global prevalence of obesity in children has increased during the past decades (1), making it an important cause of morbidity and mortality around the world. In 2015, it was estimated that ∼108 million children were obese worldwide (1). In Western Europe, the prevalence of obesity in boys and girls <20 years is 7.2% and 6.4%, respectively (2). In the Netherlands, the prevalence of obesity in children aged 4 to 17 years is 2.3% in boys and 3.3% in girls (3).

Obesity during childhood is an important risk factor for the development of various comorbidities, including dyslipidemia, hypertension, diabetes, and sleep apnea, but also nonalcoholic fatty liver disease and kidney disease (4–7). Additionally, children with obesity are likely to grow up to become adults with overweight or obesity (8). Lobstein et al. (9) have estimated that in 2025, unless we can make a larger impact on childhood obesity in the years until then, there will be ∼91 million children with obesity worldwide, of which one-half will experience one or more comorbidities, such as type 2 diabetes, hypertension, or fatty liver disease. In adults, obesity has also been associated with an increased all-cause mortality (10). In addition to the effects of obesity on health, obesity has also been associated with decreased quality of life (11), lower work productivity (12), and an increase in utilization of health care services and higher health care costs (13), thereby not only having an effect on the life of the individual with obesity, but also on society. Altogether, these effects of the obesity epidemic stress the need for early intervention to decrease the effect on the health parameters, but also economical outcomes.

Previous studies have shown that the development of these lifestyle-related comorbidities starts in young children (4, 14). Also, studies have shown that lifestyle interventions are more effective in reducing body mass index (BMI) z score in younger children with extreme obesity than in adolescents with extreme obesity (15) and that younger patients were more likely to maintain weight loss (16).

Most studies examining the effect of lifestyle interventions in children focus on the effect on BMI (z score), rather than on the effect of the intervention on health parameters; comparisons of intervention effects between different age subgroups are scarcely made. Additionally, previous studies have shown that physical activity and sedentary behavior (17) and dietary composition (18) differ between boys and girls and that boys and girls may respond differently to physical activity interventions (19). However, knowledge regarding sex differences in the effect of combined lifestyle interventions is limited.

The aim of this study is to describe the prevalence of an elaborate panel of early lifestyle-related comorbidities, including parameters of metabolic and cardiovascular health, but also liver and kidney health, in a group of school-aged children with obesity, and to compare the prevalence of these comorbidities and the effect of 1 year of interdisciplinary lifestyle intervention between children in primary and secondary school, and between boys and girls.

Methods

Participants and setting

This study was designed and conducted within the setting of the Centre for Overweight Adolescent and Children’s Healthcare at the Maastricht University Medical Centre. School-aged children (6 to 16 years; primary school, 6 to 11 years; secondary school, 12 to 16 years) with obesity and morbid obesity were referred to our center by youth health services (i.e., school doctors), general practitioners, and pediatricians, who have a key role in the recognition of obesity in the children in our region. Their approach to the care for obese children is documented in guidelines. In our center, children and their families are evaluated, monitored, and guided as described in detail previously (20). In summary, all children underwent a comprehensive assessment before the start of the intervention to exclude underlying syndromic or endocrine conditions of obesity, evaluate complications and risk factors, and gain insight in behavior and (family) functioning. A follow-up assessment including all the examinations performed during the initial assessment was offered to all children after ∼1 year of lifestyle intervention. The information gathered in the assessment was used to develop a care plan that was tailored to the needs of each family. All children and their families were offered individual guidance focusing on lifestyle improvements regarding nutrition, food habits, physical activity, sleep, and psychological and social aspects. Points of improvement that were identified during the initial assessment were used as focus points for making small, step-by-step lifestyle changes. These focus points were adapted throughout the intervention, depending on for instance the successfulness of making changes or the identification of new points of improvements. Additional support was provided if, for instance, limited pedagogical skills or financial or psychological problems were identified as barriers for lifestyle improvement. In general, the visits to the outpatient clinic started on a monthly basis, but frequency was adjusted based on personal needs (i.e., frequency was reduced in case of weight loss and maintenance, or outpatient visits were partially replaced with consultation via telephone in case of transport problems).

In addition to the individual outpatient clinic visits, group activities related to nutrition and physical activity were organized several times per year, which children and/or parents could attend on a voluntary basis. Children that were not already participating in regular sport activities were encouraged to participate in a weekly 1-hour physical activity group lesson (in addition to encouragement to increase physical activity levels at home). If possible, different elements of the intervention, such as sports activities, were located in the patient’s own neighborhood.

All school-aged children with obesity and morbid obesity that were guided in our center were included in this study. To prevent selection bias, no children referred by other hospitals were included.

Because of the continuous inflow of new participants into the intervention program, only some of the children had completed 1 year of lifestyle intervention at time of data analysis. Only children that had completed at least 1 year of lifestyle intervention were included in analyses of the intervention effects.

This study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the medical ethical committee of the Maastricht University Medical Centre. Consent was obtained by the parents and/or the child.

Anthropometric measurements

Anthropometric measurements were performed while children were barefoot and wearing underwear only. Weight was measured on a digital scale (Seca). Height was measured using a wall-mounted digital stadiometer (De Grood Metaaltechniek). BMI was calculated and BMI z scores were obtained using a growth analyzer (Growth Analyzer VE). Children were classified as obese or morbidly obese according to International Obesity Task Force criteria (21). In short, age- and sex-dependent cutoff points were used to categorize children as being obese (comparable to a BMI of 30 kg/m² in adults) or morbidly obese (comparable to a BMI of 35 kg/m² in adults). Waist circumference was measured with a nonelastic measuring tape at the midpoint between the lower margin of the last palpable rib and the top of the iliac crest, at the end of a normal expiration. Waist circumference z scores were determined according to reference values for Dutch children (22). All anthropometric measurements were measured once by trained health care personnel.

Glucose metabolism

Fasting blood glucose concentrations were determined using the Cobas 8000 modular analyzer (Roche). Fasting serum insulin levels were determined with the Immulite 1000 (Siemens). After obtaining the fasting blood sample, an oral glucose tolerance test was performed. For this test, 1.75 g of glucose per kilogram of bodyweight was dissolved into 200 mL water with a maximum of 75 g of glucose. After patients drank the glucose solution, plasma blood glucose concentrations were measured every 30 minutes for 2 hours. 
Insulin resistance was evaluated by calculating the homeostasis model assessment for insulin resistance (HOMA-IR) [HOMA-IR = fasting glucose (mmol/L) × fasting insulin (µU/L)/22.5]. HOMA-IR values higher than age- and sex-specific 75th percentiles for children with overweight or obesity were considered abnormal, based on a study by Shashaj et al. (23) that showed that these cutoff points are most accurate to identify children in which the HOMA-IR value can be considered “nonphysiological” and is suspected to occur alongside other aberrant cardiometabolic risk factors. Impaired fasting glucose was defined as a fasting glucose concentration ≥5.6 mmol/L, IGT as a 2-hour glucose concentration ≥7.8 and <11.1 mmol/L, and type 2 diabetes as a 2-hour glucose concentration ≥11.1 mmol/L (24).

Cardiovascular risk parameters

Fasting serum total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride concentrations were determined using the Cobas 8000 modular analyzer (Roche). Low-density lipoprotein (LDL) cholesterol concentrations were calculated with the Friedewald Equation (25). Dyslipidemia was defined as elevated LDL cholesterol, low HDL cholesterol, and/or elevated triglyceride concentrations according to cutoff points for children (26). Ambulatory daytime blood pressure was measured approximately 20 times with an interval of 3 minutes between measurements using the Mobil-O-Graph (I.E.M. GmbH). Mean blood pressure was calculated. Systolic and diastolic blood pressure z scores were calculated according to reference values for height and sex (27). Hypertension was defined as a systolic and/or diastolic blood pressure z score >2. The presence of the metabolic syndrome was determined according to the age-based criteria from the International Diabetes Federation (28).

Parameters reflecting liver and kidney health

Alanine transaminase (ALT) levels and creatinine concentrations were determined with the Cobas 8000 modular analyzer (Roche). For analysis in this study, the upper limit of normal for ALT was considered 22.1 U/L for girls and 25.8 U/L for boys, based on a previous study examining the healthy range of ALT concentrations in children and adolescents (29).

Estimated glomerular filtration rate (eGFR) was calculated according to the Schwartz formula ([eGFR (mL/min/1.73 m²) = 36.5 × height (cm)/plasma creatinine (μmol/L)]) (30). In this formula, the eGFR is indexed for a standardized body surface area (BSA) of 1.73 m², which is suggested to be inaccurate in children, especially in children with obesity and for longitudinal measurements (31). Therefore, the eGFR was deindexed by multiplying it according to the Schwarz formula by the estimated BSA and dividing by 1.73 (31, 32). The BSA was estimated using the Haycock formula [BSA = 0.024265 × weight (kg)^0.5378 × height (cm)^0.3964] (33). Glomerular hyperfiltration was defined as a deindexed eGFR >135 mL/min (34).

Statistical analysis

Statistical analysis was performed with IBM SPSS Statistics 22.0. All data were tested for normality with the Shapiro-Wilk test and reported as mean ± SD or median [interquartile range (IQR)], depending on the distribution. Differences in the prevalence of early lifestyle-related aberrations between primary and secondary school children were analyzed with the χ² test. Differences in parameters between age categories and sex were tested with the independent samples t test or Mann-Whitney U test, as appropriate.

Regression analyses were performed for the Δ (difference between pretreatment and posttreatment) outcomes as dependent variables, with school-age category and sex as independent variables and corrected for BMI z score at baseline and number of consultations during the intervention period. Correction for the number of consultations was performed, because previous studies have shown that the effect of lifestyle interventions on weight might be influenced by the number of consultations during the intervention period (35). In these models, an interaction term for school-age category × sex was added, but did not contribute significantly in any model and was therefore left out of the final models. Corrected regression coefficients shown are unstandardized βs for the contribution of school-age category (primary school as reference) or sex (girls as reference) to the difference between pretreatment and posttreatment outcomes.

Results

Baseline characteristics

Two hundred and ninety-nine children (176 obese, 123 morbidly obese; 46% boys) with a mean BMI z score of 3.52 ± 0.59 and median age of 12.1 (IQR, 10.1 to 14.4) years were included in this study. Of these children 149 were primary school aged and 150 were secondary school aged. 
At baseline, the boys were slightly, but not significantly, younger than the girls [respectively, 11.5 (IQR, 10.1 to 13.7) years compared with 12.4 (IQR, 9.9 to 15.2) years; P = 0.07), despite having a significantly higher BMI z score (respectively, 3.57 (IQR, 3.20 to 4.07) vs 3.28 (IQR, 2.97 to 3.70)]. Additional baseline characteristics and differences between subgroups (primary vs secondary school and boys vs girls) are presented in Table 1.

Table 1.

Baseline Characteristics, Stratified for Age and Sex Subgroups

	Total Group (N = 299)	Primary School (N = 149)	Secondary School (N = 150)	Girls (N = 162)	Boys (N = 137)
Age, y	12.1 (10.1 to 14.4)	10.1 (8.5 to 11.2)	14.4 (13.3 to 15.9)a	12.4 (9.9 to 15.2)	11.5 (10.1 to 13.7)
Sex, M/F	137/162	76/73	61/89	—	—
BMI z score	3.52 ± 0.59	3.56 ± 0.67	3.47 ± 0.50	3.28 (2.97 to 3.70)	3.57 (3.20 to 4.07)a
Waist circumference z score	5.78 (4.47 to 7.24)	5.01 (3.89 to 6.20)	6.59 (5.32 to 7.95)a	5.80 (4.62 to 7.29)	5.55 (4.41 to 7.16)
Obese/morbidly obese	176/123	89/60	87/63	93/69	83/54
Total cholesterol, mmol/L	4.30 (3.80 to 4.90)	4.30 (3.80 to 5.00)	4.30 (3.70 to 4.90)	4.40 (3.85 to 5.00)	4.20 (3.70 to 4.90)
LDL cholesterol, mmol/L	2.60 (2.10 to 3.10)	2.55 (2.10 to 3.10)	2.60 (2.10 to 3.10)	2.68 ± 0.75	2.53 ± 0.74
HDL cholesterol, mmol/L	1.20 (1.00 to 1.40)	1.20 (1.00 to 1.50)	1.20 (1.00 to 1.40)	1.20 (1.00 to 1.40)	1.30 (1.00 to 1.50)a
Triglycerides, mmol/L	1.01 (0.73 to 1.40)	0.98 (0.71 to 1.37)	1.02 (0.73 to 1.44)	1.06 (0.74 to 1.52)	0.97 (0.71 to 1.30)
Fasting glucose, mmol/L	4.20 (3.90 to 4.60)	4.25 (3.90 to 4.60)	4.20 (3.90 to 4.60)	4.20 (3.90 to 4.60)	4.30 (4.00 to 4.60)
HOMA-IR	3.14 (2.04 to 4.78)	2.43 (1.51 to 3.60)	3.99 (2.87 to 5.48)a	3.51 (2.43 to 5.07)	2.82 (1.76 to 4.49)a
Systolic blood pressure z score	0.20 (−0.40 to 0.90)	0.10 (−0.45 to 0.90)	0.30 (−0.43 to 0.93)	0.30 (−0.33 to 1.10)	0.10 (0.50 to 0.80)
Diastolic blood pressure z score	−0.70 (−1.30 to 0.10)	−0.70 (−1.55 to −0.20)	−0.50 (−1.20 to 0.30)a	−0.60 (−1.20 to 0.20)	−0.80 (−1.45 to −0.05)
ALT, U/L	23.5 (18.0 to 32.0)	24.0 (19.0 to 31.5)	23.0 (17.0 to 34.0)	22.0 (16.0 to 28.0)	26.5 (20.0 to 39.0)a
Estimated glomerular filtration rate, mL/min	121.8 ± 27.5	111.4 ± 27.0	131.8 ± 24.2a	122.8 ± 29.1	120.5 ± 25.5

Data are presented as means ± SD or median (IQR).

Abbreviations: F, female; M, male.

^a

Significant difference between primary and secondary school-aged children or between boys and girls.

Insulin resistance (37%), IGT (3%), dyslipidemia (48%), hypertension (7%), and elevated liver transaminase levels (54%) were already present in primary school children (Table 2). Insulin resistance (52%) and glomerular hyperfiltration (42%) were more prevalent in secondary school children compared with primary school children (Table 2). In girls, IGT was more prevalent than in boys (respectively, 6% vs 1%) (Table 2).

Table 2.

Prevalence of Comorbidities in Primary and Secondary School-Age Children

	All Participants (N = 299)	Primary School (6-11 Y) (N = 149)	Secondary School (12-16 Y) (N = 150)	Girls (N = 162)	Boys (N = 137)
Dyslipidemia	51.9	48.2	55.4	56.6	46.2
Increased total cholesterol	19.1	22.9	15.5	17.7	20.8
Increased LDL cholesterol	18.4	17.9	18.9	19.0	17.7
Decreased HDL cholesterol	31.3	27.9	34.5	34.2	27.7
Increased triglycerides	24.2	24.1	24.3	26.6	21.4
Insulin resistance	44.9	36.8	52.3a	47.4	42.0
Impaired fasting glucose	1.1	1.5	0.7	0.7	1.6
IGT	3.8	2.5	4.8	6.1	0.9a
Diabetes type 2	0.4	0.0	0.7	0.7	0.0
Hypertension	8.7	7.1	10.3	11.1	6.0
Systolic hypertension	8.0	7.1	9.0	9.8	6,0
Diastolic hypertension	2.4	0.7	4.1	3.9	0.8
Metabolic syndrome	17.4	16.4	17.9	20.0	14.3
Increased ALT levels	50.5	53.6	47.6	46.8	55.0
Glomerular hyperfiltration	29.4	16.4	41.8a	32.7	25.4

Data are presented as percentages.

^a

Significant difference in the prevalence of lifestyle-related diseases between primary and secondary school-aged children or between boys and girls

Effect of lifestyle intervention

The effect of 1 year of lifestyle intervention was evaluated in 82 primary school children and 75 secondary school children. The primary school children that underwent an elaborate health screening after 1 year of intervention had higher total and LDL cholesterol concentrations compared with the primary school children that dropped out of the intervention during the first year. There were no other baseline differences between these groups. Also, there were no baseline differences between secondary school children that underwent an elaborate health screening after 1 year and the children that dropped out during the first year (data not shown). The dropout rate did not differ significantly between primary and secondary school children.

After 1 year of lifestyle intervention, there was a significant reduction in BMI z score (−0.18 ± 0.40), total cholesterol concentrations [−0.25 (IQR, −0.63 to 0.30)], LDL cholesterol concentrations [−0.10 (IQR, −0.50 to 0.20)], and triglyceride concentrations [−0.05 (IQR, −0.41 to 0.19)] in the total group of children (Table 3). In primary school children, there was a significant decrease of BMI z score (−0.25 ± 0.32), total cholesterol levels [−0.30 (IQR, −0.70 to 0.10)], LDL cholesterol levels [−0.30 (IQR, −0.70 to 0.10)], and systolic blood pressure z score (−0.32 ± 1.27), and an increase in eGFR (4.8 ± 17.7) (Table 3). In secondary school children, BMI z score and eGFR were significantly reduced, respectively: −0.11 ± 0.47 and −6.3 ± 16.9. The effect of the intervention on BMI z score, LDL cholesterol concentrations, and systolic blood pressure z score was significantly greater in primary school children compared with secondary school children (Table 3). Regression analysis showed that school-age category was still a significant predictor for change in BMI z score, LDL cholesterol concentrations, systolic blood pressure z score after intervention, after correction for BMI z score at baseline, and the number of consultations during the intervention (Table 3).

Table 3.

The Effect of 1 Y of Lifestyle Intervention, Stratified for Age and Sex Subgroups

	Pretreatment	Difference Between Pretreatment and Posttreatment Outcomes
	Total Group (N = 157)	Total Group (N = 157)	Primary School at Baseline (N = 82)	Secondary School at Baseline (N = 75)	Corrected Regression Coefficients	Girls (N = 84)	Boys (N = 73)	Corrected Regression Coefficients
Sex, M/F	73/84	73/84	38/44	35/40		—	—
No. of consultations during the intervention period	—	8 (7 to 10)	9 (7 to 10)	8 (7 to 10)		8.5 (7 to 10)	8 (7 to 10)
BMI z score	3.52 ± 0.57	−0.18 ± 0.40a	−0.25 ± 0.32a	−0.11 ± 0.47a,b	0.123c	−0.05 ± 0.31	−0.33 ± 0.45a,b	−0.260c
Total cholesterol, mmol/L	4.45 (4.00 to 5.00)	−0.25 (−0.63 to 0.10)a	−0.30 (−0.70 to 0.10)a	−0.20 (−0.50 to 0.10)	0.165	−0.20 (−0.43 to 0.20)a	−0.30 (−0.70 to 0.10)a	−0.174
LDL cholesterol, mmol/L	2.65 (2.20 to 3.20)	−0.10 (−0.50 to 0.20)a	−0.30 (−0.70 to 0.10)a	−0.10 (−0.40 to 0.30)b	0.178	−0.09 ± 0.61	−0.21 ± 0.51a	−0.150
HDL cholesterol, mmol/L	1.20 (1.00 to 1.40)	0.00 (−0.10 to 0.10)	0.00 (−0.10 to 0.20)	0.00 (−0.10 to 0.10)	−0.060	0.00 (−0.10 to 0.10)	0.00 (−0.10 to 0.10)	−0.027
Triglycerides, mmol/L	1.11 (0.75 to 1.57)	−0.05 (−0.41 to 0.19)a	−0.07 (−0.42 to 0.19)	−0.05 (−0.40 to 0.19)	−0.007	−0.01 (−0.35 to 0.19)	0.14 (−0.44 to 0.19)	0.009
Fasting glucose, mmol/L	4.17 ± 0.53	0.04 ± 0.68	0.08 ± 0.66	0.01 ± 0.71	−0.077	0.08 ± 0.75	0.00 ± 0.60	−0.075
HOMA-IR	2.88 (1.99 to 4.70)	0.04 (−1.00 to 1.13)	0.39 (−0.32 to 1.23)a	−0.45 (−1.93 to 0.63)b	−1.374c	0.01 (−0.99 to 1.13)	0.06 (−1.03 to 1.16)	−0.581
Systolic blood pressure z score	0.15 (−0.40 to 0.80)	−0.05 ± 1.34	−0.32 ± 1.27a	0.24 ± 1.35b	0.555c	−0.08 ± 1.42	−0.01 ± 1.25	0.052
Diastolic blood pressure z score	−0.61 ± 1.03	−0.19 ± 1.16	−0.26 ± 1.14	−0.12 ± 1.19	0.114	−0.25 ± 1.16	−0.12 ± 1.17	0.118
ALT, U/L	25.0 (18.0 to 34.0)	−1.0 (−6.0 to 3.0)	−1.0 (−6.0 to 2.0)	0.0 (−6.0 to 4.0)	−0.446	−1.0 (−6.0 to 3.3)	0.0 (−7.5 to 3.5)	−0.849
Estimated glomerular filtration rate, ml/min	123.5 ± 26.5	−0.5 ± 18.1	4.8 ± 17.7a	−6.3 ± 16.9a,b	−2.821	−0.2 ± 19.6	−0.9 ± 16.5	−2.049

Data are presented as means ± SD or median (IQR).

^a

Significant change after 1 year of lifestyle intervention.

^b

Significant difference in the effect of lifestyle intervention between primary and secondary school-aged children or boys and girls. Corrected regression coefficients shown are unstandardized βs for the contribution of school category (primary school as reference) or sex (girls as reference) to the difference between pretreatment and posttreatment outcomes. The outcomes were corrected for BMI z score at baseline and number of consultations during the intervention period. School category outcomes were corrected for sex and vice versa.

^c

Significant contribution of either school category or sex to differences between pretreatment and posttreatment outcomes.

Comparison of sexes showed that the effect of the lifestyle intervention on BMI z score was evidently greater in boys compared with girls, with a BMI z score reduction of −0.33 ± 0.45 compared with −0.05 ± 0.31. The difference in Δ BMI z score between boys and girls was still present when analysis was stratified for primary and secondary school children (data not shown). Despite the difference in the amount of weight loss, there was no sex difference for the intervention effect on the other health parameters neither in the total group nor in stratified analysis for primary and secondary school (data not shown). Regression analysis showed that sex was still a significant predictor for changes in BMI z score after correction for BMI z score at baseline and the number of consultations during the intervention (Table 3).

Discussion

Childhood obesity increases the risk of lifestyle-related diseases such as nonalcoholic fatty liver disease, diabetes, cardiovascular disease, and renal disease at later age. In this study, we evaluated elaborate health parameters and the prevalence of comorbidities in primary and secondary school children with obesity and evaluated the effect of 1-year family-based, interdisciplinary lifestyle intervention. Remarkably, our results show that the prevalence of early lifestyle-related comorbidities hardly differs between primary school children and secondary school children with obesity and morbid obesity. Additionally, the positive effect of lifestyle intervention is greater in the younger children.

In accordance with this study, other studies from different regions across the world have also shown that these early comorbidities in children with obesity are already present from a young age. The early development of these comorbidities in our population of children with obesity and morbid obesity are comparable to the prevalence of hypertension, dyslipidemia, and glucose metabolism aberrations in a large pediatric central European population described by Reinehr et al. (14). Notable differences between our results and the results of Reinehr et al. are a higher prevalence of decreased HDL cholesterol levels and a slightly lower prevalence of IGT in our population, which, for the prevalence of abnormal HDL cholesterol concentrations, is most likely the result of a difference in cutoff values used. l’Allemand et al. (36) showed that elevated total cholesterol and triglyceride levels were more prevalent in children younger than 12 years compared with children with obesity and morbid obesity aged 12 to 20 years, and that low HDL levels were more common in the children older than 12 years. In our study, we also found that the prevalence of elevated total cholesterol levels was slightly lower and the prevalence of decreased HDL cholesterol levels was slightly higher in secondary school children, but this difference did not reach statistical significance.

Remarkable differences in the results from our study compared with studies from other regions of the world mainly concern a clear difference in glucose metabolism abnormalities. In our study, we found a low prevalence of impaired fasting glucose (1%) and type 2 diabetes (0.4%). Skinner et al. (4) describe a prevalence of impaired fasting glucose of 23% in 1005 children with obesity (aged 12 to 19 years) from the United States. Santiprabhob et al. (37) describe a prevalence of IGT of 21% in 126 children with obesity (aged 8 to 18 years) from Thailand. This could be explained partially by a difference in the degree of obesity between these populations and our study population; however, a direct comparison is not possible because of the different ways of reporting obesity classification. Additionally, these marked differences in glucose metabolism abnormalities might be (partially) due to differences in dietary habits or ethnicity (38, 39).

In our study, there was a significant decrease of BMI z score after lifestyle intervention in both primary school children and secondary school children, with a greater effect in primary school children (−0.25 ± 0.31 vs −0.11 ± 0.47 in secondary school children). In addition to a significant improvement of BMI z score, we found an improvement of LDL cholesterol concentrations and systolic blood pressure z score in primary school children, which were all significantly greater compared with secondary school children.

Previous studies have also reported a greater effect of lifestyle intervention on BMI z score in younger children (15, 16, 40, 41). Regression analysis in our study showed that school-age category was a significant predictor for change in BMI z score, regardless of BMI z score at baseline and the number of consultations during the intervention. Moreover, because both primary and secondary school children in this study were from the same population and were referred to our lifestyle intervention according to the same guidelines, we do not assume that there is a selection bias that could explain the difference in the effect of the lifestyle intervention. Perhaps a reduced influence of parents on the health behavior of adolescents may play a role in the greater effect of lifestyle interventions in younger children.

Two recent Cochrane reviews have evaluated the effects of dietary, physical activity, and behavioral randomized controlled trials for the treatment of overweight and obesity in children (6 to 11 years) (42) and adolescents (12 to 17 years) (43). These reviews show a mean decrease in BMI z score of 0.06 units in children and 0.13 units in adolescents after intervention. The design and duration of the interventions in these reviews are quite diverse, however, impairing a direct comparison with our study.

In our study, there was difference in the effect of lifestyle intervention on the BMI z score between boys and girls, with a significantly greater effect in boys. It is interesting to consider why we found this difference. Perhaps, preexisting sex differences play a role here. For instance, previous studies have shown that boys in general are more physically active than girls (17) and that boys may be better able than girls to change several physical activity aspects, like increasing activity and decreasing sedentary time (19).

Interestingly, the improvement in health parameters during the intervention did not differ significantly between boys and girls. Because their BMI z score stabilized, perhaps the girls did also improve their lifestyle. This improvement might have been enough to decrease comorbidities, but not enough to decrease their BMI z score. In that case, weight stabilization in girls might be as important as weight loss in boys.

A major strength of this study is that it was performed in a representative group of school-age children with obesity and morbid obesity in our region. Also, in addition to anthropometric data, which are often the focus in studies evaluating the effect of lifestyle interventions, we have data on elaborate health outcomes and of the effect of lifestyle intervention on these health parameters. A limitation of this study is that few objective data regarding behavioral changes resulting from the lifestyle intervention, such as changes in eating and physical activity behavior, are available, thereby limiting interpretation of which specific components of the lifestyle changes have led to the health improvements we found.

Conclusion

The early development of comorbidities is already evident in primary school-aged children with obesity and morbid obesity. The positive effect of lifestyle intervention on these comorbidities is greater in primary school children compared with secondary school children, reflecting the need to start intervention as early as possible, but also to have attention for healthy lifestyle already early on in primary school to prevent the development of obesity and the accompanying health effects.

Abbreviations:

Abbreviations:
 
• ALT

  alanine transaminase

 
• BMI

  body mass index

 
• BSA

  body surface area

 
• eGFR

  estimated glomerular filtration rate

 
• HDL

  high-density lipoprotein

 
• HOMA-IR

  homeostasis model assessment for insulin resistance

 
• IGT

  impaired glucose tolerance

 
• IQR

  interquartile range

 
• LDL

  low-density lipoprotein

Acknowledgments

Clinical Trial Information: ClinicalTrial.gov no. NCT02091544 (registered 19 March 2014).

Disclosure Summary: The authors have nothing to disclose.