Author + information
- Received April 25, 2009
- Revision received July 27, 2009
- Accepted August 6, 2009
- Published online December 15, 2009.
- Nathalie J. Farpour-Lambert, MD*,* (, )
- Yacine Aggoun, MD*,
- Laetitia M. Marchand, MS*,
- Xavier E. Martin, MS*,
- François R. Herrmann, MD, MPH† and
- Maurice Beghetti, PD*
- ↵*Reprint requests and correspondence:
Dr. Nathalie J. Farpour-Lambert, Pediatric Cardiology Unit, Department of Child and Adolescent, University Hospitals of Geneva, 6 rue Willy-Donze, 1211 Geneva 14, Switzerland
Objectives The aim of this study was to determine the effects of physical activity on systemic blood pressure (BP) and early markers of atherosclerosis in pre-pubertal obese children.
Background Hypertension and endothelial dysfunction are premature complications of obesity.
Methods We performed a 3-month randomized controlled trial with a modified crossover design: 44 pre-pubertal obese children (age 8.9 ± 1.5 years) were randomly assigned (1:1) to an exercise (n = 22) or a control group (n = 22). We recruited 22 lean children (age 8.5 ± 1.5 years) for baseline comparison. The exercise group trained 60 min 3 times/week during 3 months, whereas control subjects remained relatively inactive. Then, both groups trained twice/week during 3 months. We assessed changes at 3 and 6 months in office and 24-h BP, arterial intima-media thickness (IMT) and stiffness, endothelial function (flow-mediated dilation), body mass index (BMI), body fat, cardiorespiratory fitness (maximal oxygen consumption [VO2max]), physical activity, and biological markers.
Results Obese children had higher BP, arterial stiffness, body weight, BMI, abdominal fat, insulin resistance indexes, and C-reactive protein levels, and lower flow-mediated dilation, VO2max, physical activity, and high-density lipoprotein cholesterol levels than lean subjects. At 3 months, we observed significant changes in 24-h systolic BP (exercise −6.9 ± 13.5 mm Hg vs. control 3.8 ± 7.9 mm Hg, −0.8 ± 1.5 standard deviation score [SDS] vs. 0.4 ± 0.8 SDS), diastolic BP (−0.5 ± 1.0 SDS vs. 0 ± 1.4 SDS), hypertension rate (−12% vs. −1%), office BP, BMI z-score, abdominal fat, and VO2max. At 6 months, change differences in arterial stiffness and IMT were significant.
Conclusions A regular physical activity program reduces BP, arterial stiffness, and abdominal fat; increases cardiorespiratory fitness; and delays arterial wall remodeling in pre-pubertal obese children. (Effects of Aerobic Exercise Training on Arterial Function and Insulin Resistance Syndrome in Obese Children: A Randomized Controlled Trial; NCT00801645)
The rising prevalence of childhood obesity represents a major public health crisis, because it is associated with considerable risks to the child's present and future health (1). The appearance of pediatric forms of chronic diseases such as hypertension (HTN), early signs of atherosclerosis, or type 2 diabetes contributes to increased risks in adult life (2–4). Hypertension is indeed considered the most important cardiovascular disease (CVD) risk factor worldwide, contributing to one-half of the coronary heart disease and approximately two-thirds of the cerebrovascular disease burdens (5).
In children and adolescents, body mass index (BMI) is strongly related to high blood pressure (BP) (6–8), and the prevalence of HTN ranges from 47% to 62% in obese pediatric patients (9–11). Several factors are believed to contribute to increased vascular tone in obese individuals, including activation of the sympathetic nervous system (12–15), insulin resistance (16,17), and endothelial and smooth muscle cell dysfunctions (18–21).
The structure and function of large arteries can now be assessed by noninvasive, high-resolution ultrasound, and endothelial dysfunction is considered an early marker of atherosclerosis that precedes the plaque formation (19,22). We and others demonstrated impaired arterial endothelial and smooth muscle cell functions and the presence of arterial stiffness in obese children and adolescents (18–20,23). Vascular dysfunction develops even before puberty, in association with elevated systemic BP (21), whereas increased intima-media thickness (IMT) appears later during puberty. The appearance of HTN and first signs of atherosclerosis in young obese children forecasts a major health burden and a need for early intervention strategies.
The benefits of physical activity in the prevention and treatment of cardiovascular diseases have been very well described in adults (24). Individuals who perform regular physical activity have a lower risk of HTN (25), and both acute and chronic physical activity lower BP in patients with HTN, including those who are obese (26). On the contrary, low cardiorespiratory fitness is associated with a 1.3- and 2.6-fold increased risk for HTN in lean and overweight young adults, respectively (27). In children, a recent population-based study demonstrated that higher levels of physical activity were associated with lower BP, and results suggested that the volume was more important than the intensity (28). However, the cross-sectional design did not allow establishing the causality.
Obese children usually spend less time in moderate and vigorous physical activities and have lower cardiorespiratory fitness that their non-obese counterparts (29). The treatment usually includes exercise, dietary, and behavioral interventions. However, most randomized controlled trials were designed to assess the effects of combined lifestyle changes in obese youth (23,30), and only a few authors investigated the impact of exercise alone (31–36). In obese adolescents, they reported significant changes in body fat or visceral fat, endothelial function, IMT, parasympathetic nervous activity, lipids profile, and insulin resistance indexes. However, the effects of physical activity on systemic BP remain controversial, and to our knowledge, there is no information in children before puberty.
Therefore, the primary purpose of this study was to determine the effects of a physical activity program on systemic BP and early markers of atherosclerosis (endothelial and smooth muscle cell functions, IMT, and arterial stiffness) in pre-pubertal obese children.
Study design and subjects
We performed a randomized controlled trial including 44 pre-pubertal obese children ages 6 to 11 years. All patients were recruited at the Obesity Clinic of the Children's Hospital of Geneva, if their BMI was over the 97th age- and sex-specific percentile (37). To allow baseline comparison, we recruited 22 lean volunteers among 165 children in a local elementary school, if their BMI was within the normal range. During the first visit, we assessed pubertal stage by clinical examination according to the method of Tanner, and subjects with a stage above 1 were excluded. With these restrictive exclusion criteria, we do not have to consider the effect of sex on body fat gain at the time of puberty, generally observed in adolescent girls (38). Subjects were excluded if they: 1) were involved in any weight control, physical activity, or behavioral therapy; 2) had a familial history of dyslipidemia or essential HTN; 3) took any medications or hormones that might influence cardiovascular function, body composition, or lipid or glucose metabolism; 4) had an orthopedic affection limiting physical activity; 5) had a genetic disorder or a chronic disease; or 6) followed a therapy for psychiatric problems.
The Mother and Child Ethics Committee of the University Hospitals of Geneva approved this study, and informed written consent was obtained from both parent and child.
After the initial evaluation, the 44 eligible obese subjects were randomly assigned (1:1) to an exercise (EX, n = 22) or control (CON, n = 22) group (Fig. 1).We used closed envelopes containing 50% exercise and 50% control group assignments. The randomization process resulted in a similar girl/boy ratio in each group.
The 3-month intervention consisted of a physical activity program including 3 60-min sessions/week after school hours (total 180 min/week in addition to physical education 135 min/week), without any dietary intervention. Training sessions were supervised by 2 experienced physical education teachers and consisted of 30 min of aerobic exercise (fast walking, running, ball games, or swimming) at a heart rate corresponding to 55% to 65% of individual maximal cardiorespiratory fitness (based on baseline maximal oxygen consumption [VO2max] measures by treadmill test), followed by 20 min of strengthening exercises and 10 min of stretching and cool-down. Children wore a heart rate monitor (Polar S610, Kempele, Finland) during each training session, and watch alarms warned them if the heart rate was too low (<55% VO2max) or too high (>65% VO2max). The aerobic period was followed by strengthening exercises of the arms, legs, and trunk (2 to 3 series of 10 to 15 repetitions), with the resistance being provided by the child's own body weight and elastic bands. Subject compliance to training was determined as the proportion of attended sessions, and adherence was determined as the proportion of subjects who completed all training and testing sessions. The CON group did not participate in any intervention and was asked to maintain the current level of physical activity during the first 3 months. After the 3-month intervention, the EX group was invited to continue to exercise twice/week for a total of 6 months. The CON group was also asked to start a similar training program (twice/week) after the 3-month control period (modified crossover design), to avoid a long waiting period before receiving treatment.
At baseline, participants visited the Children's Hospital from 8:00 amto 12:00 am. Obese and lean subjects underwent identical testing, and the protocol was repeated twice at 3 and 6 months in obese patients only. Observers were blinded to subject grouping. We measured body weight to the nearest 0.1 kg with light clothes with an electronic scale (Seca 701, Hamburg, Germany) and height to the nearest 0.1 cm with a Harpenden stadiometer. We calculated BMI as weight/height squared (kg × m−2) and determined BMI z-scores (37). Total body and abdominal fat mass and fat-free mass (kg) were assessed with dual-energy X-ray absorptiometry (DXA, GE Lunar Prodigy, Lunar Corp., Madison, Wisconsin). Manual analysis, with the “regions of interest” feature, was performed on total body scans to assess abdominal fat. The upper border was defined as the distal margin of the lower ribs, and the lower border was defined as just superior to the supra-iliac crest. The lateral margins were placed outside the body, so that all abdominal but no arm tissue was included (39). The intra-class correlation for repeated measurements of body fat mass was 0.998 in our laboratory.
The brachial resting BP was measured 3 times at a 2-min interval after 10 min of rest in supine position with the back supported, with a validated automated device (Colin Press-Mate BP 8800C, San Antonio, Texas). The cuff covered at least two-thirds of the length of the upper arm, with the length of the bladder wrapping the arm circumference. The average BP was calculated, and HTN was defined as BP >95th sex-, age-, and height-specific percentiles (40).
We also performed a 24-h ambulatory BP monitoring with an automatic monitor and a standing position captor (Dyasis Integra II, Physicor S.A., Paris, France) that has been validated by the British Hypertension Society. This method is considered the gold standard for BP assessment in children (41,42). Measures were taken on the nondominant arm with previously published methods (11). We obtained valid recordings in 42 obese and 15 lean children. The BP analysis was performed off-line after the intervention, without knowing the subject group assignment or time. We calculated the 24-h mean BP and BP standard deviation scores (SDS) (43). Hypertension was defined as 24-h BP over the 95th age- and sex-specific percentile and a BP load over 25%.
Noninvasive measurements of the IMT at the common carotid artery and the endothelial and smooth muscle cell functions at the brachial artery were performed by the same investigator (Y.A.) with a real-time B-mode ultrasound imager (Vingmed CFM800C System Ltd., Hovik, Norway) with a 10-MHz linear high-resolution vascular probe (19,22). All off-line measurements of IMT were performed with an automated computerized program (Iôtec System, Iôdata Processing, Paris, France). The pulse wave of the radial artery was assessed with an applanation tonometry probe (SphygmoCor, Atcor Medical Ltd., West Ryde, Australia) to estimate central aortic pressure noninvasively. This parameter allowed determination of arterial stiffness with the incremental elastic modulus (Einc) (44). After 30 min of rest in a recumbent position, we measured endothelial (flow-mediated dilation [FMD]) and smooth muscle (nitroglycerin-mediated dilation [NTGMD]) functions of the right brachial artery with the same echographic vascular linear probe (19,44).
Cardiorespiratory fitness was measured as VO2max assessed by direct gas analysis (Vmax Spectra, Vyasis Healthcare, GE, Yorba Linda, California) during a multistage treadmill test (Marquette 2000, GE, Milwaukee, Wisconsin). After a sufficient warm-up, the subject walked or ran on the treadmill at a constant speed, which varied by age and physical capacity (from 5 to 9 km × h−1) (45). The grade of the treadmill was increased by 2.5% every 2 min until the subject was exhausted and reached the VO2max criteria.
We measured physical activity count before the intervention and at 3 and 6 months with a validated uniaxial accelerometer (Actigraph MT 6471, MTI, Pensacola, Florida) (46), worn at the right waist during a 7-day period (school week, 24 h/day, outside training periods), except during bathing or swimming. The monitor used a 1-min cycle, and at the end of each cycle, the summed value was stored in memory. The Excel software (Microsoft, Redmond, Washington) was used for data reduction and analysis. Only periods between 8:00 amand 9:00 pmwere analyzed. The “zero activity” periods of 20 min or longer were interpreted as being due to unworn accelerometers and were removed from the total activity count. Subject's data were taken into account if total daily counts/min were under 4,000 and if the monitor was worn during at least 4 days (including 1 weekend day). Leisure-time physical activity in the previous 12 months was assessed with the Modifiable Physical Activity Questionnaire for Adolescents (47).
Nutrition was assessed with a self-reported 3-day food record (2 weekdays and 1 weekend day successively) filled out by the parents. We provided a kitchen scale (precision 1 g, Seca Culina 852) to perform home measures. We calculated the total energy (kcal) and macronutrients intakes.
Blood samples were collected via venipuncture after a 10-h overnight fast. Total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and triglyceride (TG) levels (mmol × l−1) were determined by standard automotive techniques (SYNCHRON LX20, Beckman Coulter, Fullerton, California). Low-density lipoprotein cholesterol (LDL-C) was calculated with the Friedewald's formula. Plasma insulin concentrations were measured by radioimmunoassay (Access ultrasensitive insulin, Beckman Coulter Ireland, Inc., Galway, Ireland), and insulin resistance was assessed by using the homeostasis model assessment (HOMA-IR), according to the equation: HOMA-IR = fasting insulin (μU × ml−1) × fasting glucose (mmol × l−1)/22.5. High-sensitive C-reactive protein levels were assessed by nephelometry (Beckman Coulter Ireland, Inc.). Results were considered abnormal if: TC >5.2 mmol × l−1, LDL-C >3.4 mmol × l−1, HDL-C ≤0.9 mmol × l−1, TG >1.7 mmol × l−1, glucose ≥5.6 mmol × l−1, and insulin >15 μU × ml−1, according to recent pediatric guidelines (48,49).
Sample size and statistical analysis
The calculation of sample size was based on the assessment of systemic BP as the primary end point. In adults, a reduction of 5 mm Hg in diastolic BP is associated with a reduction of at least 35% of stroke and 21% of congenital heart disease (50). In obese adolescents, a 6-month exercise training program resulted in a reduction of systolic blood pressure (SBP) of 8 mm Hg (effect size 0.6) (36). Therefore, with an anticipated effect size of 0.6, a sample size of 22 subjects in each group could detect statistically significant differences at p < 0.05 with a statistical power of 80%. On the basis of previous studies, this sample size was also sufficient to identify differences in endothelial function parameters.
Data were screened for normality, with skewness and kurtosis tests. The following variables were transformed and successfully normalized: office SBP (square), total and LDL-C (reciprocal); and insulin, Einc, FMD, and NTGMD (square-root). Data were expressed as mean and SD or median and interquartile range (25th to 75th percentile), when appropriate. At baseline, we compared means of each continuous variable of obese and lean groups by analysis of variance (ANOVA) with Bonferroni post hoc tests and analysis of covariance (ANCOVA), while adjusting for sex and age entered into the model as a continuous variable. We also compared the shapes of distributions with Mann-Whitney Utests. Except for physical activity count, there were no age and sex effects in ANCOVA, and Mann-Whitney Utests yielded similar results. The associations between continuous variables were determined with Pearson correlation coefficient. To assess the relationships between changes in BP and body fat, we also performed univariate and multivariate linear regression analysis.
Then, ANOVA with a repeated-measure design was used to evaluate parameter changes over time (2 or 3 visits) in obese subjects while assessing the effects of intervention. From baseline to 3 months (randomized controlled trial), we performed a per-protocol analysis (n = 41) and an intention-to-treat (n = 44) analysis, which yielded the same results. For 3 subjects in the obese CON group with no follow-up results at 3 months, we carried forward their baseline values for all the studied parameters (n = 44). We also adjusted the same models for age, sex, and baseline outcome with repeated-measure design ANCOVA, but results were not different, except for physical activity count. We also completed an efficacy analysis including subjects who met the criteria for exposure to the intervention (>60% attendance). As groups differed in age- and sex-ratio after excluding subjects who did not meet criteria for exposure, these variables were entered in the computation.
From 3 to 6 months, we performed a per-protocol analysis only (n = 38), because it was an open study. We used the statistical software program Stata release 9.2 (Stata Corp., College Station, Texas). Differences were considered significant when p < 0.05.
Comparison between obese and lean subjects at baseline
Recruitment and randomization procedures are summarized in Figure 1. Randomization produced similar distributions of baseline characteristics and biological markers in the 2 obese groups (Table 1).Compared with the lean group, the 2 obese groups had significantly higher body weight, BMI, BMI z-score, percent of total and abdominal fat, fat-free mass, fasting insulin, and HOMA-IR and C-reactive protein levels, as well as lower HDL-C, physical activity level, and VO2max. The proportion of obese children who had dyslipidemia or increased insulin resistance indexes compared with normative references was as follows: TC (14%), HDL-C (19%), LDL-C (15%), and fasting insulin (22% between 15 and 28 μU × ml−1). None of the obese subjects had increased glucose or TG, and all lean children had normal values. Nutrition analysis (energy and macro-nutriments) was not possible due to insufficient filling-out of the food record in more than 50% of cases.
Blood pressure, cardiorespiratory fitness, and vascular parameter results at baseline are presented in Table 2.Obese children had significantly higher 24-h SBP and DBP, BP SDS, and arterial stiffness, as well as lower VO2max, FMD, and NTGMD compared with lean subjects. Office BP was significantly higher in EX than lean children. Moderate systolic HTN was present in 20 of 42 obese subjects (47.6%), and both systolic and diastolic HTN was present in 12 of 42 subjects (28.6%). The proportion of children with systolic HTN was similar between obese EX (50%) and CON (45%).
Effects of the 3-month exercise intervention in obese children
The compliance averaged 83% of the 36 scheduled exercise sessions in the obese EX group. The adherence was high; 19 of 22 obese subjects (86%) participated in 3 sessions/week. Three of 22 obese CON (14%) subjects did not attend the second visit, but there were no drop-outs in the EX group.
Changes and treatment effects at 3 months are shown in Tables 3 and 4⇓⇓(intention-to-treat analysis, n = 44). In the EX group, BMI z-score (−5.5%), whole body (−3.6%) and abdominal fat (−4.2%), TC (−3.7%), LDL-C (−4.2%), HDL-C (−5.3%), office SBP (−2.0%) and DBP (−4.1%), and 24-h SBP (−4.9%) and DBP (−3.2%) significantly decreased, whereas fat-free mass (+4.6%) and VO2max (+6.0%) increased during the intervention (p < 0.05). Mean systolic BP also decreased during the night (−7.1 ± 10.2 mm Hg, p = 0.03).
We found opposite results in the CON group, where BMI (+1.1%), whole body fat (+2.2%), abdominal fat (+2.1%), and office (+4.3%) and 24-h SBP (+3.6%) increased, whereas VO2max (−4.3%) and physical activity count (−11.6%) decreased (p < 0.05). Body weight, height, insulin, and HOMA-IR significantly augmented in the 2 obese groups.
The change differences between the obese EX and CON groups (treatment effect) in BMI, BMI z-score, whole body and abdominal fat, fat-free mass, office and 24-h SBP and DBP, VO2max, and physical activity count were all significant (p < 0.05). The prevalence of HTN reduced in EX (−12%) compared with CON (−1%) subjects. In hypertensive EX subjects (11 of 22), the reduction of 24-h SBP (−9.9 ± 14.5 mm Hg, −1.3 ± 1.5 SDS) was significantly greater than in nonhypertensive EX subjects. Changes in 24-h SBP correlated with changes in abdominal fat only (r = 0.42, p = 0.02). We did not observe any changes in biological or arterial function parameters. The efficacy analysis, after excluding the 3 obese EX subjects who participated in <60% of sessions, showed similar results.
Only 18 of 44 obese subjects had valid food records before and after the intervention (9 EX and 9 CON). We did not observe any significant between-group differences in the total energy intake either before (EX 1,813.0 ± 306.6 kcal vs. CON 1,614.0 ± 382.5 kcal) or after the intervention (EX 1,465.2 ± 298.0 kcal vs. CON 1,610.0 ± 468.2 kcal). Exercise training was associated with a decrease in total energy intake in the EX group but not in the CON group (Table 3). The EX–CON effect was not significant, and the percentage of total kilocalories from fat, protein, or carbohydrate did not change in either the EX or the CON group.
To assess the relationships between changes in BP and body fat, we performed univariate and multivariate regression analysis including all subjects (n = 44). We observed associations between 24-h SBP and BMI changes (coefficient 4.8%, p = 0.05), total body fat (coefficient 2.2%, p = 0.02) and abdominal fat (coefficient 2.0%, p = 0.003), as well as 24-h DBP and abdominal fat changes (coefficient 0.8%, p = 0.003). However, after introducing intervention and disease in the model, the intervention remained the only significant determinant of BP changes. We then adjusted the model for sex and age (or used 24-h SBP SDS as dependent variable) and found similar results. Office BP changes were not related to BMI or body fat changes.
Evaluation at 6 months
After the 3-month intervention, we encouraged obese EX participants to continue to train twice/week for a total of 6 months. Four of 22 subjects dropped-out, 19 continued, and 16 of them participated regularly (>60% of sessions) (Fig. 1). The CON group was invited to start a similar exercise program for 3 months. One of 19 obese subjects dropped-out, 18 started training, and 12 of them trained regularly.
From 3 to 6 months, we observed further changes in the EX group (n = 18) (Fig. 2)in BMI z-score (−0.03 ± 0.05, p = 0.01), whole body fat (−0.5 ± 1.1%, p = 0.02), fat-free mass (0.5 ± 0.9 kg, p = 0.02), glucose (−0.2 ± 0.3 mmol × l−1, p = 0.003), and HOMA-IR (−0.96 ± 1.02, p = 0.0003). In the CON group (n = 18), the exercise intervention also resulted in reduced BMI z-score (−0.08 ± 0.14, p = 0.009), whole body fat (−1.1 ± 2.0%, p = 0.02), abdominal fat (−1.0 ± 2.0%, p = 0.05), and office DBP (−4.4 ± 8.2 mm Hg, p = 0.02) and increased fat-free mass (1.1 ± 0.5 kg, p < 0.0001), physical activity count (77.5 ± 107.9, p = 0.03), and HDL-C (0.06 ± 0.15, p = 0.05). Body weight and height significantly increased in both groups.
From baseline to 6 months (Fig. 2), change differences between EX and CON groups were significant for SBP (−1.8 SDS, p = 0.04), DBP (−1.5 SDS, p = 0.047), arterial stiffness (−470.0 mm Hg × 102, p = 0.049), IMT (−0.02 mm, p = 0.045), and VO2max (+2.8, p = 0.02).
Hypertension and first signs of atherosclerosis develops before puberty in obese children, and therefore it is urgent to identify effective strategies to prevent the spread of CVD in this population. To our knowledge, we demonstrate for the first time that a regular physical activity program results in a significant reduction of systemic BP after 3 months and a decrease of arterial stiffness as well as a stabilization of the IMT after 6 months in pre-pubertal obese children. These changes are independent of body weight or fat reduction and are of greater magnitude in hypertensive subjects. We also show that our exercise intervention has beneficial effects on whole body and abdominal fat, fat-free mass, and cardiorespiratory fitness in obese children.
Systemic HTN is a common complication of childhood obesity. In a recent study, we reported an increased prevalence of HTN in pre-pubertal obese children, assessed by ambulatory BP monitoring (11). Elevated BP was associated with endothelial and smooth muscle cell dysfunction, arterial stiffness (21), and increased left ventricular mass (11). In this present study, we observed a significant decrease in systolic and diastolic office and 24-h BP after 3 months of exercise training. The effects on BP were clinically relevant, ranging from −7 to −12 mm Hg for SBP and from −2 to −7 mm Hg for DBP, as previously shown in adolescents (36) and adults (26). In addition, 24-h SBP- and DBP-SDS treatment effects were −1.2 and −0.5 at 3 months, respectively. These changes were significantly greater in hypertensive subjects compared with non-hypertensive subjects. In the EX group, the proportion of children with HTN reduced significantly from 50% to 37% at 3 months and to 29% at 6 months. If we hypothesize that these changes translate into those of similar magnitude in adulthood, these finding could be of public health significance. For example, in adult population a reduction of 5 mm Hg in diastolic BP is associated with a decrease of at least 35% of stroke and 21% of congenital heart disease (50). For pediatricians or general practitioners, the diagnosis of systemic HTN in young obese children might raise the critical question of whether they should prescribe antihypertensive medications or promote lifestyle modifications. Because HTN is a consequence of obesity and physical inactivity, we suggest encouraging regular exercise.
Systemic HTN in obese children might be due to impaired endothelial function or activation of the sympathetic nervous system or insulin resistance (12–21). At baseline, we showed impaired arterial endothelial and smooth muscle cell functions, increased local arterial stiffness, and insulin resistance indexes in obese compared with lean children, without significant evidence of arterial remodeling (IMT). Surprisingly, the changes in BP that we observed during the 3-month intervention were not associated with improved endothelial or smooth muscle cell functions or arterial stiffness in obese subjects. After 6 months, however, we found a significant reduction in arterial stiffness as well as a stabilization of IMT in the EX compared with the CON group. Some authors previously reported improved endothelial function after an 8-week aerobic exercise program in a small cohort of 14 adolescents, even in absence of changes in subcutaneous fat, body weight, or BMI (51). However, subjects with HTN were excluded from that study. More recently, Meyer et al. (36) demonstrated increased endothelial function (+57%) and decreased SBP, IMT, and insulin resistance indexes after a 6-month moderate exercise training program (3 sessions/week, 210 min total) in obese adolescents. Baseline SBP was elevated, similarly to our study, whereas FMD was lower compared with lean subjects. Another trial investigated the effects of a 6-week physical activity program combined with a dietary restriction in 10-year-old children and found improved endothelial function (52). However, children had normal BP at baseline. Our intervention seems to be insufficient to improve endothelial or smooth muscle function in a cohort of patients with a high prevalence of HTN. A longer program might be needed to detect changes at very early stages of atherosclerosis.
Insulin resistance has been incriminated in the development of HTN in obese patients (16,17). During the first 3 months of intervention, we did not observe any reduction in glucose and insulin levels or HOMA-IR. From 3 to 6 months, however, we reported a significant decrease of insulin resistance indexes in both groups. These findings suggest that exercise might have variable effects in young children, probably because their insulin levels are only slightly increased, close to the normal range. In addition, we found the opposite within-group differences for HDL-C in the EX group from baseline to 3 months compared with the CON group from 3 to 6 months (crossover). However, between-group differences were not significant. Because dyslipidemia was present in only a small proportion of children, we might hypothesize that our sample size was too small to detect such changes. To date, the effects of lifestyle modifications on markers of the metabolic syndrome remain controversial. Only 2 studies showed significant responses to moderate physical activity (32,36) or behavioral intervention in adolescents (30), but no studies have demonstrated effects of exercise on biological markers in obese children before puberty.
In obese patients, HTN might also be due to altered autonomous system activity (12–15), although we did not measure these parameters in our study. Gutin et al. (14,31,32,53) completed a 4-month moderate aerobic exercise randomized controlled trial in children ages 7 to 11 years and reported significant changes in cardiac parasympathetic activity as well as in fasting insulin.
The main objective of the treatment of childhood obesity is to stabilize body weight during growth, in order to reduce BMI, body fat, and comorbidities. Therapeutic programs are generally multidisciplinary and include nutritional, exercise, and behavioral components. Our baseline results confirm that physical activity level, sports participation, and cardiorespiratory fitness are significantly lower in obese compared with lean children. Our intervention included various types of activities, such as ball games, walking, and swimming, and participants showed great enthusiasm during sessions. The absence of drop-out and the high compliance and adherence rates as well as the number of children who decided to continue training suggest that our program was adapted and attractive. We assessed the physical activity level before and during the week after the 3-month intervention, which started in September and ended in December. We found that the level remained stable in the EX group, whereas it decreased in the CON group. Physical activity level usually decreases during autumn and winter in Europe, but the obese children who did the intervention maintained their level after the program, despite seasonal variations. We also observed a significant increase in cardiorespiratory fitness in the exercise group and might hypothesize that patients improved their functional capacity as well as their pleasure in participating in daily activities. We conclude that physical activity programs for obese children might be an interesting approach to increase their physical activity level and motivate them in the first phase of treatment.
We found significant changes in BMI z-scores and whole body and abdominal fat, despite an increase in body weight. These results might be explained by the gain in fat-free mass that we observed in the EX compared with the CON group. Indeed, BMI does not assess body composition, and care should be taken in using this measure as the main outcome of treatment. Gutin et al. also reported reduced percent body fat and visceral and subcutaneous abdominal adipose tissue after exercise training, whereas others did not (35,36). Savoye et al. (30) performed a family-based behavioral intervention in obese children and adolescents and found changes of similar magnitude compared with our results. Because visceral fat deposition is closely associated with CVD risk factors and type 2 diabetes, we suggest that future pediatric intervention studies assess changes on central adiposity parameters (e.g., waist circumference, visceral fat) instead of BMI.
Study strengths and limitations
This study has several strengths. We used a randomized controlled trial design, which allowed us to determine the effects of an exercise training program during 3 months. The compliance and adherence were high and the drop-out rate was low in the EX group. However, our study had also several limitations. We chose a modified crossover design because we felt that it was unethical to leave obese children without treatment for more than 3 months. This design did not permit assessing EX–CON treatment effects over the 6-month period. In addition, we assessed the energy intake by a self-reported food record and obtained only 18 of 44 valid datasets. Although the differences were not statistically significant between groups, the sample sizes (9 in each group) were small. The CON group changed little in overall calorie intake, but the EX group decreased their intake by 348 calories/day. We hypothesize that subjects who participate in a physical activity program 3 times/week spend less time at home and therefore decrease their food intake. Changes in diet could therefore be responsible for some of the changes attributed solely to exercise, suggesting that our intervention could have had an unintended but beneficial effect on diet. This could have important clinical implications but also indicates that physical activity change alone might not have the desired results without a related change in diet. However, nutritional results should be interpreted with caution, because subjects with valid diet records might be more motivated to make lifestyle changes and might not be representative of the entire group. Our study also shows that the decrease in systemic BP during the intervention is independent of body weight or fat changes, supporting a direct effect of exercise on HTN reduction.
Our study demonstrates that a regular physical activity program, including mixed aerobic and strengthening exercises at least 60 min 3 times/week (180 min total) during 3 months, significantly reduces systemic BP, BMI z-score, and total and abdominal adiposity and increases fat-free mass and cardiorespiratory fitness in obese pre-pubertal children. Changes in BP are greater in hypertensive subjects and independent of body fat reduction. The continuation of the exercise intervention during 6 months leads to a further decrease in BP, a reduction of arterial stiffness and a stabilization of the IMT. However, exercise has no effects on endothelial or smooth muscle cell functions.
We conclude that participation in physical activity programs should be encouraged in young obese children to reduce systemic BP and prevent the premature development of atherosclerosis. Future research should investigate the volume, intensity, and duration of physical activity that is needed to improve endothelial and smooth muscle functions in children with obesity.
The authors thank the subjects for volunteering for the study, Sophie Bucher Della Torre (dietician), and the staff of the Pediatric Policlinic and the Nuclear Medicine Unit for their assistance.
This work was supported by the Swiss National Science Foundation (#3200B0-103853), Bern, and the Geneva University Hospitals Research and Development Fund.
- Abbreviations and Acronyms
- analysis of covariance
- analysis of variance
- body mass index
- cardiovascular disease
- diastolic blood pressure
- incremental elastic modulus
- flow-mediated dilation
- high-density lipoprotein cholesterol
- homeostasis assessment model of insulin resistance
- intima-media thickness
- low-density lipoprotein cholesterol
- nitroglycerin-mediated dilation
- systolic blood pressure
- standard deviation score
- total cholesterol
- maximal oxygen consumption
- Received April 25, 2009.
- Revision received July 27, 2009.
- Accepted August 6, 2009.
- American College of Cardiology Foundation
- Freedman D.S.,
- Dietz W.H.,
- Srinivasan S.R.,
- Berenson G.S.
- Srinivasan S.R.,
- Myers L.,
- Berenson G.S.
- Paradis G.,
- Lambert M.,
- O'Loughlin J.,
- et al.
- Lurbe E.,
- Alvarez V.,
- Liao Y.,
- et al.
- Rocchini A.P.,
- Katch V.,
- Schork A.,
- Kelch R.P.
- Daniels S.R.,
- Kimball T.R.,
- Khoury P.,
- Witt S.,
- Morrison J.A.
- Meyer A.A.,
- Kundt G.,
- Steiner M.,
- Schuff-Werner P.,
- Kienast W.
- Aggoun Y.,
- Farpour-Lambert N.J.,
- Marchand L.M.,
- Golay E.,
- Maggio A.B.,
- Beghetti M.
- Thompson P.D.,
- Buchner D.,
- Pina I.L.,
- et al.
- Gutin B.,
- Barbeau P.,
- Owens S.,
- et al.
- Watts K.,
- Beye P.,
- Siafarikas A.,
- et al.
- Meyer A.A.,
- Kundt G.,
- Lenschow U.,
- Schuff-Werner P.,
- Kienast W.
- Urbina E.,
- Alpert B.,
- Flynn J.,
- et al.
- Aggoun Y.,
- Sidi D.,
- Levy B.I.,
- Lyonnet S.,
- Kachaner J.,
- Bonnet D.
- Blimkie C.J.,
- Cunningham D.A.,
- Nichol P.M.
- Aaron D.J.,
- Kriska A.M.,
- Dearwater S.R.,
- Cauley J.A.,
- Metz K.F.,
- LaPorte R.E.
- Daniels S.R.,
- Greer F.R.
- Williams C.L.,
- Hayman L.L.,
- Daniels S.R.,
- et al.
- Watts K.,
- Beye P.,
- Siafarikas A.,
- et al.
- Woo K.S.,
- Chook P.,
- Yu C.W.,
- et al.