Chapter  170:  Effectiveness of Weight Management Programs in Children and Adolescents

Key Questions

In conjunction with a Technical Expert Panel, we developed a set of five key research questions to evaluate the effectiveness and safety of behavioral, pharmacological, and surgical treatments for obese and overweight children and adolescents who were 2 to 18 years old. These research questions addressed various measures of the health impact of treatments to reduce or stabilize weight, including: short-term impacts on weight control (6 to12 months after enrolling in treatment); maintenance of weight changes in the medium-term (between 1 to 5 years after enrollment) or longer term (5 or more years after enrollment); adverse effects of treatment (immediate and over time); beneficial effects of treatment, aside from weight control or weight loss; and treatment components or other factors that influence the effectiveness of treatments.

Literature Searches

In 2006, the National Institute for Health and Clinical Excellence (NICE) published a comprehensive report based on a good-quality systematic review of obesity in adults and children including literature published through December, 2005. Relevant portions of this report served as a basis for our literature search, supplemented by another good-quality review of pharmacological treatments. We also conducted update searches in Ovid MEDLINE®, PsycINFO, Database of Abstracts of Reviews of Effects, the Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, and Education Resources Information Center from 2005 (2003 for pharmacological treatments) through December 11, 2007. We supplemented these literature searches (and use of existing systematic reviews) by evaluating citations from several other good-quality reviews of childhood obesity treatment, suggestions from experts, and reviewing reference lists of included trials.

We searched for trials that used a control group and evaluated behavioral and/or pharmacological treatments for weight reduction or stabilization in overweight and obese children who were 2 to 18 years old. Since we could not find any surgical trials or studies that used a control group, we searched for systematically selected case series of children or adolescents undergoing bariatric surgical treatment to determine the immediate and longer-term effectiveness and harms of different types of bariatric surgeries. We also searched for, but did not find, large observational studies to consider adverse events related to behavioral, pharmacological, and surgical weight reduction treatments.

Literature Review and Data Abstraction

Two investigators independently reviewed 2355 abstracts and 338 articles against pre-specified inclusion/exclusion criteria for each key question. Discrepancies were resolved by consensus. We required that studies be designed to promote weight management and reported weight outcomes a minimum of 6 months after treatment began, although we included immediate harms when reported. We excluded studies of children with idiosyncratic weight management issues, such as genetic conditions that affect weight or eating disorders. One pharmacological agent (mazindol) and one type of bariatric surgery (jejunal-ileal bypass) were excluded because they are no longer used. Behavioral intervention trials were required to include a minimal or no-treatment arm to establish absolute effectiveness. For evaluating specific treatment components, however, we also included comparative effectiveness trials to help clarify how specific components affect overall treatment. Trials of pharmacological treatment were required to include a pill placebo control condition. Most trials also included a behavioral intervention for both active and placebo groups. All systematically selected surgical case series were permitted. For all included articles, key elements regarding patient characteristics, treatment components, weight-related outcomes, adverse treatment effects, treatment effects on co-morbidities, and elements related to study design and execution were abstracted into standard evidence tables. For behavioral intervention trials, treatment intensity (hours of contact) was categorized as very low (less than 10 hours), low (10 to 25 hours), medium (26 to 75 hours), or high (over 75 hours). Two investigators quality rated articles using design-specific criteria, with discrepancies resolved by consensus. Articles rated poor quality were excluded, except in the case of bariatric surgeries where, due to very limited data, we retained all surgical case series.

Literature Synthesis

Data were synthesized using quantitative methods, when possible. For most questions, however, we relied on qualitative synthesis due to significant heterogeneity in setting, age range, intervention approach, weight outcome reported, and timing of outcome. We modeled typical cases to more clearly describe the magnitude of weight change in pounds. In these cases, we used growth charts published by the Centers for Disease Control and Prevention to estimate average height for the average age of the participants in a trial, and then converted body mass index (BMI) and measures of relative weight (such as percentile scores) to estimated average weight in pounds, based on average height.

Behavioral Interventions

We identified 18 fair- or good-quality trials of behavioral weight management interventions in a total of 1794 obese children and adolescents aged 5 to 18 years. All incorporated a minimal- or no-treatment control group. These trials were conducted in school settings (n=5 studies), specialty health care settings (n=5), primary care (n=2), residential treatment (n=1), child health/sports center (n=1), and the internet (n=1) with three trials conducted in unspecified settings. Behavioral weight management trials varied in other important ways, such as age of participants, intensity and length of the intervention, baseline weight, and treatment approach (e.g., approach to changing diet and physical activity, involvement of the family, role of behavioral management). We also evaluated 14 supplementary trials that did not meet our primary inclusion criteria, but were applicable to some specific key questions. Two of these reported only very short-term (<6-month) outcomes, but were relevant to the question of adverse effects. The remaining 12 all compared two intervention approaches to each other, rather than including a control group, but were relevant for assessing the importance of specific intervention components.

What are the short-term outcomes for behavioral interventions? Sixteen trials reported differences in some measure of weight immediately or within several months after treatment (6 to 12 months after enrollment); these trials enrolled children and adolescents aged 5 to 18 years whose BMI ranged from 20–24 (in trials of children 12 and younger) to 31–35 (in trials of adolescents) on average; these generally represented BMI percentiles above the 97^th percentile. Behavioral interventions in either schools or specialty health care settings produced modest weight changes, reflecting weight loss as well as weight gain prevention. Most participants remained at or above the 95^th percentile after completing the intervention. Intervention effects varied by treatment intensity and setting. In school settings, intervention trials that were mostly of medium intensity reported 0.4 to 2.07 kg/m² difference in mean BMI change from baseline between a total of 191 treated and 247 control-group participants aged 6–14 years, with a pooled estimate of -0.82 kg/m² (CI: -0.46, -1.18) lower BMI in those treated. For an 8-year old boy or girl, this would translate to about a three pound difference (assuming growth of two inches or less) and about a four pound difference for a 12 year old boy or girl under the same growth assumptions. In specialty health care settings, medium-to-high intensity intervention trials reported between 1.9 to 3.3 kg/m² difference in mean BMI change between a total of 299 treated and 126 control-group participants aged 6 to16 years. For an 8-year old boy or girl, the largest achieved BMI difference (3.3 kg/m²) would translate to about 12 to 13 pounds difference, assuming two inches of growth, and about 16.6 to 17.8 pounds difference for a 12-year old boy or girl under the same growth assumptions. For girls aged 16 years, assuming 2 inches of growth, this BMI difference would translate to about 20 pounds, while the difference would be between 22 and 23 pounds for boys aged 16 with two inches of growth. In the most intensive intervention, children and adolescents in a 10-month residential program dropped from 75 percent overweight to 25 percent overweight, compared with a slight increase in overweight in children and adolescents who were on the waiting list for this program.

How well are weight changes maintained after behavioral interventions? Five trials (three in specialty health care, one in schools, one in primary care) reported medium-term weight outcomes, 1 to 5 years since beginning the intervention. Four of these trials suggested modest differences between a total of 632 treated and control patients aged 5 to 19 years after 1 to 5 years. Three of these (one in specialty health care, one in schools, one in primary care) also reported short-term outcomes, so we could evaluate whether short-term changes were maintained. In two of three trials, short-term benefits were largely maintained 12 months later. The third study in primary care that did not maintain short-term benefits was a very low intensity (4 hours), short-duration (3 months) intervention with initially very small intervention effects. Limited evidence suggests that programs providing a lower-intensity intervention targeting maintenance after the end of primary treatment allows greater maintenance of weight loss than programs with little or no maintenance support.

Are behavioral interventions harmful to participants? We found no evidence that behavioral interventions are harmful for participants. Most studies did not report on harms, however, and those that did could address only short-term harms due to length of followup. Based on this limited evidence, studies documented no adverse effects on growth, eating disorder pathology, or mental health, and little risk of exercise-induced injuries among obese children participating in exercise programs.

Do behavioral interventions have positive effects besides weight loss? Behavioral interventions can have a number of positive effects aside from changes in weight. These include reducing adiposity, improving cardiovascular and diabetes risk factors, and increasing physical fitness. Children and adolescents participating in behavioral intervention programs, particularly those that produce greater effects on BMI (such as those in specialty healthcare settings), may also see reduced adiposity. Increased physical fitness was less commonly measured, but was improved, particularly if the treatment involved organized exercise sessions. While some studies showed an impact on a range of risk factors, results were mixed and reporting was limited. Participants in behavioral intervention programs were less obese than in pharmacological or surgical treatments, and thus may have been less likely to have elevated cardiovascular or diabetes risk factors.

Table 7

Effective behavioral interventions for overweight or (more...)

Table 7

Effective behavioral interventions for overweight or obesity

Study reference	Age range, N, Intervention hours (Intensity)	Description of intervention
Short-Term Outcomes
Carrel et al 200569 School	12–13 n=53 67.5 hrs (Medium)	Diet: Nutrition handouts describing food pyramid, recommended servings of food, appropriate portion sizes, healthier food choices, benefits of healthier eating. PA: P.E. class limited to 14 students, personalized to match skill level and encourage participation. Emphasized lifestyle-focused activities (walking, cycling, snowshoeing). Including warm-up, total movement time averaged 42 min per 45-min class (vs. 25 minutes in typical P.E. class) Beh Tx: None Family: None
Kalavainen 200772 School	6–9 n=70 45 hrs (Medium)	Diet: Recommended diet and meal pattern “in line with” general recommendations for Finnish families PA: 15 sessions over 6 months, most sessions included non-competitive physical activities aimed to develop children's motor skills and to motivate them to increase recreational physical activity. Beh Tx: Family-centered group program based on behavioral and solution-focused therapy approaches. Focus on promoting health lifestyle and well-being rather than weight loss. Details of topics covered not provided. Children given workbook and separate group meeting that included both education/counseling and PA. Family: 15 parent behavioral/solution-focused/educational sessions. Parents given treatment manuals and considered agents of change for the family.
Johnston 2007a75 School	10–14 n=71 41.5 hrs (Medium)	Diet: Provided healthy snack 5 days/wk, once/wk nutrition education, focus on healthier food choices, reading labels, controlling portion sizes, categorizing foods as “safety”, “caution”, and “danger” zone foods. Biweekly quizzes and extra tutoring to low-scoring children. PA: First 6 weeks: Activities included sports and fitness drills for building endurance, strength, and flexibility, teach children to maintain heart rate within target zone and develop basic level of fitness. Second 6 weeks: focus on skill development for activities available in neighborhood or school (e.g., basketball, soccer, jumping rope, dance, kickboxing). Beh Tx: Learned to self-monitor, set goals, address self-identified barriers to improving health. Also used token economy, children earned points for trying new fruits and vegetables, keeping bodies moving during physical activity, and meeting program and individual goals. Family: Parents invited to culturally sensitive monthly meetings to teach them how to adapt family meals and activities to facilitate health changes.
Johnston 2007b71 School	10–14 n=60 41.5 hrs (Medium)	Diet: Provided healthy snack 5 days/wk, once/wk nutrition education, focus on healthier food choices, reading labels, controlling portion sizes, categorizing foods as “safety”, “caution”, and “danger” zone foods. Biweekly quizzes and extra tutoring to low-scoring children. PA: First 6 weeks: Activities included sports and fitness drills for building endurance, strength, and flexibility, teach children to maintain heart rate within target zone and develop basic level of fitness. Second 6 weeks: focus on skill development for activities available in neighborhood or school (e.g., basketball, soccer, jumping rope, dance, kickboxing). Beh Tx: Learned to self-monitor, set goals, address self-identified barriers to improving health. Also used token economy, children earned points for trying new fruits and vegetables, keeping bodies moving during physical activity, and meeting program and individual goals. Family: Parents invited to culturally sensitive monthly meetings to teach them how to adapt family meals and activities to facilitate health changes.
Savoye et al 200774 Health Care	8–16 n=174 97.5 hrs (High)	Diet: Non-dieting approach emphasizing low-fat, nutrient-dense foods of moderate portion sizes. PA: Two 50-min sessions/wk for first 6 months, then 1 session every 2 weeks. Each session included warm-up, high-intensity aerobic exercise, cool-down. Goal to sustain 65% to 80% of age-adjusted max heart rate for duration of aerobic exercise. Also encouraged to exercise 3 additional days/week at home and to decrease sedentary behaviors. Beh Tx: One 50-min session/wk for first 6 months, then 1 session every 2 weeks. Topics included self-awareness, goal-setting, stimulus control, coping skills training, cognitive behavior strategies, contingency management. Family: Parents attended separate group during children's behavioral treatment groups. Emphasized parents' role modeling health behavior, coping skills training.
Reinehr et al 200673 Health Care	6–14 n=240 76 hrs (High)	Diet: Recommended diet of 30% fat, 15% protein, 55% carb (only 5% sugar). Categorized foods using Traffic Light system: red=“stop”, yellow=“consider the amount”, green="OK when hungry or thirsty. Total kcal went from 1459 ± 379 pre-treatment to 1250 ± 299 kcal post-treatment PA: Once per week for 12 months, consisted of ballgames, jogging, trampoline, instruction in physical activity as part of everyday life, and encouragement to reduce amount of time spend watching TV Beh Tx: In first 3 months, 6-session nutrition course and 6-session behavior therapy groups for children. Family therapy provided for the next 3 months, with up to 3-month extension as needed. Lifestyle modification approach, details of topic covered not reported. Family: 6-session parents' course for parents, 3 “Talk rounds for parents”, plus family therapy described above.
Gillis 200787 Health Care	7–16 n=27 8 hrs (Very low)	Diet: Two discussions of healthy diet; asked to record food intake once/week. No details of recommended diet reported. PA: Two sessions discussing exercise; asked to record exercise once/week. No details of exercise recommendations reported. Beh Tx: None Family: None
Nemet et al 200579 Health Care (“Child Health and Sports Center”)	Avg age 11.1 n=54 35.75 hrs (Medium)	Diet: 6 one-on-one meetings with a dietitian plus four group lectures, covering reasons for childhood obesity, nutrition information such as the food pyramid, food labels, food preparation, eating habits, regular meals. Recommend balanced diet of 5,021 to 8,368 KJ, a deficit of ~30% from baseline intake, or 15% less than estimated daily required intake. PA: Two 1-hour sessions/week for 14 weeks designed to mimic the type and intensity of exercise that children normally perform. Activities varied in duration and intensity, but usually included activities promoting endurance. Attention given to improving flexibility and coordination. Instructed to exercise at home for additional 30–45 minutes/week and to reduce sedentary activities. Beh Tx: Information on controlling the environment to minimize over-eating, coping with situations that encourage overeating. Family: Varied with child's age. Ages 6–8: parents only for first 2 meetings, children joined thereafter. Ages 8-puberty: parents and children invited to all sessions. Puberty onward: Parents and youth attend first meeting, then alternate parents and child.
Saelens et al 200277 Primary Care	12–16 n=44 3.8 hrs (Very low)	Diet: Adaptation of Traffic Light diet, goal to reduce to ~1200–1500 kcal/day. Focus on reduction in overall quantity of food and increasing healthy eating, with no prohibition of any particular foods. Computer-based assessment used to identify eating habits, develop initial recommendation/plan. Meeting with pediatrician to confirm/modify plan, 11 10–20 minute follow-up phone calls with support staff to discuss food diaries and other behavior change issues. PA: PA also assessed via computer, goals set with pediatrician, encouraged by phone counselors. Monitored PA starting with 5th phone call, goal minimum of 60 minutes of at least moderate intensity PA 5 days/week. Beh Tx: Behavioral skills covered include self-monitoring, goal setting, problem solving, stimulus control, self-reward, and preplanning. Family: Parents sent information sheets corresponding to materials received by youth, highlighting ways in which parents can be most helpful. Recommended parental skills included stimulus/environmental control, positive reinforcement, and preplanning.
Braet et al 200385 Residential	10–17 n=76 3,520 hrs (Very high)	Diet: Fed 30% fat, 15% protein, 55% carb, 1500–1800 kcal/day. Soft drinks, sweets, high-calorie food strictly regulated. PA: Minimum 4hr/wk individual training; “stimulated to exercise 10 h/wk or more if they wanted to.” Beh Tx: 12-wk small group cognitive-behavioral covering self-regulation skills, such as self-observation, self-instruction, self-evaluation, self-reward; problem-solving, coping with high-risk situations, relapse prevention. Followed by weekly personalized problem-solving sessions. Family: Children saw parents every other weekend, plus holidays. Parents received leaflets on how to prepare healthy food, “stimulated” to organize aerobic exercises during weekends and holidays.
Senediak et al 198588 Setting NR	6–12 n=45 12 hrs (Low)	Diet: Covered variety of nutritional and dietary topics, recommended diet based on Food Exchange System and Traffic Light System. PA: Children instructed to engage in at least four 30-minute aerobic exercise sessions per week. Basic conditioning exercises introduced initially, then more strenuous aerobic exercise. Also recommended other lifestyle changes (such as walking instead of riding in the car) to encourage physical activity. Beh Tx: Utilized self-monitoring, self-reinforcement and parental reinforcement, stimulus control techniques (e.g., restricting food consumption to specific times and places), attempted to modify negative cognitions that may contribute to obesity. Family: Both parents and children involved in all sessions, given materials and homework.
Maintenance Outcomes
Mellin et al198782 Health Care	12–18 n=66 24 hrs (Low)	Diet: Sustainable, small changes in diet; very-low-calorie or restrictive diets discouraged. No specific details on recommended diet. PA: Encouraged to make sustainable, small changes in exercise habits. No further details provided. Beh Tx: 14 weekly sessions; self-directed change format, encourage small, sustainable changes in relationships, lifestyle, communication, and attitudes. Details of encouraged change process not described. Family: Two parent meetings; instructed on strategies for supporting their child's weight-loss efforts, including altering family dietary and activity habits, and improving parenting and communication skills.
Flodmark et al, 199381 Health Care	10–11 n=93 I1: 12 hrs I2: 24 hrs (Low)	Diet: Counseling by pediatrician and/or dietitian; recommend 1500 to 1700 kcal, with 30% of calories from fat. PA: No recommendations described Beh Tx: None described. Family: Family therapy focused on reinforcing the resources of the family and creating and optimal emotional climate for helping the obese child. Adjustments to family hierarchy/structure, plus solution-focused therapeutic techniques.

Abbreviations: PA- physical activity; Beh TX - behavioral treatment; PE - physical education

Pharmacological Plus Behavioral Intervention

We found seven fair-to-good quality trials evaluating a pharmacological agent taken over six to twelve months along with behavioral interventions to treat obesity in a total of 1,294 obese adolescents. At baseline, participants met adult criteria for obesity, with mean entry BMI typically between 35 to 38 kg/m². All trials provided behavioral interventions for the adolescents in both treatment arms. All trials involved adolescents age 12 and older, were double-blind, and included a pill placebo control group. Five trials in a total of 715 obese adolescents examined sibutramine and two in a total of 579 examined orlistat. We also found two small trials testing the weight effects of taking the diabetes medication, metformin, for 6 to 12 months in a total of 60 obese children and adolescents with evidence of insulin resistance or hyperinsulinemia. Those reports are not directly applicable to the general population of obese adolescents.

What are the short-term outcomes for pharmacological plus behavioral interventions compared with behavioral interventions alone? Almost all the sibutramine trials found group differences in BMI change. After 6 to 12 months, adolescents treated with sibutramine plus a behavioral intervention reduced their BMI by 1.6 to 2.7 kg/m² more than those in the placebo plus behavioral intervention groups. Weight loss with orlistat was somewhat less: average BMI was 0.5 to 0.85 kg/m² lower after 6 to 12 months in the group taking orlistat plus behavioral intervention than in the placebo plus behavioral intervention group. In the trials of metformin, those taking metformin reduced their BMI by 1.3 to 1.4 kg/m² more than those taking the placebo.

How well are weight changes maintained after pharmacological treatments? No trials assessed maintenance of weight loss after the end of six or twelve months of treatment with sibutramine, orlistat, or metformin.

Are pharmacological treatments harmful to participants? Although no differences were reported in overall adverse events, serious adverse events, or discontinuation due to adverse events, adolescents taking sibutramine were more likely to develop small increases in heart rate and, in some cases, in blood pressure. Among orlistat users, mild-to-moderate gastrointestinal side effects, such as abdominal pain, oily spotting, or fecal urgency, occurred commonly (in 20 to 30 percent), with fecal incontinence reported in 9 percent of adolescents taking orlistat, compared with 1 percent of placebo participants. Limited evidence suggests no impact on growth for either medication. Neither trial of metformin in children and adolescents at risk for diabetes reported any serious adverse events, but these were very small studies.

Do pharmacological treatments have positive effects besides weight loss? Most studies suggested that both sibutramine and orlistat patients had greater reductions in adiposity than the placebo groups. Few other differences in cardiovascular or diabetes risk factors were found in those taking either medication, compared with placebo, except for reported improvements in HDL cholesterol, triglycerides, and insulin resistance/sensitivity among adolescents taking sibutramine in the single largest study. Similarly, in the single large study of orlistat, patients treated with orlistat had a small mean reduction in diastolic blood pressure. Both metformin trials reported improvements in fasting glucose and insulin measures.

What components make pharmacological treatments successful? We found insufficient data on effective pharmacological plus behavioral interventions to describe which components were most effective. Using proven behavioral treatments in conjunction with effective pharmacological agents, and ensuring their delivery, could be an important improvement.

Surgical Treatment

We identified 18 case series reporting on weight change, complications, and other outcomes from weight loss surgical interventions in a total of 612 morbidly obese adolescents, most of whom had failed previous weight management approaches. Where reported, 23 to 62 percent had one or more co-morbidities such as hypertension, diabetes, and dyslipidemia. Six of the studies explored the safety and efficacy of laparoscopic adjustable gastric banding (LAGB) and the remaining focused on gastric bypass procedures. The average ages for surgical patients in these studies ranged from 15 to 18 years. Mean baseline BMI was generally between 43 and 48 kg/m² in LAGB studies and in the high 40s to mid 50s in the gastric bypass studies. Results must be interpreted with caution, however, because loss to followup, incomplete reporting, and small samples limits our confidence in the generalizability of these results.

What are the short-term outcomes for surgical treatment? Morbidly obese adolescents undergoing laparoscopic adjustable gastric banding experienced an average BMI decline of 5.0 to 8.1 kg/m² six months after surgery, and a 9.4 to 10.2 kg/m² decline one year after surgery. Bypass procedures showed somewhat greater weight loss at one year, with average BMI reductions in the 15 to 20 kg/m² range.

How well are weight changes maintained after surgical treatments? Surgical treatments for obese adolescents have only been performed in recent years. In general, patients tend to lose the most weight at around 12 to18 months, after which their weight loss generally stabilizes. While we have only limited data on long-term outcomes, and insufficient data on all individuals, most patients seem to maintain their maximal weight loss after gastric banding (or experience a minimal amount of regain) for two to three years after surgery. One small study in 25 individuals after gastric banding found that BMI decreases were generally maintained 5 years after surgery. While we were only able to find very limited data on Roux-en-Y gastric bypass, based on 33 adolescents, BMI reductions were maintained at 5 years, with some regain suggested by 10 to 14 years. While there are clearly individuals who experience treatment failures, absolute rates for success or failure cannot be estimated with current data.

Are surgical treatments harmful to participants? Roughly 10 to 15 percent of adolescents undergoing laparoscopic adjustable banding require additional surgery for repositioning or removal of the band, but no serious adverse events or deaths were reported. Roux-en-Y gastric bypass is a more invasive procedure and, not surprisingly, appears to have higher rates of adverse effects. Serious adverse effects (involving threat to life or major organ system failures, but no deaths) occurred in approximately 5 percent of patients while in the hospital. In another study, 25 to 39 percent experienced non-life-threatening adverse events requiring additional treatment, special tests, endoscopy, or hospital readmission in the first year after surgery. Very limited numbers of cases and lack of long-term systematic follow-up limits our ability to assign absolute risks, including risk of death, over the longer term.

Do surgical treatments have positive effects besides weight loss? Not all studies measured or reported changes in co-morbidities after surgery. However, all cases of sleep apnea and most cases of reported asthma were resolved after surgery, with reported improvements in many with type II diabetes, hypertension, or dyslipidemia. More complete reporting would be very beneficial in assessing these potential health benefits that occur with weight loss after bariatric surgery in morbidly obese adolescents.

What components make surgical treatments successful? We have insufficient information to determine the relative benefits of different types of surgical approaches. Likewise, we found insufficient data to determine the impact of factors such as surgeon training or patient characteristics.

Recommendations for Future Research

While childhood overweight has been the focus of considerable research in recent years, longer-term followup is needed to confirm maintenance of treatment effects for all types of treatment, but for pharmacological and surgical treatments in particular. Longer term followup should also describe the rate and severity of longer-term adverse effects, particularly for more invasive treatments. Given the central role of behavioral treatments, much more research is needed in this area. Replication of behavioral treatment trials is needed to confirm the benefits of programs and estimate their likely effects in real-world settings. Finally, understanding important components of behavioral interventions is an ongoing need. More studies are needed in minority children and adolescents, as well as in younger children (5 years and under).

Definition and Measurement of Overweight and Obesity in Children and Adolescents

In contrast to colloquial usage, where obesity and overweight generally refer to culturally undesirable body size (“being fat”), these terms represent specific conditions with unique criteria in the medical and scientific literature. While obesity is a condition of excess body fat (adiposity), which is associated with adverse health states and risk for future disease, the medical definition of obesity in children and adolescents is not as straight forward as for adults. At present, there is no universally accepted definition that distinguishes children with normal or healthy weight from those whose level of adiposity is unhealthy. While the presence of obesity in some children and adolescents is obvious with simple observation, it is difficult to determine when a child who is not obviously overweight faces health risks from adiposity. In the absence of a clear, health-based definition of obesity, children are instead categorized as “overweight” and “obese” based on how they compare with a normative sample of children of the same age and sex.

Figure 1. Illustrative BMI percentile chart with table (more...)

   Figure 1. Illustrative BMI percentile chart with table of weight and BMI standard deviation score for selected percentiles: Boys

Figure 2. Illustrative BMI percentile chart with table (more...)

   Figure 2. Illustrative BMI percentile chart with table of weight and BMI standard deviation score for selected percentiles: Girls

Table 1

Definition of overweight and obesity terms for children (more...)

Table 1

Definition of overweight and obesity terms for children and adolescents, and adults

Children and adolescents		Adult
Overweight	85th–94th percentile BMI (age-sex specific)	Overweight 46	BMI 25–29 kg/m2
Obese	≥ 95th percentile BMI or BMI ≥ 30 kg/m2, whichever is lower5	Obesity Class I Class II Class III (also called morbid, severe)	BMI 30–34.9 kg/m2 BMI 35.0–39.9 kg/m2 BMI ≥ 40 kg/m2
Severe obesity5	> 99th percentile BMI	NIH criteria for bariatric surgery in adults40	BMI >40 kg/m2 Or BMI >35 kg/m2 with co-morbidities

Figures 1

2

Figures 1

2

Although it is not a direct measure of adiposity, BMI-for-age percentile measures in boys and girls correlate reasonably well with percentile rankings of directly measured percent body fat (correlations generally between 0.78 to 0.88).⁶ Obesity (primarily defined as BMI ≥ 95th percentile) has also been correlated with childhood health consequences and with risk factors for obesity-related morbidity in adults.⁷–⁹ Since BMI is an imperfect measure of body fat, however, categorizing children and adolescents as obese based on BMI definitions can be problematic. Recent data from the Bogalusa Heart Study found that 35 percent of children aged 5 to 17 years with BMI > 95^th did not have excess body fat.¹⁰ At or above the 99^th percentile, however, almost all (94 percent) had excess adiposity. Those with the highest BMI percentiles (≥ 99^th) were also much more likely to have two or more cardiovascular risk factors (59 percent) compared with those in the broader group at or above the 95^th percentile (39 percent). Noting these differences, experts have recently proposed distinguishing the “severely obese”, defined by the 99^th percentile, as those in particular need of clinical evaluation and treatment.^{11, 12}

Table 2

BMI at 50^th, 85^th, 95^th, and 99^th percentiles and weight (more...)

Table 2

BMI at 50^th, 85^th, 95^th, and 99^th percentiles and weight in pounds for BMI of 25, 30, 35, and 40 at ages 8, 12, and 16 years

		BMI (kg/m2) at percentiles* Children and adolescents				Weight (lbs) at BMI levels** Adults
50th Percentile for Height		Over-weight5	Obesity	Severe Obesity	Over-weight	Obesity Class I	Obesity Class II	Obesity Class III
Age (Sex)	inches	50th	85th	95th	99th	25	30	35	40
8 (Male)	50.5	15.8	17.9	20.0	25.6	91	109	127	145
8 (Female)	50.5	15.8	18.3	20.7	26.4	91	109	127	145
12 (Male)	58.5	17.8	21.0	24.2	31.8	122	146	170	195
12 (Female)	59.5	18.1	21.7	25.2	33.1	126	151	176	201
16 (Male)	68.5	20.5	24.2	27.5	33.9	167	200	234	267
16 (Female)	64	20.4	24.6	28.9	39.1	146	174	204	233

th

Estimated average height for age from 50 percentile on CDC Growth Chart “Stature-for-age percentiles: Boy (or Girls), 2 to 20 years”.

2

Pounds = (BMI × inches) /703 was used to convert from BMI to pounds.

Figures 1

2

Prevalence of Children and Adolescents Obesity in the United States

Between the early 1970s and 2003 to 2004, the prevalence of obesity (defined as age- and sex-specific BMI ≥ 95^th percentile) increased three- to six-fold, depending on age, sex, and ethnicity.¹³ In 2003 to 2004, the prevalence of obesity among 6- to 19 year-old children and adolescents was approximately 16 to 18 percent.^{14, 15} When children and adolescents who are overweight (defined as age- and sex-specific BMI in the 85^th to 94^th percentile) are also included, this prevalence increases to almost one in three children and adolescents identified as overweight or obese (31 to 33 percent).^{13, 15} Looking at the youth with the most severe levels of obesity, 3 to 6 percent of boys aged 13 to 17 years are at or above the 99^th percentile. For girls, the comparable figure is 1 to 3 percent.¹⁰

The prevalence of obesity varies somewhat with age. Children aged 6 to 11 years have the highest prevalence of obesity (18.8 percent), compared with younger children (13.9 percent) and adolescents (17.4 percent), according to data from the 2003 to 2004 National Health and Nutritional Evaluation Survey (NHANES).¹³ Males have slightly higher prevalence of obesity for all age categories. Childhood obesity is increasing all around the world, not just in the United States. A meta-analysis calculated that the annualized change in prevalence of obesity in school children in the United States from 1971 until 2000 was approximately 0.4 percentage points per year.¹⁶ Twenty-three North American, Eastern European, Western European, and Asian countries reporting comparable data also showed increases in childhood obesity, with annualized changes ranging from less than 0.1 percentage points (in Finland and the Netherlands) to over 0.7 percentage points (Singapore and East Germany).¹⁶ The estimated prevalence of overweight (including obesity) in children and adolescents in the Americas as a whole is 27.7 percent. Europe has the next-highest estimate at 25.5 percent, then Eastern Mediterranean countries (23.5 percent), followed by countries in the West Pacific (12.0 percent) and South East Asia (10.6 percent). Prevalence of overweight and obesity are low in African nations (1.6 percent).

High-Risk Groups for Child and Adolescent Obesity and Overweight

In the United States, minority children and adolescents are disproportionately obese and overweight at all ages.¹³ One large nationally representative study using NHANES data found that 43 percent of Mexican-American boys age 6 years or older were obese or overweight, which was higher than nonHispanic White (29 percent) and nonHispanic Black (31 percent) boys in the same age range.¹⁵ Native American boys are also more likely to be obese—39 percent of Native American adolescent boys in the National Longitudinal Study of Adolescent Health (Add Health) were categorized as obese in the mid-1990s, compared with 10 to 15 percent among other ethnic groups.¹⁷ NonHispanic White girls have lower prevalence of obesity or overweight (26 percent), compared with nonHispanic Black (42 percent), and Mexican American (39 percent) girls.¹⁵ These racial/ethnic disparities are consistent with prevalence figures reported by the Add Health study, which reported obesity in Black (18 percent), Hispanic (13 percent), and Native American (14 percent) adolescent girls, compared with Asian (4 percent) and nonHispanic White girls (10 percent). Statistical tests of these differences were not reported. Racial differences are also seen in the persistence of obesity into adulthood among children and adolescents aged 5 to 14 years. One study found that 65 percent of obese White girls and 84 percent of obese Black girls remained obese into adulthood. Results were similar for obese boys (71 percent of White boys versus 82 percent of Black boys).¹⁸

There is also clear correlation between income level and obesity prevalence in White children and adolescents. Obesity prevalence is highest in the lowest income bracket, and those with highest income levels have the lowest obesity prevalence.¹⁹ This correlation is less clear for Black and Hispanic ethnic groups, however, where data suggest no clear linear relationship between income and obesity.¹⁹

Children of obese parents have a higher risk of obesity,²⁰ with children with two obese parents having the highest risk of obesity.²¹ A large-scale epidemiological study published in 1976 found that by age 17, children with two obese parents had three times larger triceps skinfold measures as those with two lean parents.²¹ Compared to children without obese mothers, children with obese mothers are three to ten times more likely to be obese themselves. White and Black children of obese mothers are three times more likely to be obese, Hispanic children of obese mothers are twice as likely to be obese, and Asian children of obese mothers may be as much as ten times more likely to be obese.²² In addition, maternal obesity has been associated with earlier age of obesity onset in children.²²

Health and Psychosocial Consequences of Child and Adolescent Obesity

Although the data on the health and psychosocial consequences of obesity in children and adolescents are almost exclusively observational, and therefore causal relationships cannot be established, there is growing evidence that childhood and adolescent obesity can have a substantial health impact.^{7, 9} While most children will not experience the health consequences of persistent childhood obesity for decades, some of these consequences can occur prior to adulthood, particularly in those who are severely obese.⁹ Obese children and adolescents have a higher risk of type 2 diabetes mellitus, asthma, and nonalcoholic fatty liver disease, are more likely to have cardiovascular risk factors, such as hypertension and hyperlipidemia. These children and adolescents are also more likely to experience other adverse health-related events, such as perioperative adverse respiratory events when undergoing procedures requiring anesthesia.^{7, 9, 23} Obese children may be more likely to experience mental health and psychological issues, such as depression ²⁴ and low self-esteem,^{9, 24, 25} than nonobese children. The risk of mental health issues increases with age and is higher in girls,⁷ likely reflecting the pressures of the social environment. For severely obese children, impacts on quality of life can be severe and other serious conditions such as obstructive sleep apnea, orthopedic problems, infertility, and increased intracranial pressure can occur.^{7, 9, 11, 26}

One of the greatest concerns about childhood obesity is that it may persist into adulthood.²⁷ Adult obesity, in turn, has a detrimental effect on adult health^{2, 28, 29} and mortality.^{28, 30} Other systematic reviews have examined the persistence of obesity from childhood into adulthood.³¹ Factors associated with greater persistence of obesity from childhood into young adulthood included older age and higher BMI (above the 95^th percentile or higher). Recent data from the Bogalusa Heart Study confirm these findings.²⁷

Although it is difficult to distinguish childhood obesity's effects on morbidity and mortality independent of the effect of adult obesity, a systematic review reporting on the long-term consequences of pediatric obesity concluded that obesity-related cardiovascular disease can originate in childhood obesity.⁷ This review, and others, indicate that childhood obesity has also been associated with adverse social and economic outcomes in young adulthood.^{7, 9, 32}

Current Interventions for Child and Adolescent Obesity and Overweight

Behavioral intervention. Behaviorally based interventions are the first-line treatment for overweight and obesity in children and adolescents.¹¹ Behavioral weight management interventions promote weight loss through modifications in diet and activity level without the use of adjuncts, such as pharmacologic agents. Typical behavioral interventions aim to modify food consumption to emphasize healthy eating and reduce consumption of high calorie-low nutrient snack foods and sugary foods and beverages. A range of approaches has been used to encourage more healthy patterns of dietary intake and physical activity, which are discussed in detail elsewhere.^{5, 11, 33} Behavioral interventions often involve parents or entire families, particularly for younger children. Optimally, behavioral interventions include cognitive and behavioral management techniques to help participants initiate and sustain needed lifestyle changes, and a range of approaches have been utilized.^{33, 34} We refer to programs that focus on dietary counseling and brief lifestyle change advice without more extensive use of behavioral management principles as “behavioral counseling” interventions. We use the term “behavioral management intervention” to denote programs that are more extensive and include principles of cognitive and/or behavioral management. We use the term “behavioral intervention” as a general term to refer to both behavioral counseling and management interventions.

Pharmacologic treatment. A number of pharmacological agents are also being used to promote weight loss among obese adults as adjuncts to behavioral intervention. Weight loss drugs can be divided into two main categories based on their putative mechanism of action—appetite suppressants and lipase inhibitors. Appetite suppressants may be divided further based on the specific neurotransmitters they are thought to affect. Sibutramine and orlistat are the two most well studied weight loss drugs among adults. Sibutramine is a centrally acting appetite suppressant that selectively inhibits the reuptake of serotonin and norepinephrine, increasing their levels in the brain. Orlistat is a lipase inhibitor that is thought to promote weight loss by reversibly binding to the active center of the enzyme lipase, preventing digestion and absorption of some dietary fats. It also reduces the absorption of fat-soluble vitamins.

The United States Food and Drug Administration (FDA) has approved some medications for the treatment of obesity in adults. Only one medication has been approved for prescription use in obese children and adolescents (aged 12 and older). Medications not specifically approved for obesity treatment in children and adolescents may be considered for off-label use by physicians. The FDA approved the use of sibutramine and orlistat for the long-term treatment of obese adults in 1997 and 1999, respectively.³⁵ In 2003, the FDA approved orlistat for treatment of overweight among pediatric populations (ages ≥ 12 years).³⁶ In 2007, the FDA also approved orlistat for over-the-counter use among adults ages 18 years and older.³⁷ Several other appetite suppressants are FDA-approved only for short-term treatment of overweight adults (benzphetamine, diethylpropion, phendimetrazine, and phentermine).³⁸ Additional drugs that are not FDA-approved for treating overweight or obesity have been considered as potential weight loss agents such as some antidepressants (fluoxetine, sertraline, and bupropion), antiepileptic drugs (topiramate, zonisamide, lamotrigine), and the antidiabetic biguanide metformin. ³⁸

A recent systematic evidence review found that numerous different drugs produced modest weight loss among adults when combined with dietary recommendations: sibutramine, orlistat, phentermine, bupropion, fluoxetine, topiramate, and probably diethylpropion.³⁹ The additional weight loss attributable to these drugs has been less than five kg at 1 year. The drugs have not been compared directly against each other, and the report found no evidence that any particular drug produced more weight loss than any other. All of the drugs had side effects. Sibutramine was associated with modest increases in heart rate and blood pressure and with preventing decreases in blood pressure that may have occurred with weight loss. Orlistat is associated with numerous gastrointestinal side effects such as diarrhea, flatulence, bloating, abdominal pain, and dyspepsia.

Surgical treatment. Surgical approaches to weight loss (bariatric surgeries) have been developed to treat those in whom more conservative measures have failed. The criteria for undertaking bariatric surgery for adolescents have largely followed expert-based criteria for adults from a 1991 NIH consensus conference,⁴⁰ although expert-based criteria for selecting severely obese adolescents for bariatric surgery have also been published.⁴¹ These criteria specify that surgery be considered for persons who have attained skeletal maturity with a body mass index (BMI) greater than 40 and with high-risk co-morbid conditions responsive to weight loss, such as obstructive sleep apnea or severe diabetes mellitus. Recent followup data from severely obese adults undergoing bariatric surgeries indicate reduced risk factors, such as hypertension, dyslipidemia, or incidence of type 2 diabetes, and reduced all-cause mortality (29 percent).⁴²

Surgeries can induce weight loss through two means—restriction and malabsorption. Restrictive approaches reduce the stomach size to limit the amount of food that can be consumed at a single meal. Malabsorptive approaches bypass portions of the intestines to limit the proportion of calories absorbed from ingested food. In the case of gastric bypass, a very common bariatric procedure, restrictive and malabsorptive approaches are combined. Bariatric surgeries are associated with risks for complications, however, including death. Treatment failures can be caused by inability to tolerate surgery-related changes requiring reversals, or post-surgical changes in behavior or anatomy that in effect override the surgically induced restrictions in stomach size. With the advent of minimally invasive surgery, some risks are reduced when procedures are performed using a laparoscope instead of an open procedure (laparotomy).

Major types of bariatric surgeries include gastric banding, gastroplasties, and bypass procedures. Gastric banding positions a band outside the stomach to create a smaller pouch (15 to 30 cc) in the uppermost portion of the stomach in order to restrict food intake. While bands were fixed in circumference at the time of surgery in the past, they are now adjustable through injection of saline into an accessible subcutaneous port. Adjustable gastric bands can be adjusted over time in response to rates of symptoms and weight loss. Bypass procedures reduce caloric intake (and, unfortunately, absorption of essential nutrients) through rerouting food around a portion of the intestinal tract. The bypassed section is generally not removed, which theoretically allows for reversals. Gastric bypass is the most common bariatric surgery in the United States since other forms of bypass (jejunal-ileal; biliopancreatic diversion; biliopancreatic diversion with duodenal switch) have been associated with numerous complications.^{39, 40}

Surgeons have developed a variety of surgical approaches to gastric bypass, with some variations even for the most commonly performed type, Roux-en-Y gastric bypass (RYGB). RYGB restricts the size of the stomach to about 30 cc and then bypasses the duodenum to reduce absorption. Gastroplasties mechanically reduce the stomach's size and architecture by creating a stapled anterior gastric pouch with a reduced outlet to the remainder of the stomach. Types of gastroplasty include vertical-banded gastroplasty (VBG), which also uses a band to constrict the stomach and prevent dilatation, gastric partitioning with a band, and horizontal gastroplasty. Gastroplasties are less commonly performed, given their higher degree of recidivism than with gastric bypass, and because less invasive restrictive approaches using gastric banding are now available.

Banding approaches, particularly laparoscopic adjustable gastric banding (LAGB), are particularly appealing for adolescents since they do not involve surgical removal or realignment of the intestine and are therefore more reversible. Banding also retains the entire absorptive area of the stomach and intestines, which lowers risk of malabsorption of essential nutrients. Malabsorptive concerns are particularly important since adolescents are still developing and young females could become pregnant. Finally, banding can routinely be done laparoscopically, which reduces peri-operative complication risks. In the United States, however, FDA approval has not been granted for these devices in those under 18 years.⁴³

Potential surgically related risks and the degree of desired weight loss are factors in the choice of bariatric surgical approaches, since these may differ between bariatric procedures.⁴² Banding procedures have been more common outside the United States, but recent utilization data suggest this procedure is becoming relatively more common among obese adults and adolescents undergoing bariatric surgeries in the United States.⁴⁴ Both adjustable gastric banding and gastric bypass are currently considered for severely obese adolescents with serious obesity-related comorbid conditions who have failed medical treatment, but only when performed by highly trained and skilled bariatric surgeons in a program with close nutritional, psychological, and surgical evaluation and followup.⁴⁵

Quantitative Synthesis

For the behavioral interventions, we conducted meta-analyses of short-term (KQ1) and maintenance (KQ2) outcomes within each setting. Twelve⁶⁹–⁸⁰ of the sixteen trials reporting short-term weight outcomes were included in the meta-analysis for KQ1. Five were in school settings,⁶⁹–^{72, 75} and there were two each in specialty health care^{73, 74} and primary care^{77, 78} settings. The final trial in this analysis was the only included trial conducted on-line.⁷⁶ It was, therefore, not statistically combined with other trials, although it appears on the visual display of the meta-analysis for qualitative comparison purposes.

Four^{72, 73, 78, 81} of the five trials reporting maintenance outcomes were included in the meta-analysis of KQ2, grouped by setting. Two of these trials were conducted in specialty health care settings,^{73, 81} and one trial each was conducted in school⁷² and primary care⁷⁸ settings; all of these are presented on the summary display but not all were statistically combined with other trials. Three ^{72, 73, 78} of the trials reported both short-term and maintenance outcomes and are included in both meta-analyses.

If mean change scores from baseline for each group were not reported, we calculated an unadjusted difference between the mean baseline and mean followup scores for each group using simple subtraction. Standard deviations (SDs) of the change scores were reported in five trials with post-treatment outcomes and one trial with followup outcome. In addition, three authors who did not report them in published articles provided us with these unpublished data.^{69, 76, 80} We calculated standard deviations for trials that did not report them. Baseline BMI is highly correlated with post-treatment and follow-up BMI, and we had to take this correlation into account when calculating the standard deviations of the change scores. In order to estimate the degree of correlation, we examined data from a trial⁷⁰ that reported both the SDs of the change scores (which we were attempting to calculate) and the SDs of the baseline and post-treatment BMIs (which we would use to calculate of the SDs of the change scores). From this trial, we ascertained that the correlation between the baseline and post-treatment score was approximately 0.90. Therefore, we assumed a correlation of 0.90 for the remaining trials and calculated SDs of BMI change using the following formula:

SD_{baseline-followup} = sqrt(SD² _baseline + SD² _followup - 2*0.90*SD_baseline*SD_followup).

When given standard errors rather than standard deviations, we calculated standard deviations by multiplying the standard error by the square root of n. When given symmetric confidence limits rather than standard deviations, we determined the standard deviation using the following formula:

Std Dev = (CI width)(√n)/2*(1.96)

We used random effects models because the trials varied considerably along many dimensions that would impact both baseline BMI (e.g., age, minimum overweight inclusion criteria) and change in BMI (e.g., intensity of intervention, comprehensiveness of treatment program). All meta-analyses were conducted using the “metan” procedure of Stata 9.2 with the “random” option, and then confirmed the results using RevMan 4.2. Forest plots are taken from our RevMan output.

Trial Characteristics

Table 3

Short-term outcomes of behavioral interventions

Table 3

Short-term outcomes of behavioral interventions

Study reference setting	N randomized Age Baseline BMI	Intervention hours (I-C)/intensity Intervention Components	Short-term BMI change: Mean change (SD of change)
Graf et al 200670,86 School	N: 276 Age 6–11 BMI: I (participants): 22.8 ± 3.6 I (non-participants): 21.1 ± 2.4 C: 21.7 ± 2.7	175.5 hrs/High I: PA, BehMod, Fam C: Usual school curriculum	9-mo (post-tx) I (participants): +0.3 ± 1.3 I (non-participants): +0.5 ± 1.3 C: +0.7 ± 1.2
Carrel et al 200569 School	N: 53 Age 12–13 BMI: I: 32 ± 6 C: 30 ± 4	67.5 hrs/Medium I: PA C: Typical PE class	9-mo (post-tx) I: -0.55 ± 1.1‡ C: +0.51± 1.7‡
Kalavainen et al 200772 School	N: 70 Age 6–9 BMI: I: 23.4 ± 2.6 C: 22.9 ± 2.5	45 hrs/Medium I: PA, BehMod, Fam, C: Handouts, 2 counseling meetings	6-mo (post-tx)** I: -0.8 ± 1.0 C: 0.0 ± 1.1
Johnston et al 2007a75 School	N: 71 Age 10–14 BMI: I: 27.7 ± 5.0 C: 25.6 ± 3.4	41.5 hrs/Medium I: PA, BehMod, C: Self-help materials	6-mo (post-tx)** I: -0.16 ± 1.05 C: +0.64 ± 0.90
Johnston et al 2007b71 School	N: 60 Age 10–14 BMI: I: 25.4 ± 4.7 C: 26.7 ± 5.5	41.5 hrs/Medium I: PA, BehMod C: Self-help materials	6-mo (post-tx)** I: -0.99 ± 3.79 C: +1.08 ± 1.00
Savoye et al 200774 Health Care	N: 174 Age 8–16 BMI: I: 35.8 ± 7.6 C: 36.2 ± 6.2	97.5 hrs/High I: PA, BehMod, Fam, MHTx C: Brief semi-annual counseling	12-mo (post-tx)** I: -1.7 (3.1) C: +1.6 (3.1)
Reinehr et al 200673 Health Care	N: 240 Age 6–14 BMI: I: 27.0 (26.4, 27.6) C: 26.1 (25.2, 27.8)	76 hrs/High I: PA, BehMod, Fam, MHTx C: No treatment due to distance from clinic	12-mo (post-tx)** I: +0.1 (SD NR) C: +2.0 (SD NR)
Golley 200783 Health Care	N=111 Age 6–9 BMI: 24.3 ± 2.6 (overall)	10.3 hrs (I1), 22 hrs (I2)/Low I1: Fam, MHTx I2: PA, BehMod, Fam, MHTx C: Wait List	12-mo (7-mos post-tx): NR† BMI SDS: I1: 2.56 ± 0.79 I2: 2.43 ± 0.68 C: 2.60 ± 0.57
Gillis 200787 Health Care	N: 27 Age 7–16 BMI: I: 1.98 ± 0.21 C: 2.16 ± 0.34	8 hrs/Very low I: Case manager C: 1 counseling session	6-mo (post-tx) NR† BMI SDS: I: -0.045 ± 0.19 C: +0.075 ± 0.08
Nemet et al 200579 Child Health and Sports Center	N=54 Average age 11.1 BMI: I: 28.5 ± 4.1 C: 27.8 ± 5.0	35.75 hrs/Medium I: PA, BehMod, Fam C (n=24): Nutritional counseling	12-mo (9-mo post-tx)* I: -2.4 (SD NR) C: +0.8 (SD NR)
McCallum et al, 200778,84 Primary Care	N=163 Age 5–9 BMI: I: 20.5 ± 2.2 C: 20.0 ± 1.8	4 hrs/Very low I: BehMod, Fam C: Usual primary care treatment	9-mo (6-mo post-tx): I: +0.5 (SD NR) C: +0.8 (SD NR)
Saelens et al 200277 Primary Care	N=44 Age 12–16 BMI: I: 31.0 ± 3.5 C: 30.7 ± 3.1	3.8 hrs/Very low I: BehMod C: Usual primary care treatment	7-mo (3-mo post-tx)*: I: +0.1 (NR) C: +1.4 (NR)
Doyle et al, unpub; Celio et al 200676 E-mail, Internet	N=83 Age 12–18 BMI: I: 34.6 ± 7.8 C: 33.9 ± 6.9	16 hrs/Low I: BehMod C: Information only	8-mo (4-mo post-tx): I: -0.2 (SD NR) C: +0.4 (SD NR)
Braet et al 200385 Residential	N: 76 Age 10–17 BMI: I: 33 (SD NR), C: 33 (SD NR)	3,520 hrs/Very high I: PA, BehMod, MHTx C: Wait List	10-mo (post-tx)**: NR† %OW: I: -51 (SD NR) C: +6 (SD NR)
Rooney 200580 Community	N=98 families Age 5–12 BMI: NR	3 hrs (I1), 21 hrs (I2)/Very low; Low I1: Fam I2: Fam C: Not described	9-mo (6 mo post-tx): NR‡ I1&I2: -0.87 ± 1.27 C: -0.43 ± 1.07
Senediak et al 198588 Setting NR	N=35 Age 6–12 BMI: I: 20.5 ± 2.2 C: 20.0 ± 1.8	12 hrs (I1, I2, C)/Low I1: BehMod, Fam I2: BehMod, Fam C: Social support, relaxation, mood monitoring	6-mo (3–5 mo post-tx)*: NR† (%OW I1: -13.04 (SD NR) I2: -19.22 (SD NR) C: -5.86 (SD NR))

Note: Interventions ordered first by setting and second by intensity.

Abbreviations: I- Intervention group; C- Control group; NR-Not Reported; PA-organized physical activity sessions; BehMod-behavioral modification principles; Fam-parent participant; MHTx-mental health treatment beyond behavioral modification; post-tx-post treatment; SD-standard deviation

*

p<0.05;

**

p<0.01, bold if p<0.05

†

BMI not reported, so other outcome listed

‡

Unpublished data supplied by author

Table 4

Maintenance outcomes of behavioral interventions

Table 4

Maintenance outcomes of behavioral interventions

Study reference	N randomized Age range Baseline BMI Setting	Intervention hours (in excess of control group hours of contact) /Intensity; Intervention components	Maintenance BMI Change: Mean change (SD of change)
Kalavainen et al 200772 School	N=70 Age 6–9 BMI: I: 23.4 ± 2.6 C: 22.9 ± 2.5	45 hrs/Medium I: PA, BehMod, Fam C: Handouts, 2 counseling mtgs	18-mo (12-mos post-tx)* I: 0.1 ± 1.2 C: 0.8 ± 1.3
Reinehr et al 200673 Health Care	N=240 Age 6–14 BMI: I: 27.0 (26.4, 27.6) C: 26.1 (25.2, 27.8)	76 hrs/High I: PA, BehMod, Fam, MHTx C: No treatment d/t distance	24-mo (12-mos post-tx)** I: +1.2 (SD NR) C: +2.9 (SD NR)
Mellin et al 198782 Health Care	Age 12–18 (15.6) N=66 BMI: NR	24 hrs/Low I: PA, BehMod C: No treatment	15-mo (12-mo post-tx)†: NR‡ Percent Overweight: I: -9.9 ± 15.0 (p<0.01)** C: -0.1 ± 13.2 (n.s.)**
Flodmark 199381 Health Care	N=93 Age 10–11 BMI: I1: 25.5 ± 0.53* I2: 24.7 ± 0.36* C: 25.1 ± 0.35*	12 hrs (I1), 24 hrs (I2)/Low I1: Fam I2: Fam, MHTx C: Matched controls, no treatment	~48-mo (30–34 mo post-tx)*: I1: +1.6 (SD NR) I2: +1.1 (SD NR) C: +2.8 (SD NR)
McCallum et al, 200778,84 Primary Care	N=163 Age 5–9 BMI: I: 20.5 ± 2.2 C: 20.0 ± 1.8	4 hrs/Very low I: BehMod, Fam C: Usual primary care treatment	15-mo (12-mo post-tx): I: +1.2 (NR) C: +1.2 (NR)

Note: Interventions ordered first by setting and second by intensity.

Abbreviations: I- Intervention group; C- Control group; NR- Not Reported; PA- Intervention included organized physical activity sessions; BehMod- Intervention included instruction in behavioral modification principles; Fam- Parent was a primary participant in the intervention; MHTx- Mental Health treatment beyond behavior modification related to diet and exercise was provided; n.s.-not statistically significant; post-tx-post treatment; SD-standard deviation.

*

p<0.05 or believed likely to be p<0.05;

**

p<0.01; bold if p<0.05 or likely to be p<0.05.

†

no direct comparisons reported, but differences between paired t-tests suggests p<0.05.

‡

BMI not reported, so other outcome listed.

Two trials were conducted in primary care settings,^{77, 78} five in specialty health care settings,^{73, 74, 79, 83, 87} five in school settings,⁶⁹–^{72, 75} one in a residential setting,⁸⁵ one in a child health/sports center,⁷⁹ one using the internet,⁸⁹ and three in settings that were not described.^{80, 81, 88} Eight studies were conducted in the United States, three in Australia, three in Germany, two in Israel, and one each in Belgium, Finland, and Sweden. A total of 22 different active treatment arms were evaluated. Duration of treatment ranged from 3 to 12 months, with the exception of one study with a “rapid pace” treatment arm lasting only 4 weeks,⁸⁸ and a longer trial that lasted for 14 to 18 months.⁸¹ Treatment intensity (estimated in hours of contact) ranged from 3.8 to 3,520 hours, with 16.7 percent (n=3) providing less than 10 hours, 33.3 percent (n=6) providing 10–25 hours, and 33.3 percent (n=6) providing 26–75 hours. The remaining three trials provided considerably more than 75 hours (97.5, 175.5, and 3,520 hours).

Ten of the trials involved the parents as primary participants in the intervention.^{70, 72}–^{74, 78}–^{81, 83, 88} All but one⁷⁴ of these trials involved children aged 11 years and younger on average. Parental involvement took many forms in these trials, including weight control educational sessions (with or without their overweight child),^{70, 72, 74, 78}–^{80, 84, 86, 88} family therapy,^{73, 81} or parenting skills training.⁸³ Family involvement in the remaining eight trials ranged from no involvement to including parents in one to three counseling sessions. The trials with less parent involvement primarily targeted older children, although three included those as young as 10 years,^{71, 75, 85} and one included children as young as 7 years.⁸⁷

Participants engaged in organized physical activity sessions as part of the intervention in eleven of the trials.⁶⁹–^{75, 79, 82, 83, 85} Four additional trials^{77, 78, 88, 89} applied behavioral modification principles to help participants increase their physical activity on their own time. Three trials provided only information and encouragement for physical activity, but did not apply behavioral modification principles such as problem-solving and goal-setting to physical activity.^{80, 81, 87}

Additional trials that did not meet inclusion criteria for weight outcomes, but did for other key questions, are detailed in the sections addressing those key questions.

Study design and quality. We rated eight^{71, 72, 74, 75, 77, 78, 83, 89} of the 18 trials as good-quality. The remaining trials were rated as fair-quality. Most trials (n=14) were randomized controlled trials but three were nonrandomized controlled trials.^{73, 81, 85} It was unclear whether one fair-quality trial involved randomization.⁸² Eleven of the 16 trials using randomization failed to report whether treatment allocation was blinded. Fifteen of the 18 trials did not report whether those conducting followup assessments were blind to the treatment condition. Many of the trials were also quite small, with 12 of 18 trials including 40 or fewer participants per treatment arm. While most trials reported retention of around ninety percent or higher, but it was below 70 percent in three trials.^{74, 87, 88} One trial ⁷⁴ among these used statistical methods to compensate for attrition. Several trials tested for differential attrition statistically (none found differential attrition between treatment and control groups), but most did not. While two smaller trials^{77, 90} appeared to have differential attrition, these differences were not tested statistically. The majority of trials (13/18, 72.2 percent) were published in 2005 or later.

Short-term (6–12 month) Weight Outcomes with Behavioral Interventions (KQ1)

Sixteen trials⁶⁹–^{75, 77}–^{80, 83, 85, 87}–⁸⁹ measured short-term weight outcomes (6 to 12 months after entry into treatment). Two^{73, 83} of these trials reported actual BMI measures between groups, but tested only whether BMI trends from baseline to followup were significantly different. Most trials reported weight outcomes as post-intervention BMI or changes in BMI from baseline and compared these changes between intervention and control groups. Among trials that did not report BMI or change in BMI, two trials, reported weight outcomes as changes in BMI standard deviation scores (SDS),^{83, 87} two trials reported changes in percent overweight,^{85, 88} and one trial reported change in BMI percentile.⁸⁰ All studies involved children and/or adolescents whose BMI exceeded the 97^th percentile on average.

All trials except one⁸⁰ were consistent with a beneficial effect of treatment on BMI change compared with controls. Not all of these differences, however, were statistically significant. Programs conducted in the outpatient setting or the community generally showed only modest differences in BMI change between treatments and controls. In most cases participants remained at or above the 95^th percentile after completing the interventions. The greatest level of weight loss was seen in 76 youth aged 10 to 17 years participating in a very high-intensity (3250 hours), 10-month residential program. Average weight decreased from 75 percent overweight to 24 percent overweight in the intervention group, compared to a 6 percent increase in those on a waiting list.⁸⁵

Figure 4. Pooled analysis: Short-term effect size of (more...)

   Figure 4. Pooled analysis: Short-term effect size of behavioral interventions (KQ1)

Figure 4

Figure 4

Figure 4

Figure 4

Figure 4

Figure 4

Figure 4

Best case example from a healthcare setting. One good-quality trial conducted by Savoye and colleagues⁷⁴ illustrates a realistic best-case scenario, reporting the largest effect size of the outpatient programs included in this review and a comprehensive program in which many families with overweight children could realistically participate, if it were available to them and affordable. This year-long program (Bright Bodies Weight Management) at a pediatric obesity clinic in the United States accepted children ranging from age 8 to 16 years, with an average age of 12.1 years. Sixty-one percent of the 174 participants were girls. The Bright Bodies program involved approximately 98 hours of contact and an extensive educational program providing information on nutrition, physical activity, behavior change strategies, coping skills, and relapse prevention. They provided organized exercise sessions twice per week during the first 6 months, then once every two weeks during the next 6 months. Parents or caregivers attended all educational sessions. Children and adolescents in the intervention group began the program with an average BMI of 35.8 kg/m², which dropped by an average of 1.7 kg/m² by the end of the intervention, compared with an average increase of 1.6 kg/m² in the control group. This trial suffered from somewhat low retention (77.6 percent at 6 months and 66.7 percent at 12 months), but took statistical measures to examine and combat the effects of attrition, including comparing results in completers only with results involving multiple imputation and Last Observation Carried Forward (LOCF) data replacement methods.

To provide a more concrete example of the average impact of the Bright Bodies program, we modeled the impact on a 12-year-old girl who began the program at an assumed height of 5′0", with the average entry BMI of 35.8, and who experienced the average reduction in her BMI by 1.7 kg/m² over the course of the intervention year, while growing 2 inches (an average for this age and sex). This would amount to a change from 183 pounds to 186 pounds one year after she participated in the program, compared with an expected 21 pound weight gain and an increase of 1.6 BMI kg/m² if she had not participated.

Maintenance of Weight Changes After Behavioral Interventions (KQ2)

Three of the five interventions in trials reporting maintenance outcomes tested low-,⁸¹ medium-,⁷² or high-intensity⁷³ interventions delivered over 6 months or longer. The remaining two trials tested interventions lasting 3 months that were very low⁷⁸ or low intensity.⁸² Some of these trials also report 6-to 12-month outcomes that were described under KQ1.

Figure 5. Pooled analysis: Maintenance effect size (more...)

   Figure 5. Pooled analysis: Maintenance effect size of behavioral interventions (KQ2)

Figure 5

Adverse Effects of Behavioral Interventions (KQ3)

Table 5

Potentially harmful effects of behavioral interventions (more...)

Table 5

Potentially harmful effects of behavioral interventions for childhood overweight

Study reference	Patient characteristics	Description of intervention groups	Outcomes of all potential harmful effects examined
Height
Braet et al 200385	10–17 (13) N=76 Residential Belgium	I (n=38): 10-month residential weight loss treatment, minimal parental involvement, included organized PA C (n=38): Waiting list	No group differences in change in height at 10 months
Savoye et al 200774	Age 8–16 (12.1) N=174 Health Care USA	I (n=105): Bright Bodies Weight Management: nutrition education and behavior modification class, substantial parental involvement, included organized PA C (n=69): Brief semi-annual counseling	No group difference in changes in height at 6 months or 12 months
Golley 200783	Age 6–9 (8.2) N=111 Health Care Australia	I1 (n=37): Parenting skills training with emphasis on dietary and PA issues, no organized PA I2 (n=38): Above + Intensive weight loss education (lifestyle approach), included organized PA C (n=36): Wait-list Control, 3–4 brief phone calls to encourage retention in study	No group difference in changes in height at 12 months
Eating Pathology and Body Image
Carrel et al 200569	Age 12–13 (12.5) N=53 School USA	I (n=27): PE class emphasizing non-competitive movement activities, small class size; minimal parental involvement, included organized PA C (n=26): PE class, typical competitive, team sports emphasis	Among treatment participants, measures of “drive for thinness” and “external eating” declined, self-reported ratings of physical appearance, athletic competence, and social acceptance improved.
Saelens et al 200277	Age 12–16 Mean 14.2 ± 1.2 N=44 Primary Care USA	I (n=23): Healthy habits intervention: Primary care-based tailored weight loss intervention, minimal parental involvement, no organized PA C (n=21): Usual primary care treatment	Problematic eating/eating disorder psychopathology did not differ between groups
Doyle et al, unpub; Celio et al 200676	Age 12–18 (14.5) N=83 Internet USA	I (n=42): Internet-delivered, interactive, moderated cognitive-behavioral program, limited parental involvement, no organized PA C (n=41): Basic, non-interactive information on nutrition and physical activity	Control group showed greater decline in Shape Concern than Intervention group; no other differences in eating disorder pathology
McCallum et al, 200778,84	Age 5–9 (7.4) N=163 Primary Care Australia	I (n=82): Primary care-based wt loss intervention, including brief solution-focused intervention, parent as agent of change; no organized PA C (n=81): Usual care	No differences on child-reported ratings of body satisfaction or appearance/self-worth
Other
Mellin et al 198782	Age 12–18 (15.6) N=66 Health Care USA	I (n=37): SHAPEDOWN program, comprehensive weight loss treatment, minimal parental involvement, included organized PA C (n=29): no treatment controls	Depression improved in treatment group, did not change in control group.
Supplementary Trials, Injuries Related to Physical Activity
Sung et al 200294	Age 8–11 N=82 NR China	I (n=41): 6-week diet program + supervised physical training C(n=41) 6-week diet program, no organized PA	No training-related injuries. (Baseline BMI 25.5)
Davis et al 200693	Age 7–11 N=100 Health Care USA	I1: 13 weeks low-dose aerobic exercise (20 min/day) I2: 13 weeks High-dose aerobic exercise (42 min/day) C: No exercise control	1 bone fracture in exercising group (11 and I2 combined) (Baseline BMI 26.5)

Abbreviations: PA-physical activity; PE-physical education; tx-treatment

We found no evidence that behavioral intervention programs may be harmful. Among the eight trials, three ^{74, 83, 85} reported no group differences in change in height measured at 10 to 12 months. Four trials ^{69, 77, 78, 89} reported either favorable or no effects on several measures of eating disorder pathology or body image/physical self-concept. One trial⁸² reported that depression symptomatology improved in intervention group participants, but did not change in the control group, which represents an added benefit rather than an adverse effect. In addition, Nemet and colleagues⁷⁹ reported that no adverse events were noted, but did not describe what events they examined or how they elicited information on adverse events. In the two trials examining injuries in exercise programs, Sung and colleagues⁹⁴ reported that none of the 41 obese children in their exercise condition were injured, and only one of the 73 obese children in the trial by Davis and colleagues⁹³ fractured a bone. No children in the control groups of either of these trials reported any injuries.

Other Beneficial Outcomes of Behavioral Interventions (KQ4)

Table 6

Other positive medical outcomes reported in behavioral (more...)

Table 6

Other positive medical outcomes reported in behavioral intervention trials

Study reference	Increase in high-density lipids (HDL)	Decrease in Low-density lipids (LDL)†	Decrease in triglycerides	Decrease in systolic BP	Decrease in diastolic BP	Decrease in fasting glucose	Decrease in fasting insulin	Decrease in HOMA-IR	Adiposity (Measure)
Carrel et al 200569	—	—	—	—	—	N	IG	—	IG: DEXA; decrease in % of body fat
Flodmark 199381	—	—	—	—	—	—	—	—	IG: Triceps, subscapular, suprailiac skinfold; decrease in thickness
Gillis 200787	N	N	N	—	—	—	—	—	—
Golley 200783	N	—	N	N	N	N	N	—	IG: decrease in Waist circumference
Graf et al 200670,86	—	—	—	—	—	—	—	—	N: decrease in Waist circumference
Johnston et al 2007a75	N	N	N	N	N	N	N	—	IG: Bioelectric impedance; decrease in body fat percentage
Johnston et al 2007b71	N	IG	N	N	N	N	N	—	N: Bioelectric impedance; decrease in body fat percentage
Nemet et al 2005 79	—	—	—	—	—	—	—	—	IG: Triceps, subscapular skinfold; decrease in thickness
Reinehr et al 200673	N	IG	N	IG	N	N	IG	IG	—
Savoye et al 200774	N	N	N	—	—	N	IG	IG	IG: Bioelectric impedance; decrease in body fat percentage
Senediak et al 198588	—	—	—	—	—	—	—	—	IG: Subscapular skinfold; decrease in thickness

N - No group differences, IG - Result favors intervention group

†

HDL and LDL differences are reported separately; trials do not report on the ratio of HDL to LDL

BP- blood pressure; DEXA - Dual-energy x-ray absorptiometry; HOMA - homeostasis model assessment of insulin resistance

Measures of adiposity. Nine of these eighteen trials ⁶⁹–^{71, 74, 75, 79, 81, 83, 88} reported measures of adiposity. In most cases these trials found that the intervention groups showed greater improvement in these measures than those in the control groups. Six trials^{69, 74, 75, 79, 81, 88} found positive effects in both the primary weight outcome and either skinfold measures or body fat, as measured by bio-electrical impedance. One more trial ⁸³ that did not have positive primary weight outcomes did show improvement in adiposity (as measured by DEXA)⁶⁹ and waist circumference. ^{70, 83} The remaining two trials did not see group differences in adiposity as measured by bio-electrical impedance⁷¹ or waist circumference.⁷⁰

Health outcomes. Other outcomes explored included lipid levels, glucose tolerance, blood pressure, and physical fitness. Results for all of these outcomes were quite mixed. Reported differences were most commonly reductions in LDL cholesterol levels, reduced fasting insulin, and reduced insulin resistance. Three^{69, 73, 74} of the six^{69, 71, 73}–^{75, 83} trials reporting on fasting insulin found reductions of fasting insulin in the intervention groups relative to the control group. Two of these trials^{73, 74} also reported significant reductions in insulin resistance, as measured by the homeostasis model assessment of insulin resistance (HOMA). By contrast, none of the six trials ^{70, 71, 73}–^{75, 83} reporting lipid levels found group differences in HDL or triglyceride levels, and only two found reductions in LDL levels.^{71, 73}

None of the four trials^{71, 73, 75, 83} reporting on blood pressure found group differences on diastolic blood pressure and only one⁷³ reported reductions in systolic blood pressure. Similarly, none of the five trials ^{71, 73}–^{75, 83} reporting on glucose levels found any group differences.

Behavior changes. The interventions in these trials appeared to have a minimal impact on the intermediate outcomes of diet and activity level. While four trials⁷⁷–^{79, 87} explored dietary changes, only one⁷⁸ found group differences. The only dietary differences found in this study were that children in the intervention group reported consuming less whole milk, while consuming more skim milk and water. Five trials⁷⁷–^{80, 87} reported on changes in physical activity levels and/or sedentary behavior. Only one reported positive effects.⁷⁹ This trial provided organized physical activity sessions during the 3-month intervention, and measured the amount of sedentary and physical activity participants reported 1 year later. Participants in the intervention group reported an average of 6 fewer minutes of screen time per day and 9.1 more weighted metabolic-equivalent units of habitual activity. This suggests that long-term changes in physical activity can be sustained even after only 3 months of intervention. The remaining four trials, which showed no group differences, included one trial targeting physical activity,⁸⁰ two low-intensity primary care-based trials,^{77, 78} and a small (n=27), low-intensity trial involving weekly brief contact with a case manager.⁸⁷

Eating disorders. Finally, several trials measured constructs such as impacts on eating disorders or body image that may be a potential harm or benefit of a treatment program. No group differences were found in either of the two trials ^{77, 89} reporting on eating disorder pathology. Instead, Doyle and colleagues⁸⁹ reported reduced levels of shape concern in the intervention participants in their trial. Mellin and colleagues ⁸² found reductions in depression scores among intervention participants and no changes in depression scores in control participants. They did not, however, directly test the groups against each other. Also, Mellin and colleagues did not find group differences in change in self-esteem and both groups showed improvement in repeated measures tests.

Important Components of Behavioral Interventions (KQ5a)

We discuss our findings from this exercise, but these should be considered primarily as hypothesis-generating. The degree of variability among this small number of treatment programs, including important differences in effects due to setting, age and treatment intensity, greatly limits our ability to examine other treatment components.

Organized physical activity sessions. Programs that provided organized physical activity sessions (rather than encouraging participants to exercise at home) appeared to be more likely to improve BMI. Group differences were seen in eight of 11 programs with organized physical activity sessions. The three trials that did not see beneficial changes in BMI reported improvements in other weight or adiposity measures. We did not have sufficient data to ^†determine whether programs with organized physical activity or those that improved physical activity or fitness were more likely to have a positive impact on other health outcomes (such as fasting insulin or blood pressure). The physical activity sessions ranged from seven 1-hour sessions at 2- to 4-week intervals, which consisted of fun, noncompetitive physically active games and activities,⁸³ to twice-weekly 50-to 60-minute sessions for 6 to 9 months.^{70, 74} Efforts were generally made to present a variety of enjoyable activities, including team sports, noncompetitive games, dancing, swimming, walking, jogging, and obstacle courses. Several trials^{72, 79, 83} employed activities to help develop motor skills and one⁶⁹ reported making efforts to personalize the skill level of the activities to the skill levels of the child. One trial ⁷⁴ used exercise physiologists to facilitate the exercise sessions and help children maintain a target heart rate of 65 to 80 percent of their age-adjusted maximal heart rate.

We identified three supplementary trials in four publications that unfortunately contributed little to the elucidation of the role of physical activity sessions.⁵³–^{55, 95}

Parental involvement. The role of parental involvement in weight management programs can only be considered in the context of the child's age. None of the seven trials that focused on adolescents included parents as primary participants of the intervention. However, three of the trials^{71, 75, 82} in adolescents did invite parents to one or more intervention sessions, and all three of those trials did show positive weight outcomes. Thus, parental participation may increase the likelihood of successful weight loss in adolescents.

All eight of the trials limited to children aged 12 or younger had high levels of parental involvement, as did two of the trials that included both younger children and adolescents. Due to the lack of variability we could not explore the importance of parental involvement further than concluding that weight-loss researchers consider parental involvement crucial for successful weight loss in young children. Parental involvement took many forms in the trials with high levels of involvement. In some trials parents and children attended weight control educational sessions together,^{72, 78, 79, 88} while others provided family therapy, ^{73, 74, 80, 81} or parenting skills training ⁸³ in addition to traditional weight control topics. In one trial the children participated only in fun, physical play sessions or family activities, while only parents received instruction in weight management.⁷⁰

Similarly, few conclusions could be drawn from the five supplementary comparative effectiveness trials⁵⁷–^{60, 96} attempting to isolate the importance of child vs. parental involvement. Data suggest that it may be helpful to have both parents and children involved in interventions with young children. Parent training in child management principles may also be helpful with parents of young children. These conclusions, however, are tentative because they are based on only a few trials, with limited generalizability to the population of the United States.

Five of the supplementary trials (in six publications) examined the impact of varying types of parental involvement in weight loss interventions, four in younger children^{56, 57, 59, 60, 96} and one in adolescents.⁵⁸ Among the trials in younger children, three ^{56, 59, 96} compared interventions involving parents or children only with those involving both children and parents, with conflicting results: two^{59, 96} suggested it was most helpful to have both child and parent involved, but this was not supported by the third.⁵⁶ The fourth supplementary trial ⁵⁷ in younger children found that children had greater weight loss when parents were taught child management techniques in addition to weight management principles. This contrasts with one of our primary trials⁸³ conducted by Golley and colleagues, which taught child management techniques to parents without enhancing weight loss. However, Golley and colleagues provided only about half of the treatment hours of the supplementary trial.

One trial⁵⁸ in black adolescent girls explored the role of parental involvement in families categorized as lower to lower-middle class, largely single-parent households. Researchers randomized families to one of three groups: adolescents attending treatment sessions without mothers, mothers and daughters attending sessions together, and mothers and daughters attending separate, concurrent groups. Groups did not differ on any measure of weight loss, nor did the groups differ from their own baseline measures. The authors reported low attendance among mothers in this program, which suggests that the burden of attending a treatment program in these primarily single-parent families is likely quite high.

Behavior management techniques. Among the primary 18 trials, programs that included participant training and support in the use of behavioral management techniques were more likely to be successful than those that did not. None of the four trials^{69, 80, 81, 87} that lacked instruction in behavior management techniques were successful in improving weight outcomes. Eleven of the 14 trials that taught participants to use behavioral management techniques did show group differences in BMI or other weight outcomes. These trials all appeared to provide broad advice on using these techniques for changing diet, activity, and other related behaviors.

We identified two supplementary trials in four publications⁶¹–⁶⁴ comparing standard weight loss management programs without cognitive behavioral treatment or techniques with the same programs, adding behavioral management techniques. One of these, conducted by Epstein and colleagues,⁶¹ provided 40 hours of contact to 24 5- to 8-year-old girls and their parents over 12 months. This trial included a 5-week intensive treatment phase and once monthly maintenance contacts thereafter. Behavioral management principles were provided to parents in one of the treatment groups, but not the other. The second trial ⁶⁴ compared a group of adolescents receiving nutrition counseling from a dietitian without behavioral management techniques with a group receiving the same nutrition counseling plus an intervention delivered over the internet. This intervention was based on the treatment methods developed by Epstein and colleagues, which included behavior management techniques. The addition of behavioral management training improved weight outcomes in both of these trials at the end of the treatment phase, although the effect was not seen in long-term followup in the trial that measured weight outcomes 21 months after the end of treatment.⁶⁴

Factors Influencing the Effectiveness of Behavioral Interventions (KQ5b)

Treatment effects varied by intervention setting and by intervention intensity. Residential treatment and high-intensity interventions in specialty health care treatment settings (both inpatient and outpatient) had the largest treatment effects; medium-intensity interventions in school settings had consistent, but modest effects; some low-intensity interventions in primary care or other settings have more modest effects; and, limited data from very low-intensity primary care or internet-based interventions suggest no treatment effects. We were unable to isolate other population or environmental factors that may influence the effectiveness of a treatment because of the limited number of trials and the great heterogeneity in intervention, population, and environmental factors.

Trial Characteristics

Table 8

Characteristics of randomized clinical trials of pharmacological (more...)

Table 8

Characteristics of randomized clinical trials of pharmacological anti-obesity treatments among adolescents, by drug type

Source	Intervention	No. of months of drug treatment	No. of behavioral intervention sessions	Characteristics	No of study sites	Country	% Attrition	Qualitya	Placebo Run-in Period	Funding Source
Sibutramine
Berkowitz et al, 200397	Sibutramine (5 mg/d for 1 wk, 10 mg/d for 4 wks, then 15 mg) + BI or placebo + BI	6	19	N randomized: 82 Age: 13–17 Female: 67%	1	USA	10%	Good	Yes	NIH, hospital, pharm
Berkowitz et al, 200698	Sibutramine (10 mg/day for 6 mos, then 10–15 mg/d) + BI or placebo + BI	12	10	N randomized: 498 Age: 12–16 Female: 66%	33	USA	28%b	Good	No	Pharm
Garcia-Morales et al, 2006100	Sibutramine (10 mg/d) + BI or placebo + BI	6	8	N randomized: 51 Age: 14–18 Female: 56%	1	Mexico	22%	Fair	Yes	Pharm
Godoy-Matos et al, 2005101	Sibutramine (10 mg/d) or placebo	6	1	N randomized: 60 Age: 14–17 Female: 82%	1	Brazil	17%b	Fair	Yes	Pharm
Van Mil et al, 2007103	Sibutramine (5 mg/d for 2 wks, then 10 mg/d) + BI or placebo + BI	3c	16	N randomized: 24 Age: 12–17 Female: 54%	1	Netherlands	17%b	Fair	No	NR
Orlistat
Chanoine et al, 200599	Orlistat (120 mg, TID) + BI or placebo + BI	12	18	N: 539 Age: 12–16 Female: 67%	32	USA & Canada	35%	Good	Yes	Pharm
Maahs et al, 2006102	Orlistat (120 mg, TID) + BI or placebo + BI	6	7	N: 40 Age: 14–18 Female: 67%	1	USA	15%	Fair	No	University supported

Abbreviations: BI - behavioral intervention (with or without a behavioral management program); TID - three times daily; NR - not reported; Pharm-pharmaceutical; NIH-National Institute of Health.

a: Quality criteria are described in Appendix B Table 1.

b: Attrition rate was different between the intervention and control groups.

c: Patients were treated with BT + sibutramine or placebo for 3 mos. and then BT alone for 3 mos.

Table 10

Randomized, placebo-controlled, clinical trials evaluating (more...)

Table 10

Randomized, placebo-controlled, clinical trials evaluating pharmacological agents among special populations of obese children and adolescents and reporting weight outcomes

Source	N randomized Study design Country	Population Length of study	Intervention Drug dose	Baseline BMI	BMI results
Srinivasan et al, 2006105	N = 28 Cross-over RCT Australia	Obese children and adolescents ages 9–18 years with clinical suspicion of insulin resistance (fasting insulin: glucose > 4.5 or acanthosis nigricans) 12 months	A: Metformin for 6 months, then placebo for 6 months B: Placebo for 6 months, then metformin for 6 months Metformin dose: gradually increased (over 3 wks) up to 2 g/day vs. placebo	Total sample: 35.2 ± 5.1 kg/m2 (not reported by study group)	ΔΔ BMI SDS* -0.12 p=0.005 ΔΔ BMI -1.26 kg/m2 p=0.002
Freemark et al., 2001 104	N =32 RCT USA	Obese adolescents ages 12 to 19 years with fasting insulin concentration > 15 μU/mL; and ≥ 1 first- or second-degree relative with type 2 DM 6 months	IG: Metformin CG: Placebo Metformin dose: 500 mg, twice per day	IG: 41.5 ± 0.9 CG: 38.7 ± 1.3 (p < 0.05)	Δ BMI SDS IG: -0.12 CG: 0.23 p< 0.02 Δ BMI IG: -0.5 kg/m2 CG: 0.9 kg/m2 p-value NR

Abbreviations: BMI - Body mass index; DM - Diabetes mellitus; IG - intervention group; CG - control group; RCT - randomized controlled trial

*

ΔΔ BMI = Δ BMI_IG - Δ BMI_CG

Participants in the sibutramine and orlistat trials all met some type of BMI-based criteria for obesity (either above the age- and sex-specific 95 to 97^th percentile or above a BMI of 30 kg/m²), and mean BMI was typically 35 to 38 kg/m² at baseline. Most trials excluded those at or above the midpoint for Class III (morbid) obesity (BMI exceeding 44 kg/m²) and those who had type I or type II diabetes mellitus. The sibutramine trials also generally excluded patients who had cardiovascular disease or hypertension. About two-thirds of participants in these trials were females. The majority of trials did not report race/ethnicity of participants. However, in the two largest multi-center RCTs, almost half of the sibutramine patients were racial/ethnic minorities,⁹⁸ as were one-quarter of orlistat patients.⁹⁹ The sibutramine trial included 21 percent Black, 16 percent Hispanic, and 7 percent other nonWhite patients. The orlistat trial included 17 percent Black and 7 percent participants of other race-ethnicity. A small (n=52) sibutramine trial conducted in Mexico could have applicability to adolescents of Mexican heritage living in the United States.¹⁰⁰

The minimal behavioral intervention provided to all participants consisted of advice to follow a calorie-restricted diet (e.g., 500 kcal/day deficit) and meet physical activity goals (e.g., at least 30 min of aerobic activity per day). All but one trial¹⁰¹ also included a behavior management program, ranging in intensity from seven to 19 sessions with a dietitian, psychologist, or psychiatrist. Family members attended behavioral management sessions in only two of the seven trials.^{97, 103} The length of drug therapy lasted for either 3, 6, or 12 months (in one, four, and two trials, respectively). In the single trial evaluating 3 months of drug therapy (sibutramine), we report the follow-up results at 6 months. No other trials reported followup results describing weight patterns after the pharmacologic treatment ended.

Of the six trials that reported the source of funding, all but one trial was funded by the pharmaceutical industry, either completely or partially. Two of these pharmaceutically sponsored trials were large (about 500 participants), multi-center RCTs (over 30 study sites) conducted in the United States and Canada. One evaluated sibutramine⁹⁸ and the other evaluated orlistat.⁹⁹ The remaining trials randomized much smaller samples (n = 24 to 82), were conducted at single sites, and reported outcomes after only six months of treatment.

Study design and quality. All included studies were double-blinded, placebo-controlled RCTs of fair- or good- quality (see Appendix H ^† for quality criteria). Most trials used appropriate randomization methods and took explicit measures to conceal allocation assignment. In all of the trials, intervention and control groups were similar at baseline for age, sex, and anthropometric characteristics. Descriptions of drug protocols were clear. Descriptions of behavioral interventions were generally adequate, but much less detailed than trials evaluating behavioral interventions. Adherence to medication protocols (measured by pill counts) was 80 percent or higher in the majority of the trials. Adherence was slightly lower (72 to 73 percent) in the large multi-center orlistat RCT. In contrast, most of the trials did not report how the behavioral intervention program was supervised, whether it was delivered as intended, or any data on adherence to diet, physical activity, or other behaviors. Most of the trials specified that outcomes were assessed by personnel blinded to treatment status.

Attrition rates ranged from 10 to 35 percent. Notably, both of the large, multi-center trials had fairly high attrition. Overall attrition was 35 percent in the large orlistat trial. In the large sibutramine trial, the attrition rate was 28 percent overall and was differential between groups (24 percent in the sibutramine group and 38 percent in the control group, p=0.001). All of the trials analyzed main weight outcomes among the intent-to-treat (ITT) or modified ITT population. The modified ITT population included any participant who had at least one post baseline efficacy measurement. Missing values were replaced using the LOCF method in most trials and/or a linear mixed-effects model for repeated measures over time. One trial¹⁰⁰ excluded 10 percent of patients, even in the modified ITT population analyses, because they dropped out before one month.

Short-term (6–12 month) Weight Outcomes With Sibutramine Treatment (KQ1)

Table 9

Results of randomized controlled trials of pharmacological (more...)

Table 9

Results of randomized controlled trials of pharmacological anti-obesity treatments among adolescents, by drug type

Source	N	Baseline BMI (kg/m2)	Treatment months	Change BMI (kg/m2) p value	Physiological Outcomes		Adverse Events
Sibutramine
Berkowitz et al, 200397	43 39	I: 37.5 ± 4.0 C: 38.0 ± 3.6	6	-3.2a -1.5a p=0.001b	WC: SD LDL: NS HDL: NS TG: NS FPG: NS	Insulin: NS HOMA: NS Heart Rate: SDe Systolic BP: SDe Diastolic BP: NS	Adverse Events: NS
Berkowitz et al, 200698	368 130	I: 36.1 ± 3.8 C: 35.9 ± 4.1	12	-2.9 -0.3 p < 0.001	WC: SD LDL: NS HDL: SD TG: SD FPG: NS	Insulin: SD HOMA: SD Heart Rate: SDe Systolic BP: SDe Diastolic BP: SDe	Adverse Events: NS SAE: NS d/c med: NS Growth: NS Maturation: NS
Garcia-Morales et al, 2006100	26 25	I: 35.1± 5.3 C: 36.6 ± 5.2	6	-3.4 (-2.5, -4.2) -1.8 (-0.9, -2.6) P< 0.005*	WC: NS LDL: NS HDL: NS TG: NS	FPG NS Heart Rate: SDe Systolic BP: NS Diastolic BP: SDe	Adverse Events: NS d/c med: NS Maturation: NS Growth: NS
Godoy-Matos et al, 2005101	30 30	I: 37.5 ± 3.8 (f) 37.6 ± 4.3 (m) C: 35.8 ± 4.2 (f) 37.4 ± 1.9 (m)	6	-3.6 ± 2.5 -0.9 ± 0.9 p < 0.001	WC: SD LDL: NS HDL: NS TG: NS FPG: NS	Insulin: NS Heart Rate: NS Systolic BP: NS Diastolic BP: NS	SAE: NS d/c med: NS Other: SD
Van Mil et al, 2007103	12 12	I: 30.1 ± 4.5 C: 33.3 ± 5.0	3 + 3 mos f/uc	-0.8d -1.4d NR	% Fat Mass: NS Heart Rate: NS	Systolic BP: NS Diastolic BP: NS	Adverse Event: NS d/c med: NS Other: SD
Orlistat
Chanoine et al, 200599	357 182	I: 35.7 ± 4.2 C: 35.4 ± 4.1	12	-0.55 +0.3 p < 0.001	WC: SD Other Adiposity: SD LDL: NS HDL: NS TG: NS	FPG: NS Insulin: NS Heart Rate: NS Systolic BP: NS Diastolic BP: SDf	Growth: NS Maturation: NS Other: SD
Maahs et al, 2006102	20 20	I: 39.2 ± 1.2 C: 41.7 ± 2.6	6	-1.3 ± 1.6 -0.8 ± 3.0 NS	% Fat Mass: NS LDL: NS HDL: NS	TG: NS FPG: NS Insulin: NS	Other: SD

a: Calculated based on average BMI at baseline and average percentage change in BMI for each group (I: -8.5% ± 6.8%, C: -4.0% ± 5.4%).

b: Based on comparison of percent change in BMI between groups

*

result of ANOVA testing interaction between treatment group and time

c: Patients were treated with BT + sibutramine or placebo for 3 mos and then BT alone for 3 mos.

d: calculated based on differences reported baseline to 3 mos and 3 mos to 6 mos.

e: Relative increased rate over time in sibutramine group compared to placebo group

f: Relative reduction in rate over time in orlistat group compared to placebo group

Abbreviations: IG - Intervention group; CG - Control group; BT - Behavioral Treatment, NS - not significant; NR - not reported; WC - Waist circumference; LDL - Low-density Lipoprotein; HDL- High-density Lipoprotein; TG - triglyceride; FPG - Fasting plasma glucose; BP - Blood pressure; SD - statistically significant difference; SAE - Serious adverse events; HOMA - Homeostasis model assessment of insulin sensitivity; d/c - discontinue.

The single large trial that reported weight outcomes after 12 months of sibutramine plus a behavioral intervention also found statistically significant results in favor of the sibutramine group.⁹⁸ The mean reduction in BMI in the sibutramine group was -2.9 kg/m² compared to -0.3 kg/m² in the control group (p <0.001). As noted, this trial had higher attrition in the placebo control group (38 percent) than the sibutramine group (24 percent, p = 0.001), reducing our confidence in these findings. BMI measures over time were also analyzed using a linear mixed-effects model to predict missing values. In these analyses, the mean change in BMI between treatment and control groups was statistically significantly different at all study visits from week 1 through month 12. The difference between the changes in BMI z-scores was also statistically significant. In this trial, the mean change in body weight (± SE) at month 12 was -6.5 ± 0.31 kg in the sibutramine group versus 1.9 ± 0.56 kg in the placebo group (difference -8.4 kg, or 18.5 pounds (CI: -9.7,-7.2 kg); p < 0.001 by linear mixed-effects model).

Maintenance of Weight Changes After Sibutramine Treatment (KQ2)

No trials reported on maintenance of weight loss after sibutramine was discontinued.

Adverse Effects of Sibutramine Treatment (KQ3)

None of the sibutramine trials reported statistically significant differences between groups in the overall rates of having any adverse event, any serious adverse event, or discontinuation due to adverse events. In the large, 12-month sibutramine trial, serious adverse events were reported by 2.7 percent of patients in the sibutramine group and less than 1 percent of the control group. Only one of these events (excessive nausea and vomiting) was thought to be related to sibutramine. Two trials examined growth and maturation, including the 12-month, multi-center trial. Neither trial found a significant difference between the groups. Abdominal complaints and constipation were also found to be statistically higher in the sibutramine group in the shorter-term trials.

Other Beneficial Outcomes With Sibutramine Treatment (KQ4)

Components of Effective Sibutramine Treatment (KQ5a)

Data were largely insufficient to explore the importance of specific treatment components. Based on the limited number of trials, shorter treatment (3 as compared with 6 or 12 months) may be related to reduced beneficial effects on BMI. There are other possible explanations for these between trial differences, however, such as lack of placebo run in or differences in population or setting.

Factors Influencing the Effectiveness of Sibutramine (KQ5b)

Data were insufficient to explore the importance of population or environmental factors.

Short-term (6–12 month) Weight Outcomes With Orlistat Treatment (KQ1)

Two trials reported the weight outcomes after 6 or 12 months of orlistat therapy plus a behavioral intervention and results were mixed. The large (n=539), multi-center trial evaluating 12 months of orlistat therapy found a statistically significant difference between the change in BMI, favoring the orlistat plus a behavioral intervention group (-0.55 kg/m² vs. 0.3 kg/m², p < 0.001).⁹⁹ The absolute mean body weight increased in both groups during the 12-month trial, but increased less in the orlistat group (0.53 kg vs. 3.14 kg, p <0.001). Attrition in this trial was quite high (33 to 34 percent), but analyses of primary weight outcomes included over 98 percent of randomized participants and replaced missing data using the LOCF method. Also, baseline characteristics were not different for completers or those who dropped out within each group. Nevertheless, the high level of attrition in the trial somewhat limits its validity. A smaller trial (n=40) that evaluated the effects of six months of orlistat plus a behavioral intervention found that the orlistat group had a larger BMI reduction than the control group (-1.3 kg/m2 vs. -0.8 kg/m2), but this difference was not statistically significant.¹⁰²

Maintenance of Weight Changes After Orlistat Treatment (KQ2)

No trials reported on maintenance of weight loss after orlistat was discontinued.

Adverse Effects of Orlistat (KQ3)

Rates of serious adverse effects and discontinuation of therapy due to adverse effects were low in both trials and were not reported to be statistically different between groups. In the Chanoine and colleagues trial,⁹⁹ one or more serious adverse effects occurred in 3 percent of both groups. Discontinuation of therapy due to a serious adverse event occurred among 12 of 357 (3 percent) of orlistat patients and 3 of 182 (2 percent) patients in the placebo group. In the orlistat group, only one event was thought to be study-related: asymptomatic cholelithiasis in a 15-year-old female who had lost 15.8 kg by the time of the event. In the Maahs and colleagues trial,¹⁰² 2 of 20 patients in the orlistat group and 0 of 20 patients in the placebo group withdrew from the trial due to adverse effects. One suicide death occurred in the orlistat group to a patient who was under a psychiatrist's care. No deaths occurred in the placebo group.

Gastrointestinal (GI) side effects were very common among patients taking orlistat. Chanoine and colleagues reported that among patients taking orlistat: 50 percent reported fatty or oily stools; 20 to 30 percent reported oily spotting, oily evacuation, abdominal pain, fecal urgency, or flatus with discharge; 10 to 15 percent experienced soft stool, nausea, and increased defecation. Notably, 9 percent of orlistat patients reported fecal incontinence, compared with less than 1 percent of placebo patients. Chanoine and colleagues also reported that the GI side effects were mostly mild- to moderate-intensity and led to discontinuation of treatment among only two percent of orlistat patients. In the smaller 6-month orlistat trial, Maahs and colleagues also reported that numerous adverse gastrointestinal effects occurred significantly more frequently in the orlistat group than the placebo group, including: soft stools, oily spotting, fatty or oily stools, oily evacuation, liquid stools, cramping, flatus with discharge, and fecal incontinence. Soft stools, oily spotting, fatty or oily stools, oily evacuation, and liquid stools all occurred in over 50 percent of patients treated with orlistat. Flatus with discharge occurred in 20 to 47 percent of patients treated with orlistat (varying by study month), in contrast to 0 percent in all but the first month for the control group. Fecal incontinence occurred in 6 to 13 percent of the orlistat group at each month, in contrast to 0 percent of the control group during any month. The authors report that the oily spotting, fatty or oily stools, and cramping improved more over time in the orlistat group than in the placebo group.

Both orlistat trials measured vitamin A, D, and E levels and reported that levels were not different between groups. In the Maahs trial, quality of life measured using four different scales showed no statistically significant differences between groups over time. Possible lack of blinding in the outcome assessors, however, could have influenced these results. No between-group differences in growth, bone mineral density, and sexual maturation were reported.⁹⁹

Other Beneficial Outcomes of Orlistat Treatment (KQ4)

Chanoine and colleagues reported that both waist circumference and hip circumference decreased significantly more in those receiving orlistat and a behavioral intervention, compared with placebo plus behavioral intervention controls, at 12 months (p=0.01 for both in least squares mean (LSM) analysis). The LSM reduction for waist and hip were -2.67 and -1.52 cm, respectively, for the orlistat group, compared with -0.89 and -0.10 cm in the control group. In a subset of patients evaluated with dual-energy x-ray absorptiometry (DEXA), patients in the orlistat group lost significantly more fat mass than patients in the placebo group (-2401 g vs. -380 g; p =0.03). In contrast, percent body fat at 6 months was measured using bioelectrical impedance analysis in the Maahs trial, and no statistically significant differences were found between groups. Levels of LDL, HDL, TG, FPG, and insulin were measured in both Orlistat trials, and no significant differences were found between groups in either trial. The Chanoine and colleagues trial, however, reported a small reduction in diastolic blood pressure in the orlistat group (-0.51 mm Hg), compared to an increase in the placebo patients (+1.30 mm Hg; p=0.04). Change in systolic BP was similar in both groups and not statistically different.

Components of Effective Orlistat Treatment (KQ5a)

Data were insufficient to explore the importance of specific treatment components.

Factors Influencing the Effectiveness of Orlistat (KQ5b)

Data were insufficient to explore the importance of population or environmental factors.

Metformin Treatment in Obese Patients at High-Risk for Type 2 Diabetes

Short-Term (6–12 month) Weight Outcomes with Metformin Treatment (KQ1)

Both trials found statistically significant differences between groups for BMI or BMI SDS at 6 months, with results favoring the metformin group. Results should be interpreted with caution, however, because analyses in these trials included only patients who completed the trial (attrition rates were 9 and 21 percent), which could have caused bias.

Maintenance of Weight Changes with Metformin (KQ2)

No data were reported on maintenance of weight loss after metformin was discontinued.

Adverse Effects of Metformin (KQ3)

Trials were limited in their ability to detect adverse effects due to small sample size and limited duration. Neither trial reported any serious adverse events. One trial specifically reported that no episodes of vomiting or lactic acidosis occurred. Serum lactate, liver, and renal function parameters were reported as remaining normal or not different between groups in both trials. In both trials, some patients were reported to have nausea which, in three cases, required a 25 to 50 percent dose reduction in order to continue in the trial.

Other Beneficial Outcomes of Metformin Treatment in High-risk Obese Adolescents (KQ4)

One of the trials¹⁰⁵ found statistically significant improvements favoring the metformin group for waist circumference and subcutaneous adipose tissue, but no difference for visceral abdominal adipose tissue. These parameters were not reported in the other trial. Both trials reported improvements in fasting glucose and insulin, either between groups or only within the metformin group. Neither trial found statistically significant differences between groups for insulin sensitivity when using minimal model analyses, glucose effectiveness, acute insulin response disposition index, or glucose disposal. No lipid parameters were found to be statistically different between groups in the only trial that measures them.¹⁰⁴

Laparoscopic Adjustable Gastric Banding (LAGB)

Table 12

Other outcomes for bariatric surgical procedures

Table 12

Other outcomes for bariatric surgical procedures

Study	Failure	Resolution of comorbidities	Adverse events
Laparoscopic Adjustable Gastric Band
Angrisani 2005114 Band brand NR	≤ 25% EWL at 5 yrs: 20% (5/25)	NR	Mortality: None Laparotomic conversion: 1.7% (1/58) Overall postoperative complications: 10.3% (6/58) Band slip: 1.7% (1/58)	Gastric pouch dilation: 3.4% (2/58) Intragastric migration: 5.2% (3/58) Band removal: 10.3% (6/58) Conversion to gastric bypass or BPD: 5.2% (3/58)
Nadler 2007111 Lap-Band®	NR	NR	Perforated appendicitis within 10 days of surgery: (1.9% 1/53) Band slip: 3.8% (2/53) Hiatal hernia: 3.8% (2/53) Wound infection: 1.9% (1/53)	Mild hair loss: 9.4% (5/53) Iron deficiency: 7.5% (4/53) Nephrolithiasis, cholelithiasis: 1.9% (1/53) Gastroesophageal reflux: 1.9% (1/53)
Dolan 2003115,116 Fielding130 Lap-Band®	NR	NR	Band slip: 5.9% (1/17) Leaking port: 5.9% (1/17)
Abu-Abeid 2003126 Lap-Band®	NR	Amenorrhea: 100% High triglycerides: 100% (2/2) Abnormal cholesterol: 0% (0/1)	Perioperative complications: 0% Late complications: 0%
Silberhumer 2006129 Widhalm 2004117 Lap-Band® and SAGB®	6% (3/50) had EWL < 25% after at least 1 yr of follow-up	Diabetes mellitus II: 80% (4/5) Hypertension: 50% (6/12) Dyslipidemia: 100% (4/4) Asthma: 100% (3/3) Cholecystolithiasis: 100% (3/3)	Perioperative complications: 0% Dislocated port: 2% (1/50) Band slip: None
Yitzhak 2006#127 SAGB®	NR	100% resolution of all co-morbidities. Hypertension: 3/3 Diabetes Mellitus: 2/2 Asthma: 3/3 Obstructive sleep apnea: 10/10	Mortality: 0% Major post-operative complications: 0% Band slip: 10% (6/60) Band removal: 3.3% (2/60)
Gastric Bypass/Other procedures
Lawson 2006112 Lap RYGB	6.7% (2/30) in 1st year regained weight-up to 50% of weight lost. All patients were still overweight to severe obesity at 1 yr follow-up.	NR	2/36 were converted to an open procedure (5.6%) Minor complications (readmission < 7 days): 9/36 (25%) Moderate complications (readmission or sequelae for 7–30 days): 4/36 (11%)	Severe complication (sequelae for more than 30 days): 2/36 (5.6%), which includes 1 death 9 months post-operative due to complications from severe infectious colitis. Non-compliant with 12 mo. office visit: 23% (9/39)
Collins 2007110 Stanford 2003123 Lap RYGB	NR	Diabetes: 50% (3/6) Hypertension: 50% (3/6) Obstructive sleep apnea: 100% (2/2) no longer required constant positive airway pressure at night Polycystic ovarian syndrome: 67% (2/3) All co-morbidities: 30.1% resolved	Postoperative bleeding: 3/11 (27.3%) with 1 of these needing laparoscopic reevaluation. Marginal ulcer: 2/11 (18.2%) (1 and 18 mo postoperative) Non-compliant with vitamin regimen: 18.2% (2/11)
Sugerman 2003108 Gastric Bypass 91%	15% (5/33) regained all or most of weight lost at 5–10yrs	Diabetes Mellitus II: 100% (1/1) Hypertension: 80% (8/10) Sleep apnea: 100% (6/6)	Late complications: 21% (7/33) Incisional hernia: 18.2% (6/33) Bowel obstruction: 3% (1/33) Conversions to another type of bypass due to late weight gain or severe protein-calorie malnutrition: 6% (2/33)	Early complications: Pulmonary embolism: 3% (1/33) Major wound infection: 3% (1/33) Minor wound infection: 12% (4/33) Stomal stenoses: 9% (3/33) Marginal ulcers: 12% (4/33) No patients had evidence of impaired sexual or physical maturation.
Soper 1975120 Anderson 1980121 Open RYGB; Horizontal GP	NR	NR	Revision: 5.6% (1/18) Wound infection: 12% (3/25*) Respiratory difficulty: 12% (3/25*) Thrombophlebitis: 4% (1/25*) Upper gastrointestinal bleed: 4% (1/25*)	Urinary tract infection: 4% (1/25*) Protracted vomiting: 4%(1/25*) Incisional hernia: 16% (4/25*) *n=25, which includes 7 Prader-Willi patients
Mason 1995119 VBG	NR	NR	Mortality: None Revisions: 8.5% (4/47)
Capella 2003118 Open RYGB; VBG	NR	NR	Mortality: None Revisions: 10.5% (2/19) Cholecystectomy: 5.3% (1/19)
Strauss 2001122 Open RYGB	3 women who became pregnant regained 13–45 kg	NR	Protein-calorie malnutrition/micronutrient deficiency: 10% (1/10) Cholecystectomy: 20% (2/10) Small bowel obstruction 10 yrs postoperative: 10% (1/10) Incisional hernia: 10% (1/10)
Barnett 2005113 Open RYGB; VBG; JIB	NR	Hypertension: 100% (5/5) Asthma: 66.7% (2/3) Sleep apnea: 100% (2/2) Diabetes: 100% (1/1) Hypothyroidism: 0% (0/1)	Mortality: None Dumping syndrome: 14.3% (2/14) Surgical site infection: 7.1% (1/14) Hypoglycemia: 7.1% (1/14)
Breaux 1995124 Open RYGB; VBG; BPD	NR	Sleep apnea: 100% (11/11)	Mortality: 2 deaths at 15 mo and 3.5 yrs postoperative. Incisional hernia: 5% (1/22) Postoperative laryngeal edema: 5% (1/22)	Gallstones: 5% (1/22) Kidney stones: 5% (1/22) Nutritional deficiencies: 23% (5/22) Revision: 4.5% (1/22)
Rand 1994125 Open RYGB; VBG	NR	NR	2 cholecystectomies 1 abdominal panniculectomy No other AE reported. 3 had surgical revisions-2 were scheduled for revisions.
Papadia 2007128 BPD	NR	Hypertensive: 92% (27/33) Dyslipidemic: 100% (11/11) Hyperglycemic: NR Diabetes mellitus II: 100% (2/2)	Reoperations: 19 in 14 patients (14/68=21%) Mortality long-term: 4.4% (3/68) Protein malnutriiton 1–10 yrs post surgery: 16% (11/68) Immediate complication: 1.5% (1/68)
Tsai 2007109	NR	NR	Mortality: None Major complications: 5.5%78.3% (119/152) of major complications were respiratory

Abbreviations: AE- adverse events; EWL - Excess weight loss; RYGB - Roux-en-Y gastric bypass; VGB - Vertical banded gastroplasty; BPD - Biliopancreatic diversion; JIB - Jejunoileal Bypass; NR - Not reported; GP - Gastroplasty

Short-term (6–12 month) Weight Changes After LAGB (KQ1)

Three studies ^{111, 114, 115} reported mean decrease in BMI for the cohort at discrete time-points (6 months, 12 months after surgery, presumably for all participants who were eligible during this duration of followup). Two of these studies also provided data on the same cohort at longer term followup.^{114, 115} Two case series averaged data for participants across a broad duration of followup,^{126, 129} while a third reported weight data based on retrospective self-report only.¹²⁷

Loss to followup and the small number of cases (n=122) make any conclusions drawn from these case series tentative. Available data, however, suggest following gastric banding, patients experienced an average BMI decrease of 5.0 kg/m² (ITT) to 8.1 kg/m² (CC) at 6 months and 9.4 kg/m² (ITT) to10.2 kg/m² (CC) at 1 year. Based on one study in 17 patients,¹¹⁵ 77 percent achieved a BMI less than 35 at 1 year.

Maintenance of Weight Changes After Laparoscopic Adjustable Gastric Banding (KQ2)

Four studies reported weight outcomes measured 2 or 3 years after LAGB and results were in similar range at both time points.^{114, 115, 126, 129} Mean decrease in BMI ranged from 8.2 kg/m² (ITT) to 14.5 kg/m² (ITT) at 2 years, and 7.3 kg/m² (ITT) to 12.6 kg/m² (ITT) at 3 years. In the two studies that also measured BMI at 12 months,^{114, 115} ITT analysis suggests that on average, some weight is regained at 2 years (1.9 kg/m²) and at 3 years (2.1 kg/m²). While experience certainly varies among individuals, these data are roughly consistent with plots of repeated weight measures in individual patients from several case series that suggest BMI decreases after surgery to its nadir at 12–18 months in most patients and then stabilizes or slightly rebounds in those with longer term followup.^{115, 126} The single study with results at time points beyond 3 years suggests that mean BMI decrease was at least maintained at 5 years (ITT analysis), based on followup of 25 individuals.¹¹⁴ Estimates for 7 years followup represent only 10 individuals.

Adverse Effects of LAGB (KQ3)

Other Beneficial Outcomes of LAGB (KQ4)

Across the three LAGB studies reporting whether comorbidities “resolved” post-surgery,^{126, 127, 129} 6/13 with hypertension, 8/9 with type II diabetes, 6/7 with dyslipidemia, and 20/20 with sleep apnea were reported as resolved, as were 9/9 with asthma. Two studies also reported some improvements in quality of life, self-esteem, body image, and satisfaction with having chosen surgery, although the quality and timing of these measurements are not clear.^{127, 129}

Components of Effective LAGB Surgery (KQ5a)

Data were inadequate to examine this question.

Factors Influencing the Effectiveness of LAGB Surgery (KQ5b)

Data were inadequate to examine this question.

Roux-en-y Gastric Bypass (RYGB), Open Vertical-banded Gastroplasty (VBG), and Other Bariatric Surgeries

Eleven fair- or poor-quality case series reported weight and other beneficial outcomes in 41 adolescents after laparoscopically performed RYGB,^{110, 112} in 51 adolescents after open RYGB,^{108, 120} and in 47 adolescents after VBG.¹¹⁹ The remaining six case series (n=167 youth) provided primarily adverse effects data,^{113, 118, 122, 124, 125, 128} as weight outcomes were either self-reported or averaged across very different post-operative time periods. All but two of these series (one evaluating RYGB¹²² and one evaluating biliopancreatic diversion¹²⁸) were further limited by mixing different types of surgeries. Inpatient adverse effects but not weight outcomes, associated with 566 open, primarily RYGB, procedures in youth have also been reported from the Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample.¹⁰⁹

Short-term (6–12 month) Weight Changes After RYGB, VBG, and Other Bariatric Surgeries (KQ1)

Two fair-quality case series in three publications^{110, 112, 123} address laparoscopic RYGB performed in the United States in 41 adolescents and young adults (aged 13 to 21 years). Mean BMI before surgery was 50.5 kg/m² (SD, 2.0 kg/m²) in one group (n=11) ¹¹⁰ and 56.5 kg/m² (SD, 5.2 kg/m²) in the other (n=30).¹¹² Few other patient data were reported, although patients met or exceeded NIH adult criteria for bariatric surgery and had failed previous medical attempts at weight loss. Comorbidities were reported in the smaller case series. Diabetes (6/11), hypertension (6/11), sleep apnea (2/11), and hepatic steatosis (5/11) were most common.¹¹⁰

Laparoscopic RYGB. Measured weight outcomes were available in one of these two series¹¹² and limited to 30 of 36 patients that had accrued sufficient time post-surgically (although 3/6 not included were actually lost to followup). Among these 30 patients, mean decrease in BMI at 12 months was 20.7 kg/m² (SD, 8.1 kg/m²) and individual BMI reductions ranged from 3.3 kg/m² to 43.5 kg/m². Treatment failures in the first year (those who regained up to 50 percent of the weight lost) were reported in two of 30 patients.

Open RYGB. A large fair-quality case series conducted in the United States of 33 adolescents undergoing bariatric surgery at a single institution over 20 years reported on short-term outcomes at 12 months (but also reported outcomes 5, 10, and 14 years after surgery).¹⁰⁸ Most cases (28/33) underwent open gastric bypass, two had laparoscopic gastric bypass, and the remaining three underwent vertical banded or horizontal gastroplasty. Eligible patients were 12 to18 years of age (mean age 16 years) and met NIH bariatric surgery criteria for adults. Almost 60 percent were female, 82 percent were White, and 15 percent were Black. The mean pre-operative BMI of these patients was 52 kg/m² (SD 11 kg/m², range 28 to 91). Hypertension was present in 30 percent (10/33), sleep apnea in 18 percent (6/33), and diabetes mellitus II in 3 percent (1/33) of patients. At 1 year, mean decrease in BMI was 15.5 kg/m² according to intention-to-treat analysis (ITT) and 16 kg/m² according to complete cases (CC). The only other case series followed primarily open gastric bypass bariatric surgeries performed in 18 genetically normal adolescents (under aged 20 years, median age 19) between 1969 and 1973 at a single university hospital pediatric surgery department in the United States.¹²⁰ Median pre-operative weight in eight female and 10 male patients was 147 kg. At 6 months, the median percentage of body weight lost was 14 percent (ITT). At 10 to 15 months, the median percentage of body weight lost was 23 percent (ITT). Complete case analysis at 10–15 months was slightly more optimistic (30 percent body weight lost).

Maintenance of Weight Changes after RYGB, VBG, and other Bariatric Surgeries (KQ2)

Laparoscopic RYGB. Longer-term data on maintenance are not available on laparoscopically performed RYGB.

Open RYGB or VBG. The one large case series, conducted in the United States, of 33 adolescents also reported on longer term outcomes (5, 10, and 14 years) after surgery.¹⁰⁸ Based on complete cases, mean BMI reductions appear to be maintained or enhanced at 5, 10, and 14 years, compared with results at one year. Based on ITT analyses, mean BMI reduction is maintained at 5 years, with some regain of weight suggested by 10 to 14 years. By these time points, however, only a limited number of participants (less than 20) were eligible for followup due to the recency of the surgery and about one-third of these were not actually measured. Long-term (5 or more years) estimates of BMI reduction are very tentative, due to small numbers and the considerable difference between CC and ITT estimates, which vary by 3 to 5 kg/m². While these data provide estimates of average effects at various time points after surgery, they are not very instructive in estimating weight maintenance for individuals, particularly given the drastic reduction in those eligible for long-term followup. In terms of treatment failures, however, five patients of 33 regained most or all of their weight 5 to 10 years after surgery. One patient with horizontal gastroplasty maintained weight loss after 15 years, but then regained the weight.¹⁰⁸

One case series provided longer term outcomes after VBGs, which were performed from 1980-1994 in 47 adolescents aged 14 to 20 years. Mean BMI was decreased 8.7 kg/m² (ITT) to 12.2 kg/m² (CC) at 5 years and 6.8 kg/m² (ITT) to 9.2 kg/m² (CC) after 10 years.¹¹⁹ These results are limited due to unclear methods that may have mixed self-reported and measured weights. These results provide a point of comparison only, since this procedure is not currently widely used.

The other case series of RYGB, VGB, or other gastroplasties averaged weight outcomes from individuals measured over a broad duration of followup rather than at the same post-operative time points.^{113, 118, 122, 124} These averaged weight changes measured at short-term, medium term, and longer term followup were generally across more than 10 years. A recent study on biliopancreatic diversions performed in 68 Italian adolescents over 29 years similarly averaged weight outcomes measured between 2 and 23 years after surgery.¹²⁸ While these studies reported their outcomes for the mean followup time (5 or more years), the combination of weight outcomes over such different time periods of followup makes them of limited use in estimating weight outcomes.

Adverse Effects After Bariatric Surgeries (KQ3)

Laparoscopic RYGB. Among 47 adolescents undergoing laparoscopically performed RYGB, around 39 percent experienced some short-term complications during the first 12 months.^{110, 112} More than 25 percent (13/47) experienced minor complications (requiring a special test, treatment, endoscopy, or hospital readmission for seven days or less). Moderate complications (unanticipated intensive care unit admission, reoperation, or hospital readmission for more than seven days) occurred in about 14 percent (5/36) of patients. Severe complications (threat to life or major organ system failure) were uncommon (2/36), although one death occurred due to infectious colitis. Two of 36 patients undergoing laparoscopic RYGB had to be converted to an open procedure. During the first post-operative year, noncompliance with recommendations for multivitamin use or for clinical monitoring occurred in one-quarter (11/47) of adolescent surgical patients.

Open RYGB. Among 33 adolescents who primarily underwent open RYGB,¹⁰⁸ 30 percent (10/33) experienced early complications, including one pulmonary embolus, one major wound infection, one minor wound infection, three stomal stenoses requiring endoscopic dilatation, and four marginal ulcers requiring medical therapy. In 21 percent of patients (7/33), late complications requiring surgical treatments primarily included incisional hernias, and one of 33 patients required conversion to another type of bypass due to severe protein calorie malnutrition. In other case series^{113, 118, 124, 125} of a mixture of 89 cases undergoing open gastric bypasses and gastroplasties, two deaths were reported at 15 months and 3.5 years post-operatively: it is difficult to determine whether deaths outside the immediate post-operative period are surgery-related. Other complications included cholecystectomies or gallstones reported in six patients, nutritional deficiencies in five patients, and dumping syndrome or hypoglycemia in three patients.

Since outcomes from case series were not systematically assessed, and relied on retrospective review of medical records or patient recall, absolute rates for complications cannot be determined from these data or from another poor quality case series.¹²² However, data from the Nationwide Inpatient Sample on 566 bariatric surgeries (90 percent gastric bypasses) performed in adolescents from 1996 through 2003 found no in-hospital deaths, but did find major complications in 5.5 percent of cases. Over three-quarters (119/152) of major complications were respiratory, including aspiration, postoperative pulmonary edema, pulmonary insufficiency, acute respiratory failure, prolonged ventilation, tracheostomy, or pneumonia.¹⁰⁹

Bilio-pancreatic diversion. In a retrospective medical record review of 68 biliopancreatic diversions performed in Italy in those under aged 18 years, while immediate complications were uncommon (1/68) longer term complications were not.¹²⁸ Long-term mortality was 4.4 percent (3/68), protein malnutrition within 1 to 10 years post-operatively occurred in 11 of 68 (16 percent) patients, and 14 patients underwent 19 reoperations. These data are consistent with findings that BPD incurs higher complication and mortality rates.

Other Beneficial Outcomes After RYGB (KQ4)

Very limited data on patients after laparoscopic or open RYGB suggest decreasing need for hypoglycemic medications in 4/7 of those with diabetes, resolution of hypertension in 11/16 and no longer needing continuous positive airway pressure or resolution in 8/8 patients with sleep apnea.^{108, 110} The reported resolution of comorbidities in a series including various bariatric surgeries, including RYGB, confirms that sleep apnea resolves in all patients (13/13).^{113, 124} Very limited data supports benefits for hypertension (5/5), asthma (2/3), and diabetes (1/1).¹¹³

Components of Effective Bariatric Surgery (KQ5a)

Since the absolute number of bariatric surgeries in adolescents is small, particularly when categorized by surgical type, there are no good data that examine the effectiveness of specific factors, such as surgeon training, experience, or institutional expertise on outcomes, particularly harms. Other potentially important issues include the intensity and professional disciplines involved in both pre-operative evaluation and post-operative followup management.

Factors Influencing the Effectiveness of Bariatric Surgery (KQ5b)

Similarly, limited data prevent the examination of potentially important population or environmental factors, including degree of overweight, medical and psychological history, family factors (including parental overweight and history of parental bariatric surgery), previous nonsurgical weight loss attempts, and compliance with post-operative management.

Behavioral Interventions

Based on our review, there are effective behavioral interventions that can improve weight measures, at least over the short-term, in obese children and adolescents aged 5 to 18 years. We found no evidence addressing weight management approaches in overweight or obese children under 5 years old. Evidence-based treatments for obese children aged 5 to 12 years are limited to behavioral interventions (without pharmacological adjuncts).

Behavioral interventions for obese children and adolescents aged 5 to 18 years in either schools or in specialty health care settings can effectively produce short-term improvements in weight. Very limited evidence suggests that these improvements can be maintained (completely or somewhat) over the 12 months following the end of treatment. The amount of absolute or relative weight change associated with behavioral interventions in these settings is generally modest and varies by intervention intensity and setting.

In school setting interventions, trials reported 0.4 to 2.07 kg/m² difference in mean BMI change between those that were treated and controls at 6 to 12 months, with a pooled estimate of -0.82 kg/m² (CI: -1.18, -0.46) lower BMI in those treated. For an 8-year-old boy or girl, this BMI difference would translate to about a 3 pound difference (assuming growth of 2 inches or less), and for a 12-year old boy or girl this would translate to about a four pound difference under the same growth assumptions. In girls aged 16, this BMI difference would translate to between 4.5 and 5 pounds, depending on growth. For 16-year-old boys the difference would be between 5 and 6 pounds.

Interventions in specialty health care settings (such as pediatric obesity referral clinics) resulted in a 1.9 to 3.3 kg/m² difference in mean BMI change 6–12 months following treatment, compared with controls. For an 8-year-old boy or girl, the largest BMI difference (3.3 kg/m²) would translate to about 12 to 13 pounds (with up to 2 inches of growth). For a 12-year old boy or girl this would translate to 16.6 to 17.75 pounds difference under the same growth assumptions. In girls aged 16, this BMI difference would translate to about 20 pounds, while for boys aged 16, the difference would be between 22 and 23 pounds for two inches of growth or less.

Psychological outcomes were assessed in several trials, suggesting that interventions potentially improve depression, eating disorder pathology, and shape concern. These results, however, are based on minimal data and should be considered tentative. One included trial examined self-esteem and found no differences in change in self-esteem (both groups improved). Data were also mixed in a recent review¹³¹ on self-esteem in overweight children and adolescents.

We found no evidence of adverse effects on growth, on eating disorder pathology, or on mental health. Effects on growth found in this review are consistent with data from several noncomparative studies, including one that followed 158 children for 10 years and found that weight loss was not related to growth in height in a multivariate model controlling for child age, sex, baseline height, baseline percent overweight, and midparent height.¹³² We found little risk of exercise-induced injuries from behavioral interventions. Although these findings are reassuring, they are limited by incomplete reporting, given that fewer than half of behavioral intervention trials in children and adolescents specifically reported on any potential adverse effects. Only four trials of adolescents and two trials with both children and adolescents (representing relatively few total participants, since most trials enrolled fewer than 100 participants) reported results for any single type of adverse event. None of these found any adverse effects of treatment. The data on potential adverse effects are also further limited for children under 12 years of age. Only two studies^{83, 84} reported potential harms in participants in this age group, indicating no adverse impact on height gains in 111 children at 1 year⁸³ or on body satisfaction or appearance at 1 year in 163 children.⁸⁴ One bone fracture was reported among 107 children under aged 12 years participating in supervised exercise.^{93, 94}

Most treatment programs focused on supporting healthy lifestyle changes through establishing healthful eating habits and increasing regular physical activity. While some trials in adolescents had the explicit goal of weight reduction, trials with younger children generally aimed to reduce participants' relative level of overweight through limiting weight gain as the child grew. Many trials utilized behavioral management techniques, such as teaching parents and/or children about goal-setting, relapse prevention, problem-solving, and managing their environment to encourage healthy lifestyle. Teaching behavior management techniques and providing organized physical activity sessions seem to improve the chances of a program's success.

Physical activity is clearly an important factor in altering the balance between caloric intake and expenditure, and therefore has in important role to play in weight loss interventions. All but two interventions in the 18 main trials included either actual exercise sessions or instruction in behavioral management principles targeting exercise. It appears that organized exercise sessions increase the likelihood of treatment success, but this could not be determined conclusively since programs with organized exercise as also tended to be more intensive programs with considerably more hours of contact. Regardless of whether children and adolescents exercise under the supervision of interventionists or on their own time, improved physical fitness is likely beneficial even if it does not increase weight loss.^{133, 134}

Observational data show a relationship between sedentary behavior, such as television and electronic games, and obesity in children.¹³⁵–¹³⁷ Interventions targeting sedentary behavior have reduced weight gain in trials of obesity prevention.¹³⁸ However, the relative importance of targeting sedentary behavior in treatment of obesity could not be determined from the primary trials included in this review. In addition, Epstein and colleagues conducted three studies¹³⁹–¹⁴¹ examining the relative benefits of encouraging obese participants to decrease sedentary behavior, increase physical activity, or both. Taken as a whole, these trials did not demonstrate that any of these three approaches were clearly superior. One trial¹³⁹ found that focusing on sedentary behavior was more effective than focusing on increasing physical activity, but neither of these groups differed from the group that encouraged both approaches. Neither of the remaining two trials found that the approach to physical activity had an impact on the effectiveness of weight-loss interventions.

All programs targeting younger children involved parents, and since parents usually control most of younger children's food intake, the necessity of parental involvement is self-evident. However, since all of the trials in younger children included parents, we have no empirical basis for quantifying the importance of parental involvement in this age group. The few trials in adolescents that included parental involvement were effective. Since these interventions included many components, however, it was impossible to isolate the specific effect of parental involvement in interventions targeting adolescents.

It is difficult to determine how well the results of these trials would generalize to patients in real-world treatment settings. Several studies relied at least in part on media advertisements for recruitment, and may therefore have enrolled participants who are more motivated to lose weight than a typical obese young person. Trials that recruited via screening, actively seeking participants rather than relying on potential participants to contact them, saw only a minority of overweight or obese children actually participate in the research trial. For example, only 38 percent in Graf's study^{86, 142} and 32 percent in McCallum's trial^{78, 84} who met weight criteria actually enrolled in the trials. There may be unmeasured differences between children who did and children who did not participate that influence how well they respond to the intervention. Children and adolescents who participate may have higher levels of motivation, more free time, more concerned parents, more failed attempts at weight loss, or any number of factors that may moderate the effectiveness of the intervention.

Pharmacological Plus Behavioral Interventions

Pharmacological adjuncts to behavioral interventions have been studied only in obese adolescents aged 12 to 18 years that meet adult criteria for class II obesity (mean BMI of 35 to 40 kg/m² at trial entry), but not in less obese adolescents or in children younger than 12. Treatments with pharmacological agents (sibutramine and orlistat) delivered in combination with behavioral interventions over 6 to 12 months have been studied, but longer term results after treatment discontinuation are not available in any of the pharmacological treatment trials. This is an important limitation in our overall knowledge about their beneficial effects. Two small trials in very obese adolescents at high risk for type 2 diabetes mellitus examined the impact of metformin on glucose tolerance, insulin sensitivity, and BMI. These results are preliminary and are not directly applicable to the general population of obese adolescents.

The most informative data on sibutramine comes from a large (n=498) multicenter trial testing 12 months of sibutramine plus a behavioral intervention, compared with the behavioral intervention plus placebo. Participants receiving 10 to15 mg per day of sibutramine treatment plus a behavioral intervention decreased their BMI 2.9 kg/m² at the conclusion of treatment, corresponding to an average weight reduction of 6.5 kg (14 pounds). Trial participants receiving a behavioral intervention (plus placebo) reduced their BMI 0.3 kg/m², which correspond to a weight gain of 1.9 kg (4.2 pounds). The weight reduction possible at 12 months with effective behavioral intervention in specialty health care is similar in magnitude to the benefits achieved with 12 months of sibutramine plus some level of behavioral intervention. Direct head-to-head comparisons would allow us to confirm this impression.

Available data do not allow us to clearly determine whether behavioral interventions that produce similar effects on BMI as sibutramine also produce similar effects on other potentially beneficial outcomes. In most of the sibutramine trials, waist circumference in those receiving sibutramine was significantly reduced, on average 7 to 8 cm compared with 2 to 3 cm reductions in controls. Significant improvements in HDL cholesterol, triglycerides, and glucose tolerance measures (serum insulin and HOMA) were reported in the sibutramine treatment group in the largest multicenter trial (n=498). Trial participants receiving sibutramine were consistently more likely to develop elevated heart rates than placebo-treated participants, but had similar rates of discontinuation due to this side effect. Systolic or diastolic blood pressure (or both) were significantly elevated in about half of trials. These differences, however, were small in magnitude and are of unknown clinical significance. Few other adverse effects with sibutramine treatment were noted, except for one report of increased constipation. Limited evidence suggests no adverse effects on growth or maturation. One trial testing only three months of sibutramine (10 mg/day) plus six months of a behavioral intervention (compared with placebo and a behavioral intervention) showed modest BMI reductions at 6 months (-0.8 and -1.4 kg/m²) in both arms favoring placebo, but these were not statistically significantly different. No adverse effects were reported.

The most informative data on orlistat come from a large multicenter trial (n=539) testing 12 months of orlistat (360 mg/day) treatment plus a behavioral intervention. Mean BMI in this trial was significantly different (-0.55 kg/m²) after treatment, compared with those receiving the behavioral intervention only (who increased their mean BMI 0.3 kg/m²). This difference reflected weight gain in both groups, which was relatively attenuated in the orlistat group. From these results, it appears that the behavioral intervention component of the orlistat trials was ineffective. This could reflect the freedom at each of the 32 centers to use its own approach to the behavioral intervention aspect of the trial with no assessment of delivery.⁹⁹ Therefore, the quality or intensity of the behavioral interventions may have been lacking at some sites. Participants receiving orlistat significantly reduced their waist and hip circumference (2.7 and 1.5 cm respectively), compared with controls (0.9 and 0.1 cm reductions). Serious adverse effects were uncommon. However, mild-moderate gastrointestinal side effects (most commonly oily spotting, evacuation, abdominal pain, fecal urgency, or flatus with discharge) occurred in 20 to 30 percent of patients taking orlistat and 9 percent reported fecal incontinence. Few participants (2 percent) discontinued treatment due to these side effects, although 35 percent overall dropped out before the trial ended. The impact gastrointestinal effects would have on treatment adherences outside a trial setting is unclear. Orlistat treatment did not reduce vitamin A, D, or E levels or affect growth, bone mineral density, or sexual maturation.

Sibutramine appears to have a larger effect on weight than orlistat, although the two drugs have not been compared directly. Only orlistat has been approved for use in pediatric populations (aged 12 years or older) by the FDA. Both drugs have side effects that must be taken into account when considering treatment for an individual patient. While orlistat has a higher rate of adverse effects, the nature of these effects may be less clinically significant than those seen with sibutramine. Both drugs lack evidence of persistence of weight reduction after active treatment ends.

As with the interventions that were limited to behavioral approaches, these trials involving the addition of pharmacological agents may also be subject to limitations in how well they apply to real-world treatment. That is, adolescents participating in these trials may be more or less likely than the average overweight or obese adolescent to respond to the intervention provided. For example, they may have higher levels of motivation to lose weight and therefore do better than the average adolescent, or they may have a greater number of failed weight loss attempts, which may make them less likely to succeed than the typical overweight or obese teen in the community. The supports provided in a typical trial may also exceed those provided in a usual treatment setting.

Surgical Treatments

Some adolescents reach extremely high levels of obesity and experience substantial health problems due to increased weight. For morbidly obese adolescents with obesity-related health problems who have failed intensive efforts at medical management, surgery may offer a treatment of last resort. Case series of laparoscopic adjustable gastric banding, Roux-en-Y gastric bypass, and other bariatric surgery techniques have been reported in a relatively small number of severely obese adolescents. Surgical case series have been based primarily on retrospective medical chart reviews of patients who have received clinical care. Followup in these series can be incomplete and data collection inconsistent. Thus, both data on weight outcomes as well as other beneficial outcomes from surgery are quite limited. Adverse effect documentation may be somewhat better, particularly for serious adverse effects, since these would reflect issues requiring clinical diagnosis and/or treatment.

LAGB is logically the surgical treatment of choice in morbidly obese adolescents who are candidates for bariatric surgery, since it should be completely reversible and potentially less risky than other bariatric procedures. LAGB is done via laparoscopic rather than open surgery (laparotomy). Both absolute weight loss and risks related to the surgery, however, appear to be lower after laparoscopic adjustable gastric banding than after more invasive procedures, including gastric bypass procedures. In one LAGB series (n=53),¹¹¹estimates of mean reduction in BMI at 6 months ranged from 5.0 to 8.1 kg/m² in intention-to-treat and in complete case analyses respectively. We focus on intention-to-treat analyses as the more realistic measure of overall treatment efficacy. In two studies (n=69), estimates for mean BMI reduction at 12 months ranged from 9.4 to 10.1 kg/m². Based on limited longer term followup from the same two studies.^{114, 115} BMI reductions somewhat reversed between one and three years after surgery (from 10.1 kg/m² at 1 year to 8.2 kg/m² at 2 years and from 9.4 kg/m² at 1 year to 7.3 kg/m² at 3 years). Little data are available to estimate the proportion achieving clinically significant thresholds of weight reduction after surgery or the proportion that fail bariatric surgeries. One small study^{115, 116, 130} (n=17) reported that three-quarters of patients at 12 months and 82 percent at 24 months achieved a BMI less than 35. Similarly, a single case series^{117, 129} of 50 patients reported that only 3/50 (6 percent) did not achieve at least 25 percent body weight loss at one-year post-surgery. No perioperative mortality or major morbidity after LABG has been reported. Limited data suggested 10 to 13 percent of adolescents undergoing LABG require reoperations for band repositioning or removal. Around 10 percent may also have nutrition-related complications (mild hair loss or iron deficiency). Other miscellaneous complications were rarely noted. Very little data are available on whether comorbidities resolved after surgery. It seems clear, however, that those with sleep apnea and probably weight-associated asthma experience resolution, given the degree of weight loss induced by surgery.

A greater reduction in BMI has been seen in adolescents undergoing Roux-en-Y gastric bypass (RYGB) or vertical banded gastroplasty (VBG) procedures. In one small case series^{120, 121} of 18 adolescents whose median preoperative weight was 147 kg, median percentage of body weight lost at 10 to 15 months was 23 percent. At 12 months after RYGB surgery (performed laparoscopically or requiring a laparotomy) in two studies (n=63 adolescents), mean reductions in BMI ranged from 15.5 to 20.7 kg/m². Among 24 patients with ongoing followup,¹⁰⁸ mean BMI appeared to be maintained at 5-year followup. Followup data beyond five years are very limited (less than 20 persons eligible and fewer with measured weights). Most studies that report data on followup longer than one year after surgery are uninformative due to averaging weight measurements taken from individuals at markedly different points of time after surgery (often over 10 years apart). Further, only small numbers of patients are eligible for longer term post-surgical followup, given the rarity of performing bariatric surgery in adolescents during this time period. Treatment failures, however, have been reported even among these limited data. In one series,¹¹² two of 30 patients regained up to 50 percent of the weight lost within the first year. Five of 33 patients regained most or all of their weight 5 to 10 years after RYGB.¹⁰⁸ In both of these cases, patients met NIH inclusion criteria for adults. In a large nationally representative study of inpatient data from 566 RYGB or gastroplasty surgeries in adolescents, no in-hospital deaths were recorded, but major complications occurred in 5.5 percent of patients (two-thirds of which were respiratory). Longer-term adverse events were not captured. Other data suggest, however, that complications occur in at least 30 percent of patients during the first year after open RYGB, and in at least 39 percent in the first 12 months after laparoscopically performed RYGB. After laparoscopically performed RYGB, severe complications (death or severed organ failure) were reported in 2/36 patients and 5/36 patients experienced reoperation, unanticipated intensive care unit admission, or hospital readmission for more than seven days. About one-quarter of patients (13/47) required some special test, treatment, endoscopy, or hospital readmission for seven days or less.

At five years after VBG surgery, three-quarters of patients achieved over 25 percent excess weight loss, although this estimate was lower (61 percent) at “last followup.” ¹¹⁹ This procedure is not currently in widespread use due to higher recidivism than other surgeries and the advent of gastric banding. Although biliopancreatic diversion surgeries (with or without duodenal switching) are not currently in widespread use, it is worth noting that significant harms, including long-term mortality, were reported in 4.4 percent and protein-calorie malnutrition in 16 percent of patients within one to 10 years after surgery.¹²⁸ These data suggest this procedure may be too risky to be considered in obese adolescents.

Even more so than the children and adolescents participating in behaviorally based treatments (with pharmacological adjuncts or those without), adolescents receiving bariatric surgeries were a highly selected group of extremely obese primarily older adolescent patients (with average pre-surgical weights ranging from 284 to 297 pounds) that were often accrued over many years of practice. Many if not most had obesity-related co-morbidities. While bariatric surgeries may provide life-saving treatments for some morbidly obese adolescents, the very limited data currently available on treatment efficacy, along with the known short-term risks and unknown long-term implications of bariatric surgery, demand the utmost care and consideration before choosing these types of treatments and conducting prospective collection of long-term outcomes.^{45, 143}

Long-term Maintenance

It is unfortunate, although not surprising, that evidence of treatment maintenance is quite limited in behavioral intervention trials and surgery studies, and nonexistent in trials of pharmacological treatments. Long-term outcomes are particularly important for surgical treatments, especially in younger adolescents, in whom continuing growth and maturation are complicating factors. The effects of mechanically restricting absorption or the size of the stomach in these children, and of potentially substantial weight loss, cannot be ascertained from the adult literature.

Although this review focused on controlled trials, we searched for additional evidence that may shed light on long-term effectiveness of behavioral intervention programs. An observational study of a behavioral intervention by Epstein and colleagues reported on 10-year followup of four comparative effectiveness treatment trials in children 6 to 12 years of age that were conducted between 1981 and 1986.¹⁴⁴ It did not meet our inclusion criteria because it had no control group for comparison purposes, and it is unclear what proportion of the original participants provided 10-year followup data. Epstein and colleagues report that 30 percent of their participants were not obese at 10-year followup. It is difficult to determine, however, whether this is a higher rate of change than would be seen in a general population of obese children, many of whom likely seek assistance naturalistically in various forms. Freedman and colleagues' large scale observational study of children in Bogalusa, Louisiana²⁷ found that 22.8 percent of 9 to 11 year olds who were at or above the 95^th percentile were no longer obese an average of 16 years later, which is lower than the 30 percent found by Epstein and colleagues. On the other hand, a retrospective observational study from the UK found that 39.3 percent of obese 16-year-olds were no longer obese at age 30, which is a higher rate of remission than that reported by the Epstein study. Several differences between the populations and settings of these studies limit drawing definitive conclusions about whether children undergoing treatment programs are more or less likely to be obese at long-term follow-up. Limited as it is, the best evidence remains that described for KQ2 addressing maintenance effects after treatment, in which control groups were comparable to the treated participants and outcomes were measured consistently between the groups. Even longer-term followup of participants in these trials could be very informative.

Applicability to Vulnerable Populations

As discussed, research on treating obesity must be considered in terms of its applicability to the general population of obese children and adolescents and, in particular, those bearing the greatest burden due to higher prevalence of obesity. These vulnerable groups include racial and ethnic minorities^{13, 15} and those within lowest income groups,¹⁹ who disproportionately bear the brunt of the obesity epidemic.

Minority involvement in addressing the obesity epidemic will be essential, and as such, their involvement in obesity research is critical. Five^{71, 74, 75, 77, 89} of the behavioral intervention trials with short-term outcomes had 10 percent or more of the children and adolescents in their samples classified as Hispanic, including two trials that comprised only Mexican-American participants.^{71, 75} The remaining three reported 24.7 percent,⁷⁴ 15.9 percent,⁷⁷ and 12.5 percent⁸⁹ Hispanic samples. All of these, except the trial with the least-intensive intervention⁸⁹ found that the intervention programs improved weight outcomes. The highest-intensity trial⁷⁴ of these five reported that there were no differences in any outcome measure between ethnic groups. This, coupled with the fact that both of the trials with 100 percent Mexican-American participants were successful, indicates that behavioral interventions can have an impact in Hispanic young people. Two of the trials had more than 10 percent of their samples classified as Black,^{74, 76} one of which included 38.5 percent Black children. This trial successfully promoted weight loss⁷⁴ and reported no ethnic differences on any outcomes. The other ⁸⁹ did not improve weight loss outcomes, included 26.3 percent Black youth, and did not report on the impact of ethnicity on treatment outcome. None of the trials with maintenance outcomes reported more than minimal inclusion of Black or Hispanic children and adolescents.

We found no evidence to suggest that medication treatment is more or less effective in Black or Hispanic than in White youth. Black and Hispanic youth were present in the samples of most of the medication trials, although only three^{98, 99, 104} examined differential impact of treatment by ethnicity: large-scale trials of sibutramine,⁹⁸ orlistat,⁹⁹ and a small trial of metformin.¹⁰⁴ None of these trials found that race had an effect on response to treatment. Data on minority youth in surgical case series were reported in only two trials,^{108, 111} which involved a total of nine Black and seven Hispanic youth between the two trials. No results were reported specifically on the minority youth in either study.

Little was reported about the socioeconomic status of participants in any of the studies. Given the lack of universal access to health care, however, programs delivered through health care settings could be out of reach of many. Public school programs, however, could be available to most if not all children.

Applicability to Real-world Settings

While behavioral interventions are all ostensibly applicable to real-world settings, three of the trials^{69, 71, 75} conducted in schools involved programs that would likely be truly feasible for schools to offer during school hours as alternative health and physical education classes without extensive financial investment. All three of these programs were conducted all or mostly during school hours, and could be included in a school curriculum with some additional resources to support teacher training and planning, the acquisition of materials, and consultation with experts such as dietitians and behavioral specialists. Research on dissemination of programs such as these would be extremely valuable.

Higher intensity programs that were conducted in specialty care settings may also be feasible for many health care settings, perhaps at little extra cost. It may be possible to adapt the detailed protocols developed for use in the trials included in this review. For example, the comprehensive and effective Bright Bodies weight management program developed by Savoye and colleagues,⁷⁴ was facilitated by a registered dietitian or social worker and an exercise physiologist. A team of professionals in these or related fields would likely have the requisite training to conduct this type of program without extensive additional training. Third-party payment for these types of programs or indication of their cost-effectiveness would assist in their uptake in the real world.

Two of the behavioral intervention programs specifically addressed the use of very-low-intensity interventions (approximately four hours of total intervention time) that could be integrated into primary care.^{77, 78, 84} Only one of these improved short-term weight loss,⁷⁷ and could be feasible for implementation in some primary care practices, if it is proven to be beneficial through replication. This program relied on bachelors-level support staff to provide adjunctive care via mail and phone counseling, thus relieving the primary care provider of some of the burden of conducting the intervention. Dissemination research would be needed to truly determine the wide-spread feasibility of this and other ostensibly feasible programs.

While pharmacological treatments have been studied in multi-site clinical trials, which enhances their applicability, treatment adherence outside of the trial setting and longer term weight impacts remain unclear. And, as recommended by experts, surgical treatments should probably be delivered in centers of excellence for bariatric surgery, with adaptation to the nutritional, psychological, and medical needs of adolescents.⁴⁵

Factors Contributing to the Recent Increase in Childhood Obesity

While many experts have speculated on the causes of the recent increases in childhood obesity,^{145, 146} data are not available to conclusively determine causality. Evidence does support, however, a relationship between childhood obesity and several factors, such as overall physical activity, sedentary behaviors (e.g., watching television, playing video games, and spending time on computers), and intake of sweetened beverages.¹¹ Children (ages 2 to 17) average 4.7 hours per day “screen time” (covering cluster of activities involving television and computer screens, such as TV viewing, DVDs/videotapes, video games, computer games, e-mail and other computer activities).¹⁴⁷ Cross-sectional data show that higher prevalence of obesity is associated with more hours per day watching television.^{136, 137} Also, an obesity prevention program that reduced screen time by an average of almost ten hours per week also resulted in a BMI reduction of 0.45 kg/m² in sample of 3^rd and 4^th grade school children.¹³⁸ Environmental factors have likely reduced the amount of physical activity children get currently. For example, in 1969, 42 percent of children walked to or rode their bikes to school, while only 16 percent of children did so in 2001.¹⁴⁸ Also, enrollment in physical education classed declined from 41.6 percent in 1991 to 28.4 percent in 2003 in high school students.¹⁴⁹ Longitudinal and cross-sectional observational data have demonstrated that higher levels of physical activity tend to be associated with lower BMIs in children.^{136, 150} In one study, an increase in one hour/day of physical activity was associated with a BMI decrement of 0.22 kg/m² in boys and 0.16 kg/m² in girls after one year.¹⁵⁰

Similarly, intake of sweetened beverages has also increased and appears to contribute to childhood obesity.^{11, 151}–¹⁵³ Between the late 1970s and the late 1990s, average daily intake of sweetened beverages increased from 5 ounces to 12 ounces in 6 to 17 year-olds.¹⁵³ BMI increases by an estimated 0.01 kg/m² with every 100 grams of regular soda consumed daily in adolescent girls, but this is not true of other beverages.¹⁵¹ The odds of obesity increases by 60 percent with each additional serving of sugar-sweetened soda consumed in children.¹⁵⁴

Preventing Childhood Obesity and Overweight

While this report focuses on the effectiveness and benefits of treatments in children and adolescents who are already overweight or obese, the challenge of achieving significant weight loss (and the uncertainty as to how well any weight reduction can be maintained) reaffirms the importance of obesity prevention. Obesity prevention is a critical component of the full breadth of a public health approach to overweight and obesity among American children and adolescents. Preventive approaches address some of the factors discussed above and emphasize helping children and adolescents develop lifelong healthy habits, in order to prevent the development of overweight or obesity during childhood and into adulthood. Obesity prevention should be conceptualized broadly, to include ecological interventions as well as health promotion campaigns in schools, communities, and health care settings.

Table 14

National public health priority recommendations from (more...)

Table 14

National public health priority recommendations from IOM for childhood and adolescent obesity prevention

Recommendation 1.National Priority. Government at all levels provides coordinated leadership for the prevention of obesity in youth and children, with coordinated budgets, policies, and program requirements and with an increased and sustained commitment of federal and state funds and resources. Recommendation 2.Industry. Industry should develop and promote products, opportunities, and information that will encourage healthful eating behaviors and regular physical activity. Recommendation 3.Nutrition labeling. FDA should revise nutrition labeling and health claims approaches so that parents and youth can make informed product comparisons and decisions to achieve and maintain energy balance at a health weight. Recommendation 4.Advertising and Marketing. Industry should develop and strictly adhere to marketing and advertising guidelines that minimize the risk of obesity in children and youth, and the FTC should be the monitoring agency for compliance with these standards. Recommendation 5.Multi-Media and Public Relations Campaign. DHHS should develop and evaluate a long-term national multi-media public relations campaign focused on obesity prevention in children and youth. Recommendation 6.Community Programs. Local governments, public health agencies, schools, and community organizations should collaboratively develop and promote program to encourage healthful eating behaviors and regular physical activity, particularly for high-risk populations in order to eliminate health disparities. Recommendation 7.Built Environment. Local governments, private developers, and community groups should expand opportunities for physical activity through recreational facilities, parks, playgrounds, sidewalks, bike paths, routes for walking or biking to school, and safe streets and neighborhoods, particularly for populations at high-risk of childhood obesity. Recommendation 8.Healthcare. Pediatricians, family physicians, nurses, and other clinicians should engage in the prevention of childhood obesity, with support from professional organizations, insurers, and accrediting groups for individual and population-based obesity prevention efforts. Recommendation 9. Schools. Schools should provide a consistent environment conducive to healthful eating behaviors and regular physical activity, supported by federal and state departments of education and health and professional organizations. Recommendation 10.Home. Parents should promote healthful eating behaviors and regular physical activity for their children through breast-feeding, providing health food and beverage choices, teaching children to make healthful food and beverage choices, supporting regular physical activity, limiting recreational screen time to under 2 hours per day, monitoring and discussing weight status with the child's healthcare clinician, and serving as positive role models. Adapted from Preventing Childhood Obesity: Health in the Balance. IOM 2005.

To support the broad public health recommendations called for in the recent IOM report, international experts are engaged in ongoing activities, including summarizing available research to inform best strategies for health promotion and primary prevention of childhood obesity through policies and programs in healthcare and other community settings. The CDC is undertaking a series of reports on evidence to support obesity interventions in schools, community-settings, and health systems, which are made publicly available as they are completed.¹⁵⁵ The CDC also provides statistics on the prevalence of childhood obesity by state and year, data from the School Health Policies and Programs Study and from the Youth Behavioral Risk Factor Surveillance System, and information about state and local programs.¹⁵⁶ The National Institute for Clinical Excellence (NICE) in the United Kingdom made its comprehensive evidence-based clinical guideline on both obesity prevention and treatment in adults and children available in December, 2006.² Other systematic reviewers have published reports recently examining the effectiveness of preventive interventions and factors associated with etiology and risks. The Robert Wood Johnson Foundation's Active Living by Design Program has sponsored considerable research that has supported a link between the built environment and physical activity. Reviews of the impact of urban planning and obesity have concluded that “(1) areas with mixed land use, greater residential and commercial densities, grid street networks, and sidewalks are associated with more walking, biking, and pubic transportation usage; and (2) children with access to parks, recreation facilities, and programs are more physically active than children without access”.¹⁵⁷ Given the relatively small effects seen in most behavioral interventions, and the fact that more invasive interventions are only appropriate for a small portion of the population, prevention programs are likely to be the most effective agents in slowing the growth of childhood obesity.

Limitations in the Body of Evidence

The quality of research on treating child and adolescent obesity has improved substantially since the 2003 Cochrane review and our 2005 review which both enumerated concerns about the childhood obesity treatment literature, specifically regarding behavioral interventions. These concerns included small sample sizes, high attrition (among other quality issues), less than ideal outcome measures, and highly heterogeneous treatment approaches.⁵¹ Most (15/18) of the behavioral interventions included in our review were published since the end of the search window for these prior reviews, including seven published in 2007 and three in 2006. Several of the newly published trials have over 100 participants, although larger trials can be quite expensive. While retention remains somewhat problematic, eight of the 15 newer trials reported overall retention of 89 percent or higher. Outcome measurement has improved as well—almost all of the newer trials reported raw BMI scores or BMI SDS and all directly measured their participants rather than relying on self-report (though some did fill in missing data with self-report measures). A lingering quality issue, however, is that the blinding procedures for treatment allocation and outcomes assessment were often not described. And, research would be improved with more explicit reporting of intervention fidelity and of impacts on other outcomes (both harms and benefits, such as comorbidities), in addition to weight. Finally, while treatment trials remain quite heterogeneous, it is hoped that better reporting and growth in the research base will eventually allow determination of effective components of behavioral interventions.

While methods and reporting have improved, and the number of studies has increased, the large amount of heterogeneity in the behavioral intervention literature (e.g., populations, intervention intensity, settings, treatment components, types of outcomes assessed) makes providing summary measures of expected treatment effects still very difficult. Thus, our findings and meta-analysis should be interpreted with caution. While it appears that treatment settings were the major factor differentiating size of treatment effect, other factors such as treatment intensity and age also appear to be important and may have been inappropriately combined in our meta-analysis. Because change in BMI has a different meaning for children of different ages, it would have been preferable to analyze change in BMI SDS, which is adjusted for age and sex. However, many authors did not report BMI SDS, and because special software or look-up data are needed to calculate BMI SDS, it was not feasible to expect authors to provide this data upon request.

While larger trials of pharmacological treatments are quite recent (2005 and 2006), as are better quality surgical case series (2005 to 2007), the available treatment data for these approaches remains limited. There are only two weight-loss medications studied (sibutramine, orlistat), with few randomized trials overall, and only one large-scale trial of each of the medications. No trials were conducted among children age 11 years or younger, so no conclusions can be drawn regarding efficacy or safety for that age group. We found no data on long-term maintenance of treatment effect or safety. The longest treatment period studied was 12 months, and the only followup reported for either medication was 3 months after medication use terminated. Medication use may have either a positive or negative effect on long-term maintenance of weight changes, compared with exclusively behavioral approaches, so longer follow-up is very important. While we found sibutramine and orlistat each had one large-sample trial, these trials were not large enough to detect more rare but serious adverse effects. The high variability across trials in intensity and possibly of intervention fidelity for behavioral interventions hampered our ability to determine both the combined and independent effect of the medication.

Surgical case series are not considered to be strong evidence as these are non-comparative studies. Without a good understanding of the natural history of weight in severely obese children, it is difficult to determine if the case series are giving an accurate estimate of the effect of surgery compared with no treatment. Lack of prospective, research-designed data collection also limits the results.

The research on all types of obesity treatments remains limited for its focus on obese (or highly obese) children and adolescents. While focus on more obese adolescents is appropriate for pharmaceutical and surgical treatments, future researchers evaluating all three types of weight management approaches should address current limitations by ensuring that their studies enroll the range of obese (or overweight) children and adolescents who might benefit, and for whom the level of treatment-associated risk is appropriate. Future researchers should also address limitations in research on children under aged 6 and ensure that treatment studies enroll and evaluate race-specific effects among adequate numbers of racial and ethnic minority participants. Further data on long-term maintenance of treatment effects (benefits and harms), and better reporting of the effect of treatment on co-morbidities will address these important limitations in the currently available evidence.

Limitations in our Approach

We conducted comprehensive searches of multiple electronic databases (including those with dissertation abstracts), reviewed bibliographies, and contacted experts, but did not hand-search or otherwise review gray literature. We may not have located all relevant studies through this approach. We also did not formally assess for publication bias, given the heterogeneity of outcomes reported in included studies. Thus, it is possible that our review overestimates overweight treatment efficacy due to the “file drawer” problem whereby ineffective treatment studies are more likely to be unpublished. Finally, our review did not include all studies that others might consider relevant. We did not do a comprehensive assessment of comparative effectiveness trials, as our primary goal was to determine whether treatment worked and the size of the effects compared with no treatment. The comparative effectiveness literature was fairly extensive, and included considerable older work completed by Epstein and colleagues as well as other researchers, which represents the majority of research available for earlier reviews. We could not be confident that comparative effectiveness results would tell us about the overall effectiveness of either treatment approach tested because good, recent data could not be found on the natural history of childhood obesity. Also, there was a great deal of variability in the basic weight management approach and in the reporting of the programs, so we did not believe that effectiveness of individual components could be accurately isolated. After consultation with our Technical Expert Panel, we chose to limit our use of comparative effectiveness trials to further explore approaches (e.g., physical activity components, behavioral management techniques, and parent involvement) that seemed to be important components in those interventions that were shown to be effective when tested against minimally treated control groups.

Our examination of other beneficial outcomes was limited to studies that met our general inclusion criteria, including reporting some measure of weight change six months or more after the baseline assessment. Given the primary purpose of this review (focus on weight management) and with support of our Technical Expert Panel, we did not include trials that reported other beneficial outcomes without some measure of weight change, and therefore may have missed some reports of other beneficial outcomes.

We did not address the impact of population-based prevention programs on weight reduction in overweight or obese children. These programs are primarily targeted at preventing obesity, but since some children participating in these programs are already overweight or obese when they begin, it would be useful to know the degree to which overweight and obese children benefit. It would also be useful to know whether overweight and obese children suffer deleterious effects of such programs, such as increased dieting, increased teasing, or poorer self-esteem.

References Reason for Exclusion

Alexy U, Reinehr T, Sichert-Hellert W, Wollenhaupt A, Kersting M, Andler W. Positive changes of dietary habits after an outpatient training program for overweight children. Nutrition Research. 2006; 26(5): 202–208. Not a weight loss trial.

Amador M, Ramos LT, Morono M, Hermelo MP. Growth rate reduction during energy restriction in obese adolescents. Exp Clin Endocrinol. 1990; 96: 73–82. Setting. [PubMed]

Ambler C, Eliakim A, Brasel JA, Lee WN, Burke G, Cooper DM. Fitness and the effect of exercise training on the dietary intake of healthy adolescents. Int J Obes Relat Metab Disord. 1998; 22: 354–362. Relevance. [PubMed]

Arnold, Linda L. The effects of a program of exercise and nutrition on body composition in adolescents and young adults with moderate cognitive disabilities: A descriptive study. Dissertation Abstracts International: Section B: The Sciences and Engineering. 2005; 65(11-B): 6062. Population.

Ask AS, Hernes S, Aarek I, Johannessen G, Haugen M. Changes in dietary pattern in 15 year old adolescents following a 4 month dietary intervention with school breakfast—a pilot study. Nutrition Journal. 2006; 5: 33. Did not report relevant outcomes. [PubMed] [Free Full text in PMC]

Atlantis E, Barnes EH, Singh MA. Efficacy of exercise for treating overweight in children and adolescents: a systematic review. Int J Obes (Lond). 2006; 30: 1027–1040. Design. [PubMed]

Balagopal P, George D, Patton N. et al. Lifestyle-only intervention attenuates the inflammatory state associated with obesity: a randomized controlled study in adolescents. Journal of Pediatrics. 2005; 146(3): 342–8. Design. [PubMed]

Balagopal P, George D, Yarandi H, Funanage V, Bayne E. Reversal of obesity-related hypoadiponectinemia by lifestyle intervention: a controlled, randomized study in obese adolescents. Journal of Clinical Endocrinology & Metabolism. 2005; 90(11): 6192–7. Design. [PubMed]

Bauer C, Fischer A, Keller U. Effect of sibutramine and of cognitive-behavioural weight loss therapy in obesity and subclinical binge eating disorder. Diabetes, Obesity & Metabolism. 2006; 8(3): 289–95. Design.

Baumer JH. Obesity and overweight: its prevention, identification, assessment and management. Archives of Disease in Childhood Education & Practice. 2007; 92(3): ep92–6. Design.

Becque MD, Katch VL, Rocchini AP, Marks CR, Moorehead C. Coronary risk incidence of obese adolescents: reduction by exercise plus diet intervention. Am J Clin Nutr. 1988; 81: 605–612. Design.

Beech BM, Klesges RC, Kumanyika SK. et al. Child- and parent-targeted interventions: the Memphis GEMS pilot study. Ethn Dis. 2003; 13: S40–S53. Design. [PubMed]

Berry D, Savoye M, Melkus G, Grey M. An intervention for multiethnic obese parents and overweight children. Applied Nursing Research. 2007;63–71, 2007. Design.

Berry D, Sheehan R, Heschel R, Knafl K, Melkus G, Grey M. Family-based interventions for childhood obesity: a review (Structured abstract). SO: Journal of Family Nursing. 2004; 10: 429–449. Design.

Braet C, Van WM, Van LK. Follow-up results of different treatment programs for obese children. Acta Paediatr. 1997; 86: 397–402. Design. [PubMed]

Braet C, Winckel MV. Long-Term Follow-Up of a Cognitive Behavioral Treatment Program for Obese Children. Behavior Therapy. 2000; 31: 55–74. Design.

Brown R, Sothern M, Suskind R, Udall J, Blecker U. Racial differences in the lipid profiles of obese children and adolescents before and after significant weight loss. Clin Pediatr (Phila). 2000; 39: 427–431. Design. [PubMed]

Brownell KD, Kaye FS. A school-based behavior modification, nutrition education, and physical activity program for obese children. Am J Clin Nutr. 1982; 35: 277–283. Design. [PubMed]

Butryn ML, Wadden TA. Treatment of overweight in children and adolescents: does dieting increase the risk of eating disorders? (Structured abstract). SO: International Journal of Eating Disorders. 2005; 37: 285–293. Design.

Campbell KJ, Hesketh KD. Strategies which aim to positively impact on weight, physical activity, diet and sedentary behaviours in children from zero to five years. A systematic review of the literature. Obesity Reviews. 2007; 8(4): 327–38. Did not report relevant outcomes. [PubMed]

Chang FT, Hu SH, Wang RS. The effectiveness of dietary instruction in obese school children of southern Taiwan. Kaohsiung J Med Sci. 1998; 14: 528–535. Setting. [PubMed]

Chen W, Chen SC, Hsu HS, Lee C. Counseling clinic for pediatric weight reduction: program formulation and follow-up. J Formos Med Assoc. 1997; 96: 59–62. Setting. [PubMed]

Clemmens D, Hayman LL. Increasing activity to reduce obesity in adolescent girls: a research review (Provisional record). SO: Journal of Obstetric, Gynecologic and Neonatal Nursing. 2004; 33: 801–808. Design.

Cliff DP, Wilson A, Okely AD, Mickle KJ, Steele JR. Feasibility of SHARK: A physical activity skill-development program for overweight and obese children. Journal of Science & Medicine in Sport. 2007; 10(4): 263–7. Design. [PubMed]

Cole K, Waldrop J, D'Auria J, Garner H. An integrative research review: effective school-based childhood overweight interventions. J Spec Pediatr Nurs. 2006; 11: 166–177. Design. [PubMed]

Coleman KJ, Tiller CL, Sanchez J. et al. Prevention of the epidemic increase in child risk of overweight in low-income schools: the El Paso coordinated approach to child health. Archives of Pediatrics & Adolescent Medicine. 2005; 159(3): 217–24. Relevance. [PubMed]

Collins CE, Warren J, Neve M, McCoy P, Stokes BJ. Measuring effectiveness of dietetic interventions in child obesity: a systematic review of randomized trials. Arch Pediatr Adolesc Med. 2006; 160: 906–922. Design. [PubMed]

Collins CE, Warren JM, Neve M, McCoy P, Stokes B. Systematic review of interventions in the management of overweight and obese children which include a dietary component (Provisional record). SO: International Journal of Evidence Based Healthcare. 2007; 5: 2–53. Design.

Daley AJ, Copeland RJ, Wright NP, Roalfe A, Wales JK. Exercise therapy as a treatment for psychopathologic conditions in obese and morbidly obese adolescents: a randomized, controlled trial. Pediatrics. 2006; 118(5): 2126–34. Relevance. [PubMed]

Daley AJ, Copeland RJ, Wright NP, Wales JK. Protocol for: Sheffield Obesity Trial (SHOT): a randomised controlled trial of exercise therapy and mental health outcomes in obese adolescents [ISRCNT83888112]. BMC Public Health. 2005; 5: 113. Relevance. [PubMed] [Free Full text in PMC]

Danielzik S, Pust S, Landsberg B, Muller MJ. First lessons from the Kiel Obesity Prevention Study (KOPS). International Journal of Obesity. 2005; 29(Suppl2): S78–S83. Relevance. [PubMed]

Davee AM, Blum JE, Devore RL. et al. The vending and a la carte policy intervention in Maine public high schools. Preventing Chronic Disease. 2005; 2(Spec no): A14. Design. [PubMed]

Davis CE, Hunsberger S, Murray DM. et al. Design and statistical analysis for the Pathways study. Am J Clin Nutr. 1999; 69: 760S–763S. Relevance. [PubMed]

DeJongh ED, Binkley TL, Specker BL. Fat mass gain is lower in calcium-supplemented than in unsupplemented preschool children with low dietary calcium intakes. American Journal of Clinical Nutrition. 2006; 84(5): 1123–7. Relevance. [PubMed] [Free Full text in PMC]

DeMattia L, Lemont L, Meurer L. Do interventions to limit sedentary behaviours change behaviour and reduce childhood obesity? A critical review of the literature. Obesity Reviews. 2007; 8(1): 69–81. Design. [PubMed]

Dennison BA, Russo TJ, Burdick PA, Jenkins PL. An intervention to reduce television viewing by preschool children. Arch Pediatr Adolesc Med. 2004; 158: 170–176. Relevance. [PubMed]

Dicken KR, Bell MM. Pedometers as a means to increase walking and achieve weight loss. SO: Journal of the American Board of Family Medicine : JABFM. 2006; 19: 524–525. Population.

Donnelly JE, Jacobsen DJ, Whatley JE, Hill JO, Swift LL, Cherrington A, Polk B, Tran ZV, Reed G. Nutrition and physical activity program to attenuate obesity and promote physical and metabolic fitness in elementary school children. Obes Res. 1996; 4(3): 229–243. Prevention trial. [PubMed]

Dreimane D, Safani D, MacKenzie M. et al. Feasibility of a hospital-based, family-centered intervention to reduce weight gain in overweight children and adolescents. Diabetes Research & Clinical Practice. 2007; 75(2): 159–68. Design. [PubMed]

Ebbeling CB, Feldman HA, Osganian SK, Chomitz VR, Ellenbogen SJ, Ludwig DS. Effects of decreasing sugar-sweetened beverage consumption on body weight in adolescents: a randomized, controlled pilot study. Pediatrics. 2006; 117(3): 673–80. Relevance. [PubMed]

Ebbeling CB, Garcia-Lago E, Leidig MM, Seger-Shippee LG, Feldman HA, Ludwig DS. Altering portion sizes and eating rate to attenuate gorging during a fast food meal: effects on energy intake. Pediatrics. 2007; 119(5): 869–75. Relevance. [PubMed]

Ebbeling CB, Leidig MM, Feldman HA, Lovesky MM, Ludwig DS. Effects of a low-glycemic load vs low-fat diet in obese young adults: a randomized trial. JAMA. 2007; 297(19): 2092–102. Population. [PubMed]

Ebbeling CB, Leidig MM, Sinclair KB, Hangen JP, Ludwig DS. A reduced-glycemic load diet in the treatment of adolescent obesity. Archives of Pediatrics & Adolescent Medicine. 2003; 157: 773–779. Design. [PubMed]

Economos CD, Hyatt RR, Goldberg JP. et al. A community intervention reduces BMI z-score in children: Shape Up Somerville first year results. Obesity. 2007; 15(5): 1325–36. Relevance. [PubMed]

Edwards B. Childhood obesity: a school-based approach to increase nutritional knowledge and activity levels. Nurs Clin North Am. 2005; 40: 661–6ix. Did not report relevant outcomes. [PubMed]

Eliakim A, Kaven G, Berger I, Friedland O, Wolach B, Nemet D. The effect of a combined intervention on body mass index and fitness in obese children and adolescents - a clinical experience. Eur J Pediatr. 2002; 161: 449–454. Did not report relevant outcomes. [PubMed]

Epstein LH, Kuller LH, Wing RR, Valoski A, McCurley J. The effect of weight control on lipid changes in obese children. Am J Dis Child. 1989; 143: 454–457. Precedes search period. [PubMed]

Epstein LH, McCurley J, Wing RR, Valoski A. Five-year follow-up of family-based behavioral treatments for childhood obesity. J Consult Clin Psychol. 1990; 58: 661–664. Design. [PubMed]

Epstein LH, McKenzie SJ, Valoski A, Klein KR, Wing RR. Effects of mastery criteria and contingent reinforcement for family-based child weight control 3838. Addictive Behaviors. 1994; 19: 135–145. Design. [PubMed]

Epstein LH, Paluch RA, Raynor HA. Sex differences in obese children and siblings in family-based obesity treatment. Obesity Research. 2001; 9: 746–753. Design. [PubMed]

Epstein LH, Saelens BE, O'Brien JG. Effects of reinforcing increases in active behavior versus decreases in sedentary behavior for obese children. Int J Behav Med. 1995; 2: 41–50. Did not report relevant outcomes. [PubMed]

Epstein LH, Valoski A, McCurley J. Effect of weight loss by obese children on long-term growth. Am J Dis Child. 1993; 147: 1076–1080. Design. [PubMed]

Epstein LH, Valoski A, Wing RR, McCurley J. Ten-year outcomes of behavioral family-based treatment for childhood obesity. Health Psychol. 1994; 13: 373–383. Design. [PubMed]

Epstein LH, Valosk Ai, Wing RR, McCurley J. Ten-year follow-up of behavioral, family-based treatment for obese children. JAMA. 1990; 264(19): 2519–2523. Precedes search period. [PubMed]

Epstein LH, Valoski AM, Kalarchian MA, McCurley J. Do children lose and maintain weight easier than adults: a comparison of child and parent weight changes from six months to ten years. Obes Res. 1995; 3: 411–417. Did not report relevant outcomes. [PubMed]

Epstein LH, Wing RR, Koeske R, Andrasik F, Ossip DJ. Child and parent weight loss in family-based behavior modification programs. J Consult Clin Psychol. 1981; 49: 674–685. Design. [PubMed]

Epstein LH, Wing RR, Koeske R, Valoski A. Effect of parent weight on weight loss in obese children. J Consult Clin Psychol. 1986; 54: 400–401. Did not report relevant outcomes. [PubMed]

Epstein LH, Wing RR, Koeske R, Valoski A. Effects of diet plus exercise on weight change in parents and children. J Consult Clin Psychol. 1984; 52: 429–437. Precedes search period. [PubMed]

Epstein LH, Wing RR, Koeske R, Valoski A. Long-term effects of family-based treatment of childhood obesity. J Consult Clin Psychol. 1987; 55: 91–95. Precedes search period. [PubMed]

Epstein LH, Valoski A, Koeske R, Wing RR. Family-based behavioral weight control in obese young children. J Am Diet Assoc. 1986; 86(4): 481–484. Design. [PubMed]

Epstein LH, Valoski AM, Vara LS, McCurley J, Wisniewski L, Kalarchian MA, Klein KR, Shrager LR. Effects of decreasing sedentary behavior and increasing activity on weight change in obese children 3814. Health Psychology. 1995; 14(2): 109–115. Design. [PubMed]

Epstein LHR, Paluch RA, Gordy CC, Dorn J. Decreasing sedentary behaviors in treating pediatric obesity. Arch Pediatr Adolesc Med. 2000; 154(3): 220–226. Design. [PubMed]

Epstein LH, Paluch RA, Saelens BE, Ernst MM, Wilfley DE. Changes in eating disorder symptoms with pediatric obesity treatment. J Pediatr. 2001; 139(1): 58–65. Design. [PubMed]

Epstein LH, Paluch RA, Kilanowski CK, Raynor HA. The effect of reinforcement or stimulus control to reduce sedentary behavior in the treatment of pediatric obesity. Health Psychol. 2004; 23(4): 371–380. Design. [PubMed]

Epstein LH, Paluch RA, Roemmich JN, Beecher MD. Family-based obesity treatment, then and now: Twenty-five years of pediatric obesity treatment. Health Psychol. 2007; 26(4): 381–391. Design. [PubMed] [Free Full text in PMC]

Epstein LH, Wing RR, Valoski A, Penner BC. Stability of food preferences during weight control. A study with 8- to 12-year-old children and their parents. Behav Modif. 1987; 11: 87–101. Did not report relevant outcomes. [PubMed]

Figueroa-Colon R, von Almen TK, Franklin FA, Schuftan C, Suskind RM. Comparison of two hypocaloric diets in obese children. Am J Dis Child. 1993; 147: 160–166. Design. [PubMed]

Figueroa-Colon R, Franklin FA, Lee JR, von Almen TK, Suskind RM. Feasibility of a clinic-based hypocaloric dietary intervention implemented in a school setting for obese children. Obes Res. 1996; 4(5): 419–429. Design. [PubMed]

Flodmark CE, Marcus C, Britton M. Interventions to prevent obesity in children and adolescents: A systematic literature review. International Journal of Obesity. 2006; 30(4): 579–589. Relevance. [PubMed]

Flores R. Dance for health: improving fitness in African American and Hispanic adolescents. Public Health Rep. 1995; 110(2): 189–193. Outcomes < 6 months. [PubMed] [Free Full text in PMC]

Foster GD, Wadden TA, Brownell KD. Peer-led program for the treatment and prevention of obesity in the schools. J Consult Clin Psychol. 1985; 53: 538–540. Design. [PubMed]

Gately PJ, Cooke CB, Barth JH, Bewick BM, Radley D, Hill AJ. Children's residential weight-loss programs can work: a prospective cohort study of short-term outcomes for overweight and obese children. Pediatrics. 2005; 116(1): 73–7. Design. [PubMed]

Gately PJ, King NA, Greatwood HC. et al. Does a High-protein Diet Improve Weight Loss in Overweight and Obese Children? Obesity. 2007; 15(6): 1527–34. Design. [PubMed]

Gibson LJ, Peto J, Warren JM, dos SS I. Lack of evidence on diets for obesity for children: a systematic review. International Journal of Epidemiology. 2006; 35(6): 1544–52. Design. [PubMed]

Golan M, Crow S. Targeting parents exclusively in the treatment of childhood obesity: long-term results. Obes Res. 2004; 12: 357–361. Information provided in another publication. [PubMed]

Goldfield GS, Mallory R, Parker T. et al. Effects of modifying physical activity and sedentary behavior on psychosocial adjustment in overweight/obese children. Journal of Pediatric Psychology. 2007; 32(7): 783–93. Design. [PubMed]

Goldfield GS, Epstein LH, Kilanowski CK, Paluch RA, Kogut-Bossler B. Cost-effectiveness of group and mixed family-based treatment for childhood obesity. Int J Obes Relat Metab Disord. 2001; 25(12): 1843–1849. Design. [PubMed]

Gortmaker SL, Peterson K, Wiecha J. et al. Reducing obesity via a school-based interdisciplinary intervention among youth: Planet Health. Arch Pediatr Adolesc Med. 1999; 153: 409–418. Relevance. [PubMed]

Graf C, Koch B, Bjarnason-Wehrens B. et al. Who benefits from intervention in, as opposed to screening of, overweight and obese children? Cardiology in the Young. 2006; 16(5): 474–80. Did not report relevant outcomes. [PubMed]

Graves T, Meyers AW, Clark L. An evaluation of parental problem-solving training in the behavioral treatment of childhood obesity. J Consult Clin Psychol. 1988; 56: 246–250. Design. [PubMed]

Grey M, Berry D, Davidson M, Galasso P, Gustafson E, Melkus G. Preliminary testing of a program to prevent type 2 diabetes among high-risk youth. J Sch Health. 2004; 74(1): 10–15. Design. [PubMed]

Harrell JS, Gansky SA, McMurray RG, Bangdiwala SI, Frauman AC, Bradley CB. School-based interventions improve heart health in children with multiple cardiovascular disease risk factors. Am J Clin Nutr. 1998; 102: 371–380. Relevance.

Harvey-Berino J, Rourke J. Obesity prevention in preschool native-american children: a pilot study using home visiting. Obes Res. 2003; 11: 606–611. Relevance. [PubMed]

Heymsfield SB, van-Mierlo CA, van-der-Knaap HC, Heo M, Frier H I. Weight management using a meal replacement strategy: meta and pooling analysis from six studies (Structured abstract). SO: International Journal of Obesity. 2003; 27: 537–549. Population.

Hills AP, Parker AW. Obesity management via diet and exercise intervention. Child Care Health Dev. 1988; 14: 409–416. Design. [PubMed]

Huang JS, Norman GJ, Zabinski MF, Calfas K, Patrick K. Body image and self-esteem among adolescents undergoing an intervention targeting dietary and physical activity behaviors. Journal of Adolescent Health. 2007; 40(3): 245–51. Relevance. [PubMed] [Free Full text in PMC]

Ildiko V, Zsofia M, Janos M. et al. Activity-related changes of body fat and motor performance in obese seven-year-old boys. Journal of Physiological Anthropology. 2007; 26(3): 333–7. Setting. [PubMed]

Intense diet, behavior, and physical activity intervention effective for obese children. J Fam Pract. 2005;54:579. Design.

Israel A, Solotar LC, Zimand E. An Investigation of Two Parental Involvement Roles in the Treatment of Obese Children. Int J Eat Disord. 1990; 9(5): 557–564. Design.

Jago R, Jonker ML, Missaghian M, Baranowski T. Effect of 4 weeks of Pilates on the body composition of young girls. Preventive Medicine. 2006; 42(3): 177–80. Relevance. [PubMed]

Jago R, Baranowski T, Baranowski JC, Thompson D, Cullen KW, Watson K, Liu Y. Fit for Life Boy Scout badge: outcome evaluation of a troop and Internet intervention. SO: Preventive medicine. 2006; 42(3): 181–187. Prevention Trial.

Jelalian E, Mehlenbeck R, Lloyd-Richardson EE, Birmaher V, Wing RR. ‘Adventure therapy’ combined with cognitive-behavioral treatment for overweight adolescents. International Journal of Obesity. 2006; 30(1): 31–9. Design. [PubMed]

Jelalian E, Wember YM, Bungeroth H, Birmaher V. Practitioner review: bridging the gap between research and clinical practice in pediatric obesity. Journal of Child Psychology & Psychiatry & Allied Disciplines. 2007; 48(2): 115–27. Design.

Jiang JX, Xia XL, Greiner T, Lian GL, Rosenqvist U. A two year family based behaviour treatment for obese children. Archives of Disease in Childhood. 2005; 90(12): 1235–8. Setting. [PubMed] [Free Full text in PMC]

Johnston, Craig A, Steele, Ric G. Treatment of Pediatric Overweight: An Examination of Feasibility and Effectiveness in an Applied Clinical Setting. Journal of Pediatric Psychology. 2007; 32(1): 106–110. Design. [PubMed]

Jones RA, Okely AD, Collins CE. et al. The HIKCUPS trial: a multi-site randomized controlled trial of a combined physical activity skill-development and dietary modification program in overweight and obese children. BMC Public Health. 2007; 7: 15. Did not report relevant outcomes. [PubMed] [Free Full text in PMC]

Kelly AS, Steinberger J, Olson TP, Dengel DR. In the absence of weight loss, exercise training does not improve adipokines or oxidative stress in overweight children. Metabolism: Clinical & Experimental. 2007; 56(7): 1005–9. Design. [PubMed]

Kitzmann, Katherine M, Beech, Bettina M. Family-Based Interventions for Pediatric Obesity: Methodological and Conceptual Challenges From Family Psychology. Journal of Family Psychology. 2006; 20(2): 175–189. Design. [PubMed]

Klijn PH, van der Baan-Slootweg OH, van Stel HF. Aerobic exercise in adolescents with obesity: preliminary evaluation of a modular training program and the modified shuttle test. BMC Pediatrics. 2007; 7: 19. Design. [PubMed] [Free Full text in PMC]

Korsten-Reck U, Kromeyer-Hauschild K, Wolfarth B, Dickhuth HH, Berg A. Freiburg Intervention Trial for Obese Children (FITOC): results of a clinical observation study. International Journal of Obesity. 2005; 29(4): 356–61. Did not report relevant outcomes. [PubMed]

Lansky D, Vance MA. School-based intervention for adolescent obesity: analysis of treatment, randomly selected control, and self-selected control subjects. J Consult Clin Psychol. 1983; 51: 147–148. Design. [PubMed]

Lauer RM, Obarzanek E, Hunsberger SA, Van Horn L, Hartmuller VW, Barton BA, Stevens VJ, Kwiterovich PO Jr, Franklin FA Jr, Kimm SY, Lasser CL, Simons-Morton DG. Efficacy and safety of lowering dietary intake of total fat, saturated fat, and cholesterol in children with elevated LDL cholesterol: the Dietary Intervention Study in Children. Am J Clin Nutr. 2000; 72(5 Suppl): 1332S–1342S. Not a weight loss trial. [PubMed]

Levine MD, Ringham RM, Kalarchian MA, Wisniewski L, Marcus MD. Is family-based behavioral weight control appropriate for severe pediatric obesity? Int J Eat Disord. 2001; 30: 318–328. Design. [PubMed]

Lytle LA, Stone EJ, Nichaman MZ. et al. Changes in nutrient intakes of elementary school children following a school-based intervention: results from the CATCH Study. Prev Med. 1996; 25: 465–477. Relevance. [PubMed]

Maffeis C, Castellani M. Physical activity: an effective way to control weight in children? Nutrition Metabolism & Cardiovascular Diseases. 2007; 17(5): 394–408. Design.

Manios Y, Moschandreas J, Hatzis C, Kafatos A. Evaluation of a health and nutrition education program in primary school children of Crete over a three-year period. Prev Med. 1999; 28: 149–159. Relevance. [PubMed]

McLean N, Griffin S, Toney K, Hardeman W. Family involvement in weight control, weight maintenance and weight-loss interventions: a systematic review of randomised trials (Provisional record). SO: International Journal of Obesity. 2003; 27: 987–1005. Design.

Meyer AA, Kundt G, Lenschow U, Schuff-Werner P, Kienast W. Improvement of early vascular changes and cardiovascular risk factors in obese children after a six-month exercise program. Journal of the American College of Cardiology. 2006; 48(9): 1865–70. Did not report relevant outcomes. [PubMed]

Moore, Brie A. and O'Donohue, William T. 225–270. 2005. Design.

Moreno LA. Interventions to improve cardiovascular risk factors in obese children. Journal of Pediatric Gastroenterology & Nutrition. 2006; 43(4): 433–5. Design. [PubMed]

Mo-suwan L, Pongprapai S, Junjana C, Puetpaiboon A. Effects of a controlled trial of a school-based exercise program on the obesity indexes of preschool children. Am J Clin Nutr. 1998; 68: 1006–1011. Setting. [PubMed]

Muller MJ, Asbeck I, Mast M, Langnase K, Grund A. Prevention of obesity—more than an intention. Concept and first results of the Kiel Obesity Prevention Study (KOPS). Int J Obes Relat Metab Disord. 2001; 25(Suppl 1): S66–S74. Prevention trial. [PubMed]

Nemet D, Berger-Shemesh E, Wolach B, Eliakim A. A combined dietary-physical activity intervention affects bone strength in obese children and adolescents. International Journal of Sports Medicine. 2006; 27(8): 666–71. Design. [PubMed]

Nova E, Varela P, Lopez-Vidriero I, Toro O, Cena MJ, Casas J, Marcos A. A one-year follow-up study in anorexia nervosa. Dietary pattern and anthropometrical evolution. Eur J Clin Nutr. 2001; 55(7): 547–554. Population. [PubMed]

Nuutinen O, Knip M. Long-term weight control in obese children: persistence of treatment outcome and metabolic changes. Int J Obes Relat Metab Disord. 1992; 16: 279–287. Relevance. [PubMed]

O'Dea JA, Abraham S. Improving the body image, eating attitudes, and behaviors of young male and female adolescents: a new educational approach that focuses on self-esteem. Int J Eat Disord. 2000; 28: 43–57. Relevance. [PubMed]

Owens S, Gutin B, Allison J. et al. Effect of physical training on total and visceral fat in obese children. Med Sci Sports Exerc. 1999; 31: 143–148. Design. [PubMed]

Patrick K, Calfas KJ, Norman GJ. et al. Randomized controlled trial of a primary care and home-based intervention for physical activity and nutrition behaviors: PACE+ for adolescents. SO: Archives of pediatrics & adolescent medicine. 2006; 160: 128–136. Relevance.

Peterson KE, Fox MK. Addressing the epidemic of childhood obesity through school-based interventions: what has been done and where do we go from here? Journal of Law, Medicine & Ethics. 2007; 35(1): 113–30. Design.

Poland BD. Learning to ‘walk our talk’: the implications of sociological theory for research methodologies in health promotion. Can J Public Health. 1992; 83(Suppl 1): S31–S46. Relevance. [PubMed]

Ray R, Lim LH, Ling SL. Obesity in preschool children: an intervention programme in primary health care in Singapore. Ann Acad Med Singapore. 1994; 23(3): 335–341. Design. [PubMed]

Reinehr T, Kersting M, Alexy U, Andler W. Long-term follow-up of overweight children: after training, after a single consultation session, and without treatment. J Pediatr Gastroenterol Nutr. 2003; 37(1): 72–74. Quality. [PubMed]

Reinehr T, de SG, Toschke AM, Andler W. Long-term follow-up of cardiovascular disease risk factors in children after an obesity intervention. American Journal of Clinical Nutrition. 2006; 84(3): 490–6. Did not report relevant outcomes. [PubMed]

Reinehr T, de SG, Wabitsch M. Changes of cardiovascular risk factors in obese children effects of inpatient and outpatient intervention. Journal of Pediatric Gastroenterology & Nutrition. 2006; 43(4): 506–11. Information provided in another publication. [PubMed]

Resnicow K, Davis R, Rollnick S. Motivational interviewing for pediatric obesity: Conceptual issues and evidence review. Journal of the American Dietetic Association. 2006; 106(12): 2024–33. Design. [PubMed]

Resnicow K, Yaroch AL, Davis A. et al. GO GIRLS!: results from a nutrition and physical activity program for low-income, overweight African American adolescent females. Health Educ Behav. 2000; 27: 616–631. Design. [PubMed]

Reybrouck T, Vinckx J, Van den BG, Vanderschueren-Lodeweyckx M. Exercise therapy and hypocaloric diet in the treatment of obese children and adolescents. Acta Paediatr Scand. 1990; 79: 84–89. Did not report relevant outcomes. [PubMed]

Robbins LB, Gretebeck KA, Kazanis AS, Pender NJ. Girls on the move program to increase physical activity participation. Nursing Research. 2006 Jun;55(3): 206–16. Design. [PubMed]

Robinson TN. Reducing children's television viewing to prevent obesity: a randomized controlled trial. JAMA. 1999; 282: 1561–1567. Relevance. [PubMed]

Rocchini AP, Katch V, Anderson J. et al. Blood pressure in obese adolescents: effect of weight loss. Am J Clin Nutr. 1988; 82: 16–23. Design.

Rocchini AP, Katch V, Schork A, Kelch RP. Insulin and blood pressure during weight loss in obese adolescents. Hypertension. 1987; 10: 267–273. Design. [PubMed]

Rodearmel SJ, Wyatt HR, Barry MJ. et al. A family-based approach to preventing excessive weight gain. Obesity. 2006; 14(8): 1392–401. Design. [PubMed]

Rolland-Cachera MF, Thibault H, Souberbielle JC, Soulie D, Carbonel P, Deheeger M, Roinsol D, Longueville E, Bellisle F, Serog P. Massive obesity in adolescents: dietary interventions and behaviours associated with weight regain at 2 y follow-up. Int J Obes Relat Metab Disord. 2004; 28(4): 514–519. Design. [PubMed]

Rosenbaum M, Nonas C, Weil R. et al. School-based intervention acutely improves insulin sensitivity and decreases inflammatory markers and body fatness in junior high school students. Journal of Clinical Endocrinology & Metabolism. 2007; 92(2): 504–8. Design. [PubMed]

Salmon J, Booth ML, Phongsavan P, Murphy N, Timperio A. Promoting Physical Activity Participation among Children and Adolescents. Epidemiologic Reviews. 2007; 29: 144–59. Design. [PubMed]

Sasaki J, Shindo M, Tanaka H, Ando M, Arakawa K. A long-term aerobic exercise program decreases the obesity index and increases the high density lipoprotein cholesterol concentration in obese children. Int J Obes. 1987; 11: 339–345. Design. [PubMed]

Schwartz RP, Hamre R, Dietz WH. et al. Office-based motivational interviewing to prevent childhood obesity: a feasibility study. Archives of Pediatrics & Adolescent Medicine. 2007; 161(5): 495–501. Did not meet quality criteria. [PubMed]

Schwingshandl J, Sudi K, Eibl B, Wallner S, Borkenstein M. Effect of an individualised training programme during weight reduction on body composition: a randomised trial. Arch Dis Child. 1999; 81: 426–428. Design. [PubMed] [Free Full text in PMC]

Shaibi GQ, Cruz ML, Ball GD. et al. Effects of resistance training on insulin sensitivity in overweight Latino adolescent males. Medicine & Science in Sports & Exercise. 2006; 38(7): 1208–15. Design. [PubMed]

Sharma M. School-based interventions for childhood and adolescent obesity. Obesity Reviews. 2006; 7(3): 261–9. Design. [PubMed]

Sherry B. Food behaviors and other strategies to prevent and treat pediatric overweight. International Journal of Obesity. 2005; 29(Suppl 2): S116–26. Design. [PubMed]

Singh AS, Paw MJ, Brug J, van MW. Short-term effects of school-based weight gain prevention among adolescents. Archives of Pediatrics & Adolescent Medicine. 2007; 161(6): 565–71. Relevance. [PubMed]

Snethen JA, Broome ME, Cashin SE. Effective weight loss for overweight children: a meta-analysis of intervention studies. Journal of Pediatric Nursing. 2006; 21(1): 45–56. Design. [PubMed]

Sondike SB, Copperman N, Jacobson MS. Effects of a low-carbohydrate diet on weight loss and cardiovascular risk factor in overweight adolescents. J Pediatr. 2003; 142: 253–258. Design. [PubMed]

Sothern MS, Despinasse B, Brown R, Suskind RM, Udall JN Jr, Blecker U. Lipid profiles of obese children and adolescents before and after significant weight loss: differences according to sex. South Med J. 2000; 93: 278–282. Design. [PubMed]

Sothern MS, Hunter S, Suskind RM, Brown R, Udall JN Jr, Blecker U. Motivating the obese child to move: the role of structured exercise in pediatric weight management. South Med J. 1999; 92: 577–584. Design. [PubMed]

Sothern MS, Loftin JM, Udall JN. et al. Safety, feasibility, and efficacy of a resistance training program in preadolescent obese children. Am J Med Sci. 2000; 319: 370–375. Design. [PubMed]

Sothern MS, Schumacher H, von Almen TK, Carlisle LK, Udall JN. Committed to kids: an integrated, 4-level team approach to weight management in adolescents. J Am Diet Assoc. 2002; 102: S81–S85. Design. [PubMed]

Sothern, Udall JN Jr, Suskind RM, Vargas A, Blecker U. Weight loss and growth velocity in obese children after very low calorie diet, exercise, and behavior modification. Acta Paediatr. 2000; 89: 1036–1043. Design. [PubMed]

Southard DR, Southard BH. Promoting physical activity in children with MetaKenkoh. Clinical & Investigative Medicine - Medecine Clinique et Experimentale. 2006; 29(5): 293–7. Design. [PubMed]

Spieth LE, Harnish JD, Lenders CM. et al. A low-glycemic index diet in the treatment of pediatric obesity. Arch Pediatr Adolesc Med. 2000; 154: 947–951. Design. [PubMed]

Sung RY, Yu CW, Chang SK, Mo SW, Woo KS, Lam CW. Effects of dietary intervention and strength training on blood lipid level in obese children. Arch Dis Child. 2002; 86: 407–410. Design. [PubMed] [Free Full text in PMC]

Tanas R, Marcolongo R, Pedretti S, Gilli G. A family-based education program for obesity: a three-year study. BMC Pediatr. 2007; 7(1): 33. Design. [PubMed] [Free Full text in PMC]

Temple JL, Wrotniak BH, Paluch RA, Roemmich JN, Epstein LH. Relationship between sex of parent and child on weight loss and maintenance in a family-based obesity treatment program. International Journal of Obesity. 2006; 30(8): 1260–4. Design. [PubMed]

Vido L, Facchin P, Antonello I, Gobber D, Rigon F. Childhood obesity treatment: double blinded trial on dietary fibres (glucomannan) versus placebo. Padiatr Padol. 1993; 28: 133–136. Design. [PubMed]

Viner R, Nicholls D. Managing obesity in secondary care: a personal practice. Arch Dis Child. 2005; 90: 385–390. Design. [PubMed] [Free Full text in PMC]

Viski-Stalec N, Stalec J, Kati R, Podvorac D, Katovi D. The impact of dance-aerobics training on the morpho-motor status in female high-schoolers. Collegium Antropologicum. 2007; 31(1): 259–66. Setting. [PubMed]

Warschburger P, Fromme C, Petermann F, Wojtalla N, Oepen J. Conceptualisation and evaluation of a cognitive-behavioural training programme for children and adolescents with obesity. Int J Obes Relat Metab Disord. 2001; 25(Suppl 1): S93–S95. Design. [PubMed]

Williams CL, Strobino BA, Bollella M, Brotanek J. Cardiovascular risk reduction in preschool children: the “Healthy Start” project. J Am Coll Nutr. 2004; 23: 117–123. Relevance. [PubMed]

Williams CL, Strobino BA, Brotanek J. Weight control among obese adolescents: A pilot study. International Journal of Food Sciences & Nutrition. 2007; 58(3): 217–30. Design. [PubMed]

Young KM, Northern JJ, Lister KM, Drummond JA, O'Brien WH. A meta-analysis of family-behavioral weight-loss treatments for children. Clinical Psychology Review. 2007; 27(2): 240–9. Design. [PubMed]

Young-Hyman D, Schlundt DG, Herman L, De LF, Counts D. Evaluation of the insulin resistance syndrome in 5- to 10-year-old overweight/obese African-American children. Diabetes Care. 2001; 24: 1359–1364. Relevance. [PubMed]

Zemel MB, Richards J, Mathis S, Milstead A, Gebhardt L, Silva E. Dairy augmentation of total and central fat loss in obese subjects.[see comment]. International Journal of Obesity. 2005; 29(4): 391–7. Relevance. [PubMed]