Surgery for the treatment of obesity in children and adolescents
Abstract
Background
Child and adolescent overweight and obesity have increased globally and are associated with significant short‐ and long‐term health consequences.
Objectives
To assess the effects of surgery for treating obesity in childhood and adolescence.
Search methods
For this update, we searched Cochrane Central Register of Controlled Trials, MEDLINE, Latin American and Caribbean Health Science Information database (LILACS), World Health Organization International Clinical Trials Registry Platform (ICTRP)and ClinicalTrials.gov on 20 August 2021 (date of the last search for all databases). We did not apply language restrictions. We checked references of identified studies and systematic reviews.
Selection criteria
We selected randomised controlled trials (RCTs) of surgical interventions for treating obesity in children and adolescents (age < 18 years) with a minimum of six months of follow‐up. We excluded interventions that specifically dealt with the treatment of eating disorders or type 2 diabetes, or which included participants with a secondary or syndromic cause of obesity, or who were pregnant.
Data collection and analysis
We used standard methodological procedures expected by Cochrane. Two review authors independently extracted data and assessed the risk of bias using the Cochrane Risk of Bias 2.0 tool. Where necessary, we contacted authors for additional information.
Main results
With this update, we did not find any new RCTs. Therefore, this updated review still includes a single RCT (a total of 50 participants, 25 in both the intervention and comparator groups). The intervention focused on laparoscopic adjustable gastric banding surgery, which was compared to a control group receiving a multi‐component lifestyle programme. The participating population consisted of Australian adolescents (a higher proportion of girls than boys) aged 14 to 18 years, with a mean age of 16.5 and 16.6 years in the gastric banding and lifestyle groups, respectively. The trial was conducted in a private hospital, receiving funding from the gastric banding manufacturer. For most of the outcomes, we identified a high risk of bias, mainly due to bias due to missing outcome data.
Laparoscopic gastric banding surgery may reduce BMI by a mean difference (MD) of ‐11.40 kg/m2 (95% CI ‐13.22 to ‐9.58) and weight by ‐31.60 kg (95% CI ‐36.66 to ‐26.54) compared to a multi‐component lifestyle programme at two years follow‐up. The evidence is very uncertain due to serious imprecision and a high risk of bias. Adverse events were reported in 12/25 (48%) participants in the intervention group compared to 11/25 (44%) in the control group. A total of 28% of the adolescents undergoing gastric banding required revisional surgery. The evidence is very uncertain due to serious imprecision and a high risk of bias. At two years of follow‐up, laparoscopic gastric banding surgery may increase health‐related quality of life in the physical functioning scores by an MD of 16.30 (95% CI 4.90 to 27.70) and change in health scores by an MD of 0.82 (95% CI 0.18 to 1.46) compared to the lifestyle group. The evidence is very uncertain due to serious imprecision and a high risk of bias. No data were reported for all‐cause mortality, behaviour change, participants’ views of the intervention and socioeconomic effects.
Finally, we have identified three ongoing RCTs that are evaluating the efficacy and safety of metabolic and bariatric surgery in children and adolescents.
Authors' conclusions
Laparoscopic gastric banding led to greater body weight loss compared to a multi‐component lifestyle program in one small study with 50 participants. These results have very limited application, primarily due to more recent recommendations derived from observation studies to avoid the use of banding in youth due to long‐term reoperation rates. This systematic review update still highlights the lack of RCTs in this field. The authors are concerned that there may be ethical barriers to RTCs in this field, despite the lack of other effective therapies for severe obesity in children and adolescents and the significant morbidity and premature mortality caused by childhood obesity. Nevertheless, future studies, whether pre‐registered and planned non‐randomised or pragmatic randomised trials, should assess the impact of the surgical procedure and post‐operative care to minimise adverse events, including the need for post‐operative adjustments and revisional surgery. Long‐term follow‐up is also critical to comprehensively assess the impact of surgery as participants enter adulthood.
PICOs
Plain language summary
Surgery for the treatment of obesity in children and adolescents
Review question
How effective is bariatric surgery in safely reducing weight in obese children and adolescents?
Background
Across the world, more children and adolescents are developing overweight and obesity. As children and adolescents with overweight and obesity are more likely to suffer from health problems, more information is needed about how best to treat this problem.
Study characteristics
We did not find any new trials compared to the former version. Therefore, this work still includes one randomised controlled trial with a total of 50 participants (25 in both the intervention and comparator groups) and a follow‐up of two years. The surgery used was 'laparoscopic adjustable gastric banding' (gastric band placed around the entrance of the stomach by means of keyhole surgery). The control group received a program consisting of reduced energy intake (individualised diet plans ranging between 800 and 2000 kcal per day, depending on age and weight), increased activity (target of 10,000 steps per day) with a structured exercise schedule of at least 30 minutes a day and behavioural modification.
Key results
Australian adolescents (higher proportion of girls than boys) with an average age of 16.5 and 16.6 years in the gastric banding and control group participated. The study authors reported an average reduction in weight of 34.6 kg at two years, representing a change in body mass index units (kg/m²) of 12.7 for the gastric banding group; and an average reduction in weight of 3.0 kg representing a change in body mass index units of 1.3 for the control intervention. Side effects were reported in 12 of 25 (48%) participants in the intervention group and in 11 of 25 (44%) in the control group. A total of 28% of the adolescents undergoing gastric banding required a 'revisional procedure' (surgery because of complications from the gastric banding surgery). No data were reported for all‐cause mortality, behaviour change, participants’ views of the intervention and socioeconomic effects. At two years, the gastric banding participants performed better than the lifestyle participants in two of eight health‐related quality of life concepts as measured by the Child Health Questionnaire (physical functioning score (94 versus 78, community norm 95) and change in health score (4.4 versus 3.6, community norm 3.5).
Quality of the evidence
Our results are limited to two years of follow‐up and are based on just one small Australian study with high risk of bias, which was conducted in a private hospital and received funding from the gastric banding manufacturer. There remains insufficient RCT evidence to inform the recommendations of clinical guidelines. Current guidelines are reliant on the growing body of evidence from observational data.
Currentness of evidence
This evidence is up to date as of August 2021.
Authors' conclusions
Summary of findings
| Surgery compared with usual care(non‐surgical treatment) for children and adolescents with obesity | |||||
| Population: children and adolescents with obesity Settings: community, clinic Intervention: laparoscopic adjustable gastric banding surgery Comparison: multi‐component lifestyle programme | |||||
| Outcomes | Laparoscopic adjustable gastric banding surgery | Multi‐component lifestyle programme | Relative effect | No of participants | Certainty of the evidence |
|---|---|---|---|---|---|
| BMI loss (kg/m²) Follow‐up: two years | 12.7 lower (11.3 lower to 14.2 lower) | 1.3 lower (0.4 lower to 2.9 lower) | MD 11.40 less (13.22 less to 9.58 less) | 50 (1 RCT) | ⊕⊝⊝⊝ |
| Weight loss (kg) Follow‐up: two years | 34.6 lower (30.2 lower to 39.0 lower) | 3.0 lower (2.1 lower to 8.1 lower) | MD 31.60 less (36.66 less to 26.54 less) | 50 (1 RCT) | ⊕⊝⊝⊝ |
| Adverse events (revisional procedure) Follow‐up: two years | 280 per 1000 | 0 per 1000 | ‐ | 42 (1 RCT) | ⊕⊝⊝⊝ |
| Health‐related quality of life (CHQ (8 subscores); scale 0 to 100, where 0 indicates the worst possible health state and 100 the best possible health state)c Physical functioning Change in health Follow‐up: two years | Physical functioning: 94.4 (91.8 to 97.0) Change in health: 4.38 (4.1 to 4.7) | Physical functioning: 78.1 (68.7 to 87.5) Change in health: 3.56 (3.09 to 4.03) | Physical functioning: MD 16.30 higher (4.90 higher to 27.70 higher) Change in health: MD 0.82 higher (0.18 higher to 1.46 higher) | 42 (1 RCT) | ⊕⊝⊝⊝ |
| All‐cause mortality | Not reported | ||||
| Morbidity (metabolic syndrome)d Follow‐up: two years | 0 per 1000 | 222 per 1000 | RR 0.08 (0.00 to 1.47) | 42 (1 RCT) | ⊕⊝⊝⊝ |
| Socioeconomic effects | Not reported | ||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). BMI: body mass index; CHQ: child health questionnaire; CI: confidence interval; MD: mean difference; RCT: Randomised controlled trial; RR: risk ratio | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded one level for serious imprecision (one study with a small number of participants). | |||||
Background
From 1975 to 2016, the body mass index (BMI) in children and adolescents increased for girls by 0.30 kg/m2 (95% credibility interval 0.20 to,0.40) and for boys by 0.40 kg/m2 (95% credibility interval 0.30 to 0.50) per decade in most countries of the world, presenting a global public health crisis (NCD‐RisC 2017). Obesity prevalence increased in every country, with 50 million girls (95% credibility interval 24 to 89) and 74 million boys (95% credibility interval 39 to 125) affected globally. These BMI increases have also accelerated due to the COVID‐19 pandemic for children in the USA and European countries (Jarnig 2022; Lange 2021; Vogel 2022; Yang 2022).
The impact of severe obesity is also a major concern in the paediatric population (Pinhas‐Hamiel 2022). Although the International Obesity Task Force (IOTF) published an international definition for severe paediatric (morbid) obesity in 2012 (Cole 2012), severe obesity prevalence is often reported using country‐specific cut‐points, making international comparisons difficult. However, current data suggest that severe obesity affects 6.10% (standard error 0.70) of both boys and girls in the USA between two and 19 years of age (Fryar 2020), and between 1.00% (95% confidence interval 0.70 to 1.30) and 5.50% (95% confidence interval 4.90 to 6.10) of European children, with the prevalence being worse in boys compared to girls (Spinelli 2019).
Inequalities in overweight and obesity prevalence have also been documented. Generally, socioeconomically disadvantaged children in high‐income countries (Buoncristiano 2021; Knai 2012; Shrewsbury 2008), and children of higher socioeconomic status in low‐ and middle‐income countries (Buoncristiano 2021; Dinsa 2012), are at greater risk of developing overweight (Bridger 2021). However, this relationship may vary by population demographics (e.g. age, gender, ethnicity) and environment (e.g. country, urbanisation) (Wang 2012). The prevalence of obesity has been shown to vary by urbanisation, with higher prevalence in regions with lower population density (Ogden 2018), and data from several regions show substantial ethnic variation in child populations in Europe (de Wilde 2018; Pedersen 2016; Strugnell 2020), the USA (Min 2018), and New Zealand (Gibb 2019). The prevalence of severe obesity also varies by demographic characteristics, such as socioeconomic status, ethnicity, or urbanisation (de Wilde 2018; Ogden 2018; Pinhas‐Hamiel 2020). It may result in a greater risk of adverse cardio‐metabolic events and severe obesity in adulthood (Bendor 2020; Chung 2018; Michalsky 2015).
Description of the condition
Obesity is defined as an abnormal or excessive accumulation of body fat (WHO 2000). It is often measured by the BMI in adults. As children and adolescents have not completed linear growth, international or region‐specific and age‐ and sex‐adjusted BMI percentiles are used to define overweight and obesity in this age group (Cole 2012; de Onis 2007; Krebs 2007; Kuczmarski 2002; Styne 2017). Compared to their healthy‐weight peers, children and adolescents with obesity have a higher risk for comorbidities such as pre‐diabetes and diabetes, dyslipidaemia, high blood pressure, metabolic syndrome, non‐alcoholic fatty liver disease (NAFLD), asthma and sleep apnoea (Sharma 2019). The risk of developing comorbidities, such as high blood pressure, high triglyceride, elevated HbA1 values and low HDL (high‐density lipoprotein) cholesterol levels, has also been shown to increase as the severity of obesity increases (Skinner 2015). As in adulthood, obesity in childhood and adolescence can increase the risk of both short‐ and longer‐term health consequences. The early onset of obesity during childhood and adolescence can persist into adulthood (Ward 2017). The condition can also affect psychosocial well‐being, and young people with obesity are susceptible to reduced self‐esteem and quality of life due to weight stigma, which might lead to a vicious cycle (Puhl 2020).
In addition, paediatric obesity is associated with adverse health outcomes in later life, such as cardiovascular disease (Meyer 2021) and complications of type 2 diabetes mellitus (Bjerregaard 2018; Weihrauch‐Blüher 2019).
Description of the intervention
Given the serious implications associated with childhood and adolescent obesity, effective treatment is imperative. The primary aim of treatment (i.e. weight reduction or deceleration of weight gain) and the most suitable intervention approach varies and is dependent on the child's age and degree of excess weight, amongst other considerations. Multicomponent lifestyle intervention should be the starting point for all children and adolescents. If not effective, given the chronic relapsing nature of obesity, further treatment options (i.e. pharmacotherapy or surgery) should be considered to obtain sustainable, significant improvement in obesity and related comorbidities (Farpour‐Lambert 2015; Kelly 2018; Luca 2015; van der Baan‐Slootweg 2014).
Metabolic and bariatric surgery is an established treatment for adults with severe obesity (Phillips 2018; Rives‐Lange 2022). However, indications for bariatric surgery in youth differ between existing clinical guidelines, with less strict indications proposed by the American Academy of Paediatrics (AAP) (Armstrong 2019) and the American Society for Metabolic and Bariatric Surgery (ASMBS) (Pratt 2018) compared to the European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) (Nobili 2015). According to the experiences of adolescents who underwent bariatric surgery, this type of therapy should be offered to adolescents with severe obesity, even though it is described as a "rough" journey (Nordin 2018; van de Pas 2021). Therefore, in some adolescents with severe obesity, both adolescents and clinicians may consider surgery to be a pragmatic last solution to reduce BMI and associated comorbidities and improve health‐related quality of life.
Adverse effects of the intervention
Metabolic and bariatric surgery is a major surgical intervention with a risk of serious operative and perioperative complications and mortality. Depending on the type of surgery, interventions can cause early and late complications, such as nutritional, vitamin and mineral deficiencies, internal or incisional hernia, reflux disease, cholelithiasis, wound infections, or ulceration (Arterburn 2020). The restrictive or malabsorptive nature of some forms of metabolic and bariatric surgery is an additional consideration in growing children, with guidelines largely agreeing that eligible candidates must be adolescents with severe obesity who have reached or nearly reached physical maturity (DAG 2019; Nobili 2015). In contrast, current American Guidelines (e.g. ASMBS, AAP) advise not to rely only on physical maturity or the adolescent's age as indication standards (Armstrong 2019; Pratt 2018). Additional considerations in adolescents may include developmental issues around the ability to consent and the need for family support. Severe obesity can be a comorbidity in some children with learning disabilities, who may have limited ability to both consent and adhere to dietary regimes required for safe surgery and postoperative care. Given this, consideration of patient and parent compliance is an important issue surrounding surgery. Contraindications to surgery include pregnancy or breastfeeding, medically correctable causes of obesity, substance abuse, and a disability that may prevent adherence to postoperative management (DAG 2019; Nobili 2015).
How the intervention might work
Metabolic and bariatric surgery changes the neurohormonal signalling that triggers hunger, satiety, and metabolism through changes in gastrointestinal hormones involved in appetite regulation, such as peptide YY (PYY), glucagon‐like peptide 1 (GLP‐1) or ghrelin (Nguyen 2017; Xu 2021). This leads to decreased caloric intake, changes in food choices and improved metabolism. A number of different surgical procedures exist that are commonly used in children.
-
Laparoscopic sleeve gastrectomy involves removing two‐thirds of the stomach, leaving a 'banana' shaped stomach. This results in decreased ghrelin (less hunger) and increased GLP‐1 and PYY levels (improving insulin resistance and leading to early satiety). Weight loss and improvement in type 2 diabetes, sleep apnoea, cardiovascular risk and fatty liver disease occur rapidly the following surgery.
-
Roux‐en‐Y gastric bypass involves dividing the proximal stomach, leaving an egg‐sized pouch, and then connecting an end of the small intestine called the roux limb, which bypasses the stomach and proximal small intestine. The resultant change in gastrointestinal hormonal secretion, bile salts and microbiome results in decreased food intake, improved metabolism, and significant weight loss. It also results in significant and rapid improvement in type 2 diabetes mellitus, sleep apnoea, cardiovascular risk and fatty liver disease.
-
Laparoscopic adjustable banding involves the placement of an adjustable band just below the gastro‐oesophageal junction to create a small gastric pouch that restricts food intake (Nguyen 2017). It does not produce the beneficial neurohormonal changes that the other procedures do and has not shown significant long‐term efficacy in most patients. The only randomised study done in this field in children, however, was carried out using the adjustable gastric band.
For adolescents undergoing bariatric surgery in the USA, the most common procedures include vertical sleeve gastrectomy (70.6%) and laparoscopic Roux‐en‐Y gastric bypass (27.7%) procedures (Griggs 2018).
Compared to bariatric surgery in children and adolescents with severe obesity, there is more evidence from trials in adults with severe obesity. Bariatric surgery in adults has been shown to improve quality of life and life expectancy by three to nine years (Carlsson 2020; Schauer 2015).
Why it is important to do this review
Since the last update of the review in 2015 (Ells 2015), the prevalence of severe obesity in children and adolescents has increased, and the application of surgical techniques has changed (Griggs 2018). For example, the use of the technique evaluated in the only randomised controlled trial (RCT) (O'Brien 2010) included in the Ells 2015 review (laparoscopic adjustable gastric banding) decreased from 419 in 2009 to less than 10 in 2014 (Griggs 2018). Furthermore, observational data showed high failure rates and increased needs for reintervention (Peña 2017); as such, the ASMBS guideline no longer recommends adjustable gastric banding (Pratt 2018). Furthermore, observational data for other procedures, such as gastric bypass in adolescents, now have longer‐term follow‐up data available (three years (Shoar 2017), or five years (Inge 2019)). Therefore, we updated the search for published and ongoing RCTs to summarise the recent trial evidence for bariatric and metabolic surgery in children and adolescents, to derive further implications for research and clinical practice.
Objectives
To assess the effects of surgery for treating obesity in children and adolescents.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials (RCTs).
Types of participants
We included study groups consisting of obese participants, with a mean age of less than 18 years at the commencement of the intervention. We excluded pregnant females and the critically ill, as well as children with obesity due to a secondary or syndromic cause (e.g. Prader‐Willi syndrome).
Types of interventions
We investigated the following comparisons of intervention versus control or comparator.
-
Surgery compared with placebo.
-
Surgery compared with usual care (non‐surgical treatment).
-
Surgery + other therapy compared with placebo + other therapy.
-
Surgery + other therapy compared with usual care (non‐surgical treatment) + other therapy compared.
Concomitant therapies were the same in the intervention and comparator groups.
Types of outcome measures
Primary outcomes
-
BMI and weight loss.
-
Adverse events.
Secondary outcomes
-
Health‐related quality of life and self‐esteem.
-
All‐cause mortality.
-
Morbidity (changes in disease status).
-
Measures of body fat distribution.
-
Behaviour change.
-
Participants' views of the intervention.
-
Socioeconomic effects.
Method and timing of outcome measurement
-
BMI: defined as the weight (kg) divided by height (m) squared, and weight loss defined as a loss in weight in kg from baseline, measured at baseline, 6, 12 and 24 months.
-
Adverse events: defined as an adverse outcome that occurs during or after the intervention but is not necessarily caused by it, and measured at baseline, 6, 12 and 24 months.
-
Health‐related quality of life and self‐esteem: evaluated by a validated instrument such as the Paediatric Quality of Life Inventory and measured at baseline, 6, 12 and 24 months.
-
All‐cause mortality: defined as any death that occurred during or after the intervention and measured at baseline, 6, 12 and 24 months.
-
Morbidity: defined as illness or harm associated with the intervention or the condition and measured at baseline, 6, 12 and 24 months.
-
Measures of body fat distribution: defined by the use of validated tools, such as dual energy X‐ray absorptiometry (DXA), waist circumference, skinfold thickness, waist to hip ratio, dual x‐ray absorptiometry or bioelectrical impedance analysis, and measured at baseline, 6, 12 and 24 months.
-
Behaviour change: defined as validated measures of diet or physical activity (Bryant 2014), and measured at baseline, 6, 12 and 24 months.
-
Participants' views of the intervention: defined as documented accounts from participant feedback and measured at baseline, 6, 12 and 24 months.
-
Socioeconomic effects: defined as a validated measure of socioeconomic status, such as parental income or educational status, and measured at baseline, 6, 12 and 24 months.
Summary of findings table
We present a summary of findings table reporting the following outcomes, listed according to priority.
-
BMI and weight loss.
-
Adverse events.
-
Health‐related quality of life.
-
All‐cause mortality.
-
Morbidity.
-
Socioeconomic effects.
Search methods for identification of studies
Electronic searches
For this update, we searched the following sources from 1 January 2015 to 20 August 2021 (date of last search for all databases) and placed no restrictions on the language of publication.
-
Cochrane Central Register of Controlled Trials (CENTRAL) via Cochrane Register of Studies Online (CRSO).
-
MEDLINE (Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid MEDLINE(R) 1946 to Present).
-
LILACS (Latin American and Caribbean Health Science Information database).
-
World Health Organization International Clinical Trials Registry Platform (ICTRP) (www.who.int/trialsearch).
-
ClinicalTrials.gov (www.clinicaltrials.gov).
We did not include Embase in our search, as RCTs indexed in Embase are now prospectively added to CENTRAL via a highly sensitive screening process (Cochrane 2022). For detailed search strategies see Appendix 1.
Searching other resources
We tried to identify other potentially eligible trials or ancillary publications by searching the reference lists of retrieved included trials, systematic reviews, meta‐analyses and health technology assessment reports.
Data collection and analysis
Selection of studies
To determine the studies to be assessed further, two review authors (GT, JB) independently scanned the abstract, title, or both, of every record retrieved by the searches. We investigated all potentially relevant articles as full text. Where differences in opinion existed, we resolved them by discussion and consensus with a third review author (LJE). If it was not possible to resolve the disagreement, we added the article to those 'Studies awaiting classification' and contacted study authors for clarification. We present an adapted PRISMA flow diagram reporting the process of study selection (Page 2021).
Data extraction and management
For studies that fulfilled inclusion criteria, two review authors (LJE, KR, or EM) independently extracted key participant and intervention characteristics and reported data on efficacy outcomes and adverse events using a standard data extraction form supplied by the Cochrane Metabolic and Endocrine Disorders (CMED) Group. Disagreements were to be resolved by discussion, or if required by a third review author (GA). Details of study characteristics are provided in Characteristics of included studies, and of participants' characteristics in Table 1. Details of intervention characteristics are shown in Appendix 2, and baseline characteristics are provided in Appendix 3 and Appendix 4. Characteristics on endpoints are shown in Appendix 5, Appendix 6, Appendix 7, Appendix 8, Appendix 9.
| Intervention(s) and comparator(s) | Sample sizea | Screened/eligible | Randomised | Safety | ITT/analysed (N) | Finishing study | Randomised finishing study | Follow‐upb | |
|---|---|---|---|---|---|---|---|---|---|
| I: gastric banding procedure + lifestyle advice | The study was powered assuming that, using an ITT analysis, more than 60% of participants of the gastric banding group would achieve an excess weight loss of more than 50% at 2 years and that less than 10% of the lifestyle group would achieve this weight lossc. Using these expected proportions, study authors required 17 participants in each the study group to provide an 80% power and a 2‐sided P value of 0.05. On the basis of a possible loss of 30% after randomisation, 50 adolescents were recruited. | 163/84 | 25 | 25 | 25/25d | 24 | 96 | 24 months | |
| C: lifestyle programme | 25 | 25 | 25/25d | 18 | 72 | ||||
| total: | 50 | 50 | 50/50 | 42 | 84 | ||||
| Grand total | All interventions | 25 | 24 | ||||||
| All c omparators | 25 | 18 | |||||||
| All interventions and c omparators | 50 | 42 | |||||||
aAccording to power calculation in study publication or report
bDuration of intervention or follow‐up, or both, under randomised conditions until end of study
cActual numbers were 84% in the intervention and 12% in the comparator group
dPrimary analysis only (weight change data)
C: comparator; I: intervention; ITT: intention‐to‐treat
We provide information about potentially‐relevant ongoing studies including trial identifier in the Characteristics of ongoing studies table and in Appendix 5 'Matrix of study endpoints (publications and trial documents)'. We tried to find the protocol for each included study, either in databases of ongoing trials, in publications of study designs, or both. We sent an email request to the author of the included study to enquire whether further unpublished data relating to the study were available, whether the trial was ongoing, and whether they were involved with any new studies in this area (Appendix 10).
Dealing with duplicate publications and companion papers
In the event of duplicate publications and companion papers of a primary study, we tried to maximise yield of information by simultaneous evaluation of all available data. In case of doubt, we gave priority to the publication reporting the longest follow‐up associated with our primary or secondary outcomes.
Assessment of risk of bias in included studies
Two review authors (GT, JB) independently assessed the risk of bias for the results of the main outcomes (those included in the summary of findings table, see below) in each study using a recently developed revision of the Cochrane Risk of bias tool (the Risk of Bias (RoB) 2 tool) (Higgins 2022a; Sterne 2019). We resolved disagreements by consensus or by consulting a third review author (LJE). If adequate information was unavailable from the publications, trial protocols, clinical study reports or other sources, we contacted the study authors for more details to request missing data on risk of bias items. We assessed the risk of bias according to the following domains, focusing on the effect of assignment to the intervention at baseline:
-
the randomisation process;
-
deviations from intended interventions;
-
missing outcome data;
-
measurement of the outcome;
-
selection of the reported results.
Answers to signalling questions and supporting information collectively lead to a domain‐level judgement of either 'low risk', 'some concerns', or 'high risk' of bias. These domain‐level judgements informed an overall risk of bias judgement for a single result in the form of (a) 'low risk' if we judged all domains to be 'low risk'; (b) 'some concerns' if we judged all domains to have 'some concerns'; or (c) 'high risk' if we judged one or more domain to be 'high risk', or if we judged four domains to have 'some concerns'. We provided a quote from the study report together with a justification for our judgement in the risk of bias table. We summarised the risk of bias judgements across different studies for each of the domains listed. We sourced trial registries, protocols and analysis plans for the assessment of selective reporting. Where information on the risk of bias related to unpublished data or correspondence with a trialist, we noted this in the risk of bias table.
When considering treatment effects, we took into account the risk of bias for the studies that contributed to that outcome. We constructed summary assessments of the risk of bias for each important outcome (across domains), within and across studies (Higgins 2022a; Sterne 2019).
We used the RoB 2 Excel tool to manage the data supporting the answers to the signalling questions and risk of bias judgements (available at www.riskofbias.info/). The data are available in the Open Science Framework (osf.io/7tydm/).
Measures of treatment effect
For dichotomous outcomes we calculated odds ratio (OR) or risk ratio (RR) and corresponding 95% confidence interval (CI). For continuous outcomes we calculated the mean difference (MD) and corresponding 95% CI.
Unit of analysis issues
We planned to take into account the level at which randomisation occurred, such as cross‐over trials, cluster‐randomised trials and multiple observations for the same outcome.
Dealing with missing data
If feasible, we obtained relevant missing data from authors. We evaluated important numerical data such as the number of screened, eligible, and randomised participants, as well as intention‐to‐treat (ITT), as‐treated and per‐protocol (PP) populations. We investigated attrition rates (e.g. dropouts, losses to follow‐up, withdrawals), and we critically appraised issues concerning missing data and imputation methods (e.g. last observation carried forward (LOCF)).
If standard deviations for outcomes had not been reported, we would have imputed these values by assuming the standard deviation of the missing outcome to be the average of the standard deviations from those studies where this information was reported. If more than one study had been available, we would have investigated the impact of this imputation on the point estimate using a sensitivity analysis.
Assessment of heterogeneity
If more than one paper had been identified and there was substantial clinical, methodological or statistical heterogeneity, we would not have reported study results as meta‐analytically pooled effect estimates. Heterogeneity would have been identified by visual inspection of the forest plots and by using a standard Chi² test with a significance level of α = 0.1, in view of the low power of this test. If more than one study had been identified, we would have examined heterogeneity using the I² statistic (which quantifies inconsistency across studies) to assess the impact of heterogeneity on the meta‐analysis (Higgins 2002; Higgins 2003); where an I² statistic of 75% or more indicates a considerable level of inconsistency (Higgins 2022b). If heterogeneity had been found, we would have attempted to determine potential reasons for it by examining individual study and subgroup characteristics.
We expected the following characteristics to introduce clinical heterogeneity.
-
Differences in the age of the study population.
-
Differences in the study population demographics.
-
Differences in the types of surgery performed.
-
Differences in BMI at baseline.
Assessment of reporting biases
If we included 10 studies or more for a given outcome, we planned to use funnel plots to assess small study effects. Due to there being several potential explanations for funnel plot asymmetry, we planned to interpret results carefully (Sterne 2011).
Data synthesis
Unless there was good evidence for homogeneous effects across studies, we planned to primarily summarise low‐risk of bias data by means of a random‐effects model (Wood 2008). We planned to interpret random‐effects meta‐analyses with due consideration of the whole distribution of effects, ideally by presenting a prediction interval (Higgins 2009). A prediction interval specifies a predicted range for the true treatment effect in an individual study (Riley 2011). In addition, if statistical analyses had been possible, these would have been performed according to the statistical guidelines provided by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022b)).
Subgroup analysis and investigation of heterogeneity
We planned to carry out the following subgroup analyses, and wanted to investigate interaction.
-
Length of follow‐up.
-
Impact and nature of maintenance periods.
-
The impact of comparator or control: whether concomitant therapy or no treatment (true control).
-
The impact of population demographics.
Sensitivity analysis
We planned to perform sensitivity analyses in order to explore the influence of the following factors (when applicable) on effect size by restricting the analysis to the following.
-
Published studies.
-
Taking into account risk of bias, as specified in the Assessment of risk of bias in included studies section.
-
Very long or large studies to establish how much these studies dominate the results.
-
Studies using the following filters: diagnostic criteria, language of publication, source of funding (industry versus other), country.
We also planned to test the robustness of the results by repeating the analysis using different measures of effect size (RR, OR etc) and different statistical models (fixed‐effect and random‐effects models).
Summary of findings and assessment of the certainty of the evidence
We presented the overall certainty of the evidence for each outcome specified below according to the GRADE approach, which takes into account issues related to internal validity (overall risk of bias, inconsistency, imprecision, publication bias) and external validity (such as directness of results). Two review authors (LJE, KR, or GT) independently rated the certainty of the evidence for each outcome. We resolved any differences in assessment by discussion or by consultation with a third review author (EM).
We presented a summary of the evidence in a summary of findings table. This provides key information about the best estimate of the magnitude of effect, in relative terms and as absolute differences for each relevant comparison of alternative management strategies; the numbers of participants and studies addressing each important outcome; and a rating of overall confidence in effect estimates for each outcome. We created the summary of findings table using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2022), using GRADEpro software (GRADEpro GDT).
We justified all decisions to downgrade the certainty of the evidence by using informative footnotes, and we used GRADE guidelines for informative statements (Santesso 2016; Santesso 2020).
We planned to create summary of findings tables for the following comparisons and outcomes.
-
Comparison
-
Surgery compared with usual care (non‐surgical treatment)
-
-
Outcomes
-
BMI, weight loss, health‐related quality of life, all‐cause mortality, morbidity, socioeconomic effects (follow‐up: two years)
-
Results
Description of studies
For a detailed description of studies, see Characteristics of included studies, Characteristics of excluded studies, and Characteristics of ongoing studies.
Results of the search
The update search on 20 August 2021 identified 1034 records (881 from database searches and 153 from trial registry searches). From these, we identified 20 full‐text publications and protocols for further examination, of which three trials met the inclusion criteria for ongoing studies. We excluded the other 17 studies because they did not meet the inclusion criteria or were not relevant to the review question (see Figure 1 for the amended PRISMA flow diagram). After screening the full texts of the selected publications, we did not identify any new finished or published trials that met our inclusion criteria. Therefore, this updated review only includes the one RCT that had already been included in the previous review by Ells 2015.
Ongoing studies
NCT01172899 reports the recruitment of 14‐ to 16‐year‐old adolescents with obesity to assess the efficacy of gastric banding in a Dutch population, with completion anticipated in December 2022. In this trial, 30 participants were randomly allocated to a lifestyle intervention plus gastric banding and 30 participants to a lifestyle intervention.
ACTRN12609001004257 reports the recruitment of 44 of 50 planned 12‐ to 17‐year‐old participants with obesity to assess the efficacy of a Bioenterics Intragastric Balloon (BIB) in an Australian population. This study started recruitment in 2009 but does not report the end date. For this trial, interim results were published as a conference abstract. No significant differences between the groups were shown for percentage change in body weight and BMI z‐score. The interim analysis included five participants in the BIB group; no adverse event was reported during insertion, the intervention phase, or at removal (ACTRN12609001004257).
NCT02378259 is the most recently registered trial and aims to recruit 13‐ to 16‐year‐olds to assess the efficacy of bariatric surgery in Sweden. Twenty‐five participants were randomised to bariatric surgery and 25 participants to calorie restriction by a meal replacement product. As stated in the published protocol, 23 of the 25 participants were treated by Laparoscopic Roux‐en‐Y Gastric Bypass and two by sleeve gastrectomy. The primary completion was anticipated for June 2022, with study completion in June 2034.
The last version of this review (Ells 2015) identified another ongoing trial from France (clinicaltrials.gov/ct2/show/NCT01700738). For this trial, results were only published for one of the two planned groups (Pourcher 2015). Therefore, we did not consider this trial as an ongoing RCT for this version of the review.
Included studies
A detailed description of the characteristics of included studies is presented elsewhere (see Characteristics of included studies and appendices). The following is a succinct overview.
Source of data
With this update, we could not include any additional RCTs. The only published RCT that meets the inclusion criteria (O'Brien 2010) was included in the previous version of this review (Ells 2015). Furthermore, we list three ongoing trials that were identified from trial registry searches (ACTRN12609001004257; NCT01172899; NCT02378259). These were also identified as ongoing studies by Ells 2015. We could not include these ongoing trials as outcome data are not yet available. Details of these studies are provided in the Characteristics of ongoing studies studies.
Comparisons
O'Brien 2010 compared laparoscopic adjustable gastric banding surgery to a multi‐component lifestyle modification program, consisting of individual calorie reduction diet plans, increased physical activity through pedometer targets, structured exercise schedules, advice to reduce sedentary activity and support through consultation with a health care practitioner every six weeks.
Overview of study populations
A total of 50 participants were included in the trial; 25 participants were randomised to intervention and 25 to control groups. Twenty‐four (98%) participants finished the study in the intervention compared to 18 (72%) participants in the control group.
Study design
The included study was a randomised parallel group superiority trial. Given the nature of the intervention under investigation, it was not possible to blind to participants or personnel delivering the interventions. However, outcome assessors were also unblinded. The duration of the intervention was two years, conducted between August 2006 and September 2008. The study was not terminated early.
Settings
The study was undertaken in a specialist weight management clinic either in the community or the Royal Children’s Hospital, Melbourne, with surgery occurring at a private hospital.
Participants
The participating population consisted of Australian adolescents aged 14 to 18 years, with a mean age of 16.5 and 16.6 years in the banding and lifestyle groups, respectively. All participants demonstrated substantial physiological maturity with secondary sexual characteristics and most had also completed bone growth. This study contained a higher proportion of girls than boys in each arm of the intervention: 36% of the banding group were males and 28% of the lifestyle group were males. No further demographic information was reported. The mean BMI at baseline was 42.30 (standard deviation (SD) 6.10) kg/m² in the banding group compared to 40.40 (SD 3.10) kg/m² in the lifestyle group. Entry criteria are outlined in the Characteristics of included studies table. Major exclusion criteria were intellectual disability and syndromic obesity.
Diagnosis
Participants in the O'Brien 2010 study were required to have a BMI greater than 35 and identifiable medical complications such as metabolic syndrome, physical limitation (such as an inability to play a sport), or psycho‐social difficulties such as low self‐esteem.
Interventions
This study employed a two‐month run‐in program, which all participants undertook prior to randomisation. The program involved the implementation of best practice guidance on healthy eating and physical activity. The surgical intervention consisted of the gastric band placement followed by detailed guidance on post‐operative eating and activity.
Outcomes
The one included study assessed 50 participants and reported data for all primary and some secondary endpoints. This study did not report all‐cause mortality, behaviour change, participants' views of the intervention, socioeconomic effects and costs. For a summary of all outcomes assessed in the study, see Appendix 5.
Excluded studies
Fifteen articles had to be excluded from the update search after careful evaluation of the full publication (Asaad 2018; Bjork 2021; de Oliveira 2018; Dewberry 2019; Ebell 2017; Miller 2017; Misra 2020; Ospanov 2019; Ponce 2015; Schiavon 2020; Shah 2021; Sullivan 2017; Trastulli 2017; Varma 2019; Zitsman 2020; see Figure 1).
Risk of bias in included studies
For details on the risk of bias of the included study, see the Characteristics of included studies table and each of the analyses. We investigated the risk of bias for all reported outcome measures that met our inclusion criteria.
Randomisation process
We judged the included study to have a low risk of bias for randomisation. We identified a mismatch in information between the study publication and data in the trial register, but judged both methods to be sufficiently random. In addition, we did not identify substantial differences in baseline characteristics between the intervention and comparator groups.
Deviations from the intended interventions
O'Brien 2010 explicitly stated that the study was not blinded. However, blinding of the participants and personnel delivering the intervention was not possible given the nature of this study. The authors stated that ITT analyses were performed for the primary outcome (weight change). Therefore, we rated the risk of bias due to deviations from the intended interventions as low for BMI and weight loss. We rated the risk of bias in this domain as high for the other outcomes as the authors applied complete case analyses.
Missing outcome data
O'Brien 2010 reported on withdrawals and losses to follow‐up, with one loss to follow‐up in the banding intervention and two losses to follow‐up and five withdrawals in the lifestyle intervention (due to family problems and being unsatisfied with progress). ITT analysis was performed for the primary outcome (weight change) only, and all secondary outcomes (health‐related quality of life outcomes) were assessed by analysis of completers as the study was only powered to detect changes in the primary outcome measure. Consequently, we considered that there was a high risk of bias due to missing outcome data for objectively measured outcomes (such as the primary outcome of weight loss) because of a substantial loss to follow‐up and differences in dropouts. There was a higher dropout rate in the lifestyle group, with a considerable number of participants justifying it by unhappiness with progress, and there were no details of ITT analyses (e.g. the method of imputation). Additionally, we considered the risk of bias for subjective measures (health‐related quality of life) to be high as no ITT analysis was conducted, and disparate attrition rates probably influenced this outcome measure. In addition, the O'Brien 2010 study lacked analyses correcting for bias or sensitivity analyses.
Measurement of the outcome
The impact of not blinding posed a high risk for the subjectively reported health‐related quality of life measures. Therefore, we judged the impact of no blinding on bias due to measurement of the outcome to be high risk for the subjective outcome health‐related quality of life. For the objective outcomes, we judged that the measurement of outcomes was standardised, so we did not assume that the measurement of outcomes differed between the groups. Therefore, we judged the risk of bias due to the measurement of the outcome to be low for objectively measured outcomes.
Selection of the reported result
No detailed analysis plan was published in the trial register, but comparing the study publication and protocol information in the trial register revealed some differences in health‐related quality of life. Therefore, we judged there to be a high risk of bias due to the selection of this reported result; however, there was little information for other results, so we rated this domain to have 'some concerns'.
Effects of interventions
See: Summary of findings 1 Summary of findings
Baseline characteristics
For details of baseline characteristics, see Appendix 3 and Appendix 4.
Gastric banding program versus lifestyle program
The included study examined the effects of laparoscopic gastric banding surgery compared to a form of lifestyle programme (O'Brien 2010). This study measured weight change as the primary outcome.
Primary outcomes
BMI and weight loss
The study authors reported, for a total of 50 participants providing outcome data, a mean reduction in weight of 34.60 kg (95% CI 30.20 to 39.00) at two years, representing a change in BMI units of 12.70 (95% CI 11.30 to 14.20) for the surgery intervention; and a mean reduction in weight of 3.00 kg (95% CI 2.10 to 8.10) representing a change in 1.30 BMI units (95% CI 0.40 to 2.90) for the lifestyle intervention. Laparoscopic gastric banding surgery may reduce BMI by a mean difference (MD) of ‐11.40 kg/m2 (95% CI ‐13.22 to ‐9.58) (Analysis 1.1) and weight by ‐31.60 kg (95% CI ‐36.66 to ‐26.54) (Analysis 1.2), compared to a multi‐component lifestyle programme at two years follow‐up for a total of 50 participants. The certainty of the evidence is very uncertain due to serious imprecision and a high risk of bias.
Adverse events
The gastric banding placement occurred without any complications during the perioperative period or within 30 days. The mean length of hospital stay was 26 hours (range from 23 hours to 32 hours). A total of 28% of the 42 participants providing outcome data required a revisional procedure (Analysis 1.3).
Adverse events were reported in both groups, with 13 events reported in 12 participants in the surgery intervention compared to 18 events reported in 11 participants in the lifestyle group; adverse events in the surgery group included six proximal gastric enlargements, two needlestick injuries to tubing, one cholecystectomy, one hospital admission for depression, one loss to follow‐up and two unplanned pregnancies. Adverse events in the lifestyle group included one hospital admission for depression and intracranial hypertension, one cholecystectomy, seven loss to follow‐up and two unplanned pregnancies.
Over the two‐year study period, the surgical group (n = 25) had a mean of 20 visits with a physician (range 10 to 31) per participant and a mean of 9.5 adjustments made to the volume of saline in the band (range 5 to 18) per participant. In the non‐surgical group (n = 25), adolescents visited the adolescent physician, study dietitian, study nurse practitioner, or other physicians a mean of 16 (range 7 to 31) times. There was also a mean of five telephone consultations per participant and each participant had six sessions with a personal trainer.
Secondary outcomes
Health‐related quality of life
Health‐related quality of life was assessed by the Child Health Questionnaire (CHQ), a family of generic quality of life instruments that have been designed and validated for children 5 to 18 years of age. Parents and children (ages 10 to 18 years) may self‐administer the CHQ after instructions from the administrator. The CHQ measures 14 unique physical and psychosocial concepts. The parent form is available in two lengths: 50 or 28 items. Scores can be analysed separately, the CHQ profile scores, or combined to derive an overall physical and psychosocial score, the CHQ summary scores.
Score interpretation: the range on subscales and the overall scale is 0 to 100, where 0 indicates the worst possible health state and 100 the best possible health state. A normative sample was not available for comparison of paediatric patient‐reported health‐related quality of life. Poor health‐related quality of life has been defined as two standard deviations below the mean of the normative sample or a physical functioning or psychosocial health summary score less than 30.
Eight of the subscores of the CHQ are shown in Appendix 11. The subscores for behavioural, emotional, and physical limitations are not shown because these items did not differ from community values at entry into the study and were not different within or between groups over the two‐year follow‐up period. No statistically significant differences existed in any measures between groups at the commencement of the study. For the 42 participants providing outcome data at two years follow‐up, laparoscopic gastric banding surgery may increase health‐related quality of life in the physical functioning scores by an MD of 16.30 (95% CI 4.90 to 27.70) (Analysis 1.4) and change in health scores by an MD of 0.82 (95% CI 0.18 to 1.46) (Analysis 1.5) compared to the lifestyle group. The certainty of the evidence is very uncertain due to serious imprecision and a high risk of bias.
Morbidity
Morbidity was associated with metabolic syndrome, a weak surrogate endpoint for illness or harm associated with the intervention or the condition itself. At study entry, 36% of the participants in the gastric banding group and 40% in the lifestyle group were diagnosed with metabolic syndrome. For the 42 participants providing outcome data at two years follow‐up, none of the 24 study completers (0%) in the gastric banding group had metabolic syndrome compared to four of 18 completers (22%) in the lifestyle group who still had metabolic syndrome. Therefore, laparoscopic gastric banding surgery may decrease the risk for morbidity by a RR of 0.08 (95% CI 0.00 to 1.47) (Analysis 1.6) compared to a lifestyle programme at two years follow‐up. The certainty of the evidence is very uncertain due to serious imprecision, a high risk of bias and indirectness.
Measures of body fat distribution
Waist circumference was reduced by 28.20 cm in the gastric banding group and by 3.50 cm in the lifestyle group in the 42 participants providing outcome data at two years (MD ‐24.70 cm, 95% CI ‐33.10 to ‐16.30; P < 0.001; Analysis 1.7).
Other outcomes
All‐cause mortality, behaviour change, self‐esteem, participants' views of the intervention and socioeconomic effects were either not investigated or not reported in the included study.
Discussion
Summary of main results
This review reports the findings from one RCT (50 participants). The intervention focused on laparoscopic adjustable gastric banding surgery, which was compared to a control group receiving a multi‐component lifestyle program. The study authors were unable to blind their participants, personnel and outcome assessors, which may have resulted in a high risk of bias in deviations from the intended interventions and measurement of the outcome. At two years' follow‐up, laparoscopic gastric banding surgery may reduce BMI, weight, and the risk of morbidity, and may improve health‐related quality of life in the physical functioning scores and change in health scores compared to a multi‐component lifestyle programme. For participants who underwent bariatric surgery, the performance of revisional procedures was necessary in a substantial number of cases. The certainty of evidence is very uncertain due to serious imprecision, a high risk of bias, or indirectness. In addition, there are three ongoing trials that evaluate the efficacy and safety of metabolic and bariatric surgery in children and adolescents.
Overall completeness and applicability of evidence
We did not find further RCTs to provide additional data to that identified by the previous update (Ells 2015). Whilst the included trial by O'Brien 2010 reported on weight, health‐related quality of life and adverse events, further data on the participants' socioeconomic status and ethnic origin may have enhanced the wider applicability of the findings. Eating small meals slowly is central to avoiding problems after the gastric banding procedure. This was repeatedly stressed during the O'Brien 2010 trial. For adolescents, additional education and supervision of eating may help reduce the need for revision surgery. Recruitment methods were used to minimise bias towards one or other treatment but may have drawn on a subset of the community attracted by the availability of free treatment. The O'Brien 2010 trial was also powered to measure differences in weight outcomes rather than differences in other health measures or adverse events. Adolescents and parents must understand the importance of carefully adhering to recommended eating behaviours and seeking early consultation if symptoms of reflux, heartburn, or vomiting occur. As importantly, they should be in a setting where they can maintain contact with health professionals who understand the care process. The authors state their uncertainty as to whether the study population accurately reflects the general adolescent population living with obesity since it may have attracted a subset of the community amenable to the availability of free treatment. In addition, the only included trial provides evidence for a surgical technique that is no longer recommended by the ASMBS due to high complication (10.50%) and reintervention (14.70%) rates and a lack of safety data (Pratt 2018), and which is currently not yet approved by the US Food and Drug Administration (FDA) for people < 18 years of age (Pratt 2018). Therefore, gastric bands should be considered for metabolic and bariatric surgery with caution.
We list three ongoing trials (ACTRN12609001004257; NCT01172899; NCT02378259). One ongoing trial is examining the effect of Bioenterics Intragastric Balloon (BIB) insertion over six months compared to usual care (a 10‐week multidisciplinary lifestyle modification programme) in an Australian adolescent population (ACTRN12609001004257). We contacted the study authors twice but did not receive a reply. In the absence of further details, we were unable to include the trial. This finding was reflected in a recent repeated meta‐research analysis, suggesting that the nonpublication of RCTs still occurs (Speich 2022). Another Swedish ongoing study, started in August 2014, planned to have completed data collection by June 2022 for the primary outcome BMI. This study aims to examine the impact of bariatric surgery (Roux‐en‐Y‐gastric bypass (23 participants) or sleeve gastrectomy (two participants)) compared to an intensive lifestyle treatment (25 participants) in 13‐ to 16‐year‐old adolescents (NCT02378259). Although another ongoing trial in the Netherlands (NCT01172899) is also comparing laparoscopic gastric banding (as assessed by O'Brien 2010), it has yet to reach completion so no further narrative or quantitative comparisons could be made.
In line with the previous update (Ells 2015), and the other reviews in this series examining interventions for the treatment of child and adolescent obesity (Al‐Khudairy 2017; Colquitt 2016; Ells 2018; Loveman 2015; Mead 2016; Mead 2017), the study design was limited to RCTs to provide the least‐biased estimate of effect size (Rosen 2006).
Additionally, important aspects ‐ mainly regarding safety and the long‐term outcome ‐ have not been addressed (or not sufficiently addressed) by O'Brien and colleagues. Post‐surgical interventions, such as nutritional supplementation, are recommended to reduce adverse events after metabolic and bariatric surgery. Observational data of 85 adolescents who underwent Roux‐en‐Y gastric bypass surgery showed that those not adhering to the recommendations for nutritional supplementation had a higher chance of nutritional deficiencies (Henfridsson 2019). This is supported by evidence from studies in adults undergoing metabolic and bariatric surgery (Ha 2021). Strategies to improve adherence to post‐surgery management need further development to prevent future deficiencies (Anvari 2021).
For the efficacy and safety of weight loss interventions, the knowledge of long‐term outcomes is crucial. A systematic review summarised evidence from non‐randomized trials on metabolic and bariatric surgery in adolescents (Ruiz‐Cota 2019). For BMI or weight, most studies reported weight regain until the latest reported follow‐up at 12 years, remission rates for comorbidities, i.e., dyslipidaemia, musculoskeletal problems, hypertension, and type 2 diabetes mellitus, were 75%‐85% and rates for iron deficiency and anaemia were high (up to 70% and 50%). Other possible complications were reported insufficiently (Ruiz‐Cota 2019). A further recent publication reporting ten‐year outcomes of sleeve gastrectomy in over 2500 children and adolescents demonstrated improvements in cardiovascular outcomes (e.g., hypertension, dyslipidaemia), significant weight loss and a low number of adverse events, such as nausea or vomiting, or neuropathy (Alqahtani 2021).
Moreover, it is important not to ignore the significant morbidity of (severe) obesity in childhood and the risk of inadequately treating children with severe obesity. Type 2 diabetes mellitus in children is associated with a much more rapid progression of beta‐cell loss, cardiovascular disease, renal impairment, retinopathy, and neuropathy than in adults or children with type 1 diabetes (Barrett 2020). Furthermore, elevated BMI in adolescents has been shown to significantly increase the risk of cardiovascular mortality in adults in an American study, as well as all‐cause mortality in a Swedish study (Lindberg 2020; Ryder 2020; Twig 2016). Obesity during childhood is also associated with a higher cancer mortality rate in adults (Nuotio 2022). In addition, adult studies report significantly improved life expectancy in people with severe obesity treated with metabolic and bariatric surgery, especially those with type 2 diabetes mellitus, versus those who continue to pursue lifestyle interventions (Carlsson 2020; Schauer 2015).
Quality of the evidence
With this update, we assessed the risk of bias with the Cochrane Risk of Bias 2.0 tool. Whilst the included study was well conducted and provides much‐needed evidence in this field, further studies are required to strengthen the evidence base. Although blinding would have reduced the risk of bias, we acknowledge the logistical challenges of blinding in such studies. The effect of blinding on treatment outcomes is still not clear, which is why blinding continues to be recommended in RCTs (Moustgaard 2020). It would also have been useful if O'Brien 2010 had additionally reported the exact baseline‐adjusted group difference in change scores.
Potential biases in the review process
As only one published study was comprehensively assessed in this review, no potential biases in the review process arose.
Agreements and disagreements with other studies or reviews
The findings from this review are limited due to the lack of other RCTs in adolescents. However, metabolic and bariatric surgery is supported by multiple systematic reviews of adolescent obesity surgery that also considered evidence from observational studies (Qi 2017; Selvendran 2018; Trooboff 2019). The degree of weight loss and improvements to health‐related quality of life reported in the O'Brien 2010 study are concordant with those reported in recent reviews of metabolic and bariatric surgery in young people (Qi 2017). However, there is significantly better weight loss, improved quality of life and improvement in comorbidities seen with Roux‐en‐Y gastric bypass and sleeve gastrectomy than with adjustable gastric banding in the adolescent systematic reviews (Qi 2017; Selvendran 2018; Trooboff 2019). However, these reviews also show the inadequacy of alternative therapies in successfully treating severe obesity, including pharmacotherapy, intensive lifestyle, and exercise programmes (Selvendran 2018).
To our knowledge, no other systematic review has been published that included any RCTs other than the O'Brien 2010 trial. However, several systematic reviews have been published that (additionally) summarised evidence from observational studies. As mentioned before, due to missing evidence from RCTs, evidence from non‐randomised trials should be considered given the rising prevalence of severe paediatric obesity and surgical procedures in children and adolescents with obesity since the publication of the O'Brien 2010 study (Griggs 2018).
We excluded trials that had participants with severe obesity due to a secondary or syndromic cause, such as Prader‐Willi Syndrome. The risks and benefits of metabolic and bariatric surgery must be carefully assessed for people with such conditions, given their other comorbidities (Gantz 2022). In addition, we excluded pregnant adolescents, which is in line with clinical guidelines that do not recommend metabolic and bariatric surgery during pregnancy or for at least 12 months after pregnancy (Pratt 2018).
As we did not find further evidence from RCTs in adolescents, indirect evidence from studies in (young) adults warrants discussion. These observational studies investigate differences in the effects of metabolic and bariatric surgery between adolescents and adults. For example, Lennerz 2014 did not show a significant difference between adolescents and younger adults for short‐term BMI reduction following various types of bariatric and metabolic surgery (gastric banding, gastric bypass, sleeve gastrectomy, gastric balloon, biliopancreatic diversion and gastric pacemaker). This finding was also demonstrated by Benedix 2017, who compared adolescents (mean age 19.50 ± 1.50 years) and middle‐aged adults (mean age 44.2 ± 11 years) from Germany for two years after laparoscopic sleeve gastrectomy, and the Teen‐LABS study in the USA (Inge 2019), which compared middle‐aged adults with adolescents five years after surgery. Further, remission of diabetes and hypertension was more likely for adolescents than adults (Inge 2019). Similar results between the two age groups were shown for the remission of hypertriglyceridemia, low HDL cholesterol, and the need for any post‐surgery intra‐abdominal operation (Inge 2019).
O'Brien 2010 also reported improvements in cardiovascular morbidity in the gastric banding group. Improvements in cardiovascular morbidity following metabolic and bariatric surgery in adolescents have been reported in multiple studies of people undergoing Roux‐en‐Y gastric bypass, sleeve gastrectomy and adjustable gastric banding (Qi 2017). The improved well‐being measure reported in the O'Brien 2010 study aligns with observational evidence that also demonstrates the beneficial effects of stapled bariatric procedures on psychosocial outcomes and quality of life in adolescents (Trooboff 2019).
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 1: BMI loss
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 2: Weight loss
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 3: Adverse events
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 4: Health‐related quality of life (physical functioning)
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 5: Health‐related quality of life (change in health)
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 6: Morbidity (changes in disease status)
Comparison 1: Surgery compared with usual care (non‐surgery intervention), Outcome 7: Measures of body fat distribution
| Surgery compared with usual care(non‐surgical treatment) for children and adolescents with obesity | |||||
| Population: children and adolescents with obesity Settings: community, clinic Intervention: laparoscopic adjustable gastric banding surgery Comparison: multi‐component lifestyle programme | |||||
| Outcomes | Laparoscopic adjustable gastric banding surgery | Multi‐component lifestyle programme | Relative effect | No of participants | Certainty of the evidence |
|---|---|---|---|---|---|
| BMI loss (kg/m²) Follow‐up: two years | 12.7 lower (11.3 lower to 14.2 lower) | 1.3 lower (0.4 lower to 2.9 lower) | MD 11.40 less (13.22 less to 9.58 less) | 50 (1 RCT) | ⊕⊝⊝⊝ |
| Weight loss (kg) Follow‐up: two years | 34.6 lower (30.2 lower to 39.0 lower) | 3.0 lower (2.1 lower to 8.1 lower) | MD 31.60 less (36.66 less to 26.54 less) | 50 (1 RCT) | ⊕⊝⊝⊝ |
| Adverse events (revisional procedure) Follow‐up: two years | 280 per 1000 | 0 per 1000 | ‐ | 42 (1 RCT) | ⊕⊝⊝⊝ |
| Health‐related quality of life (CHQ (8 subscores); scale 0 to 100, where 0 indicates the worst possible health state and 100 the best possible health state)c Physical functioning Change in health Follow‐up: two years | Physical functioning: 94.4 (91.8 to 97.0) Change in health: 4.38 (4.1 to 4.7) | Physical functioning: 78.1 (68.7 to 87.5) Change in health: 3.56 (3.09 to 4.03) | Physical functioning: MD 16.30 higher (4.90 higher to 27.70 higher) Change in health: MD 0.82 higher (0.18 higher to 1.46 higher) | 42 (1 RCT) | ⊕⊝⊝⊝ |
| All‐cause mortality | Not reported | ||||
| Morbidity (metabolic syndrome)d Follow‐up: two years | 0 per 1000 | 222 per 1000 | RR 0.08 (0.00 to 1.47) | 42 (1 RCT) | ⊕⊝⊝⊝ |
| Socioeconomic effects | Not reported | ||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). BMI: body mass index; CHQ: child health questionnaire; CI: confidence interval; MD: mean difference; RCT: Randomised controlled trial; RR: risk ratio | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded one level for serious imprecision (one study with a small number of participants). | |||||
| Intervention(s) and comparator(s) | Sample sizea | Screened/eligible | Randomised | Safety | ITT/analysed (N) | Finishing study | Randomised finishing study | Follow‐upb | |
|---|---|---|---|---|---|---|---|---|---|
| I: gastric banding procedure + lifestyle advice | The study was powered assuming that, using an ITT analysis, more than 60% of participants of the gastric banding group would achieve an excess weight loss of more than 50% at 2 years and that less than 10% of the lifestyle group would achieve this weight lossc. Using these expected proportions, study authors required 17 participants in each the study group to provide an 80% power and a 2‐sided P value of 0.05. On the basis of a possible loss of 30% after randomisation, 50 adolescents were recruited. | 163/84 | 25 | 25 | 25/25d | 24 | 96 | 24 months | |
| C: lifestyle programme | 25 | 25 | 25/25d | 18 | 72 | ||||
| total: | 50 | 50 | 50/50 | 42 | 84 | ||||
| Grand total | All interventions | 25 | 24 | ||||||
| All c omparators | 25 | 18 | |||||||
| All interventions and c omparators | 50 | 42 | |||||||
| aAccording to power calculation in study publication or report | |||||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 BMI loss Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 1.2 Weight loss Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 1.3 Adverse events Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Totals not selected | |
| 1.4 Health‐related quality of life (physical functioning) Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 1.5 Health‐related quality of life (change in health) Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 1.6 Morbidity (changes in disease status) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Totals not selected | |
| 1.7 Measures of body fat distribution Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| O'Brien 2010 |  | |  | | | | ||||||
| 1.1 Quote from publication: "Randomization was performed using a computer‐derived random allocation sequence to allow orderly admission into both programs. There was no stratification or blocking," Information specified in study register (ACTRN12605000160639): "Subjects were block randomized into 3 unequal blocks. Each block contained equal numbers in both treatment arms. The sequence within blocks was determined by the staff member in control of concealment by drawing the allocation out of a hat and the number in each blocking group was known only to this staff member" Comment: randomization sufficiently random 1.2 Quote from author (via email): "The trial coordinator went over the trial details with the prospective participant (and parent), had the informed consent for the trial signed, phoned the allocation centre, received the allocation and informed the participant immediately" Information specified in study register (ACTRN12605000160639): "Intervention recorded and stored in numbered, sealed opaque envelopes organized and maintained by staff member not involved in patient care or scheduling and opened in sequence as randomized." Comment: allocation was concealed 1.3 Comment: no baseline differences detected. | ||||||||||||
| 2.1/ 2.2 Quote from publication: "The study was not blinded" Quote from information specified in study register (ACTRN12605000160639): "Open (masking not used)" 2.6 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data." Comment: ITT analysis was used to estimate the effect of assignment to intervention. | ||||||||||||
| 3.1 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data. Longitudinal data analysis was performed using SAS software to estimate weight measures allowing for missing data" Comment: 24/25 in bariatric surgery arm at FU and 18/25 in lifestyle arm with 4 patients dropping out 'because unhappy with progress' 3.2 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data. Longitudinal data analysis was performed using SAS software to estimate weight measures allowing for missing data" Comment: 24/25 in bariatric surgery arm at FU and 18/25 in lifestyle arm with 4 patients dropping out 'because unhappy with progress' 3.3/ 3.4 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data. Longitudinal data analysis was performed using SAS software to estimate weight measures allowing for missing data" Comment: 24/25 in bariatric surgery arm at FU and 18/25 in lifestyle arm with 4 patients dropping out 'because unhappy with progress'" | ||||||||||||
| 4.1 Quote from publication: "The primary end point of the study was whether participants could lose 50% excess weight. We used the Centers for Disease Control and Prevention (CDC) growth charts and defined excess weight as the weight above the 85th percentile of BMI for age and sex. We calculated the total weight loss (kg), percentage of total weight lost, percentage of excess weight lost, change in BMI, and BMI z score. BMI z scores are the number of standard deviations that a patient's BMI deviates from the reference mean BMI for that age group. Anthropometric measures included neck, waist, and hip circumference. Adjustment for change in height was made when appropriate." Comment: appropriate assessment of the outcome. 4.2 Quote from publication: "The primary end point of the study was whether participants could lose 50% excess weight. We used the Centers for Disease Control and Prevention (CDC) growth charts and defined excess weight as the weight above the 85th percentile of BMI for age and sex. We calculated the total weight loss (kg), percentage of total weight lost, percentage of excess weight lost, change in BMI, and BMI z score. BMI z scores are the number of standard deviations that a patient's BMI deviates from the reference mean BMI for that age group. Anthropometric measures included neck, waist, and hip circumference. Adjustment for change in height was made when appropriate." Comment: probably no difference in objective outcome assessment between groups 4.3 Quote from publication: "the study was not blinded." Comment: Outcome assessors were aware of the intervention. 4.4 Comment: probably no difference in objective outcome assessment between groups | ||||||||||||
| 5.1 Comment: no changes in analysis plans detected. 5.2 Comment: all assessed outcomes were reported 5.3 Comment: all assessed outcomes were reported | ||||||||||||
| High RoB for at least one RoB domain | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| O'Brien 2010 | | | | | | | ||||||
| 1.1 Quote from publication: "Randomization was performed using a computer‐derived random allocation sequence to allow orderly admission into both programs. There was no stratification or blocking," Information specified in study register (ACTRN12605000160639): "Subjects were block randomized into 3 unequal blocks. Each block contained equal numbers in both treatment arms. The sequence within blocks was determined by the staff member in control of concealment by drawing the allocation out of a hat and the number in each blocking group was known only to this staff member" Comment: randomization sufficiently random 1.2 Quote from author (via email): "The trial coordinator went over the trial details with the prospective participant (and parent), had the informed consent for the trial signed, phoned the allocation centre, received the allocation and informed the participant immediately" Information specified in study register (ACTRN12605000160639): "Intervention recorded and stored in numbered, sealed opaque envelopes organized and maintained by staff member not involved in patient care or scheduling and opened in sequence as randomized." Comment: allocation was concealed 1.3 Comment: no baseline differences detected. | ||||||||||||
| 2.1/ 2.2 Quote from publication: "The study was not blinded" Quote from information specified in study register (ACTRN12605000160639): "Open (masking not used)" 2.6 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data." Comment: ITT analysis was used to estimate the effect of assignment to intervention. | ||||||||||||
| 3.1 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data. Longitudinal data analysis was performed using SAS software to estimate weight measures allowing for missing data" Comment: 24/25 in bariatric surgery arm at FU and 18/25 in lifestyle arm with 4 patients dropping out 'because unhappy with progress' 3.2 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data. Longitudinal data analysis was performed using SAS software to estimate weight measures allowing for missing data" Comment: 24/25 in bariatric surgery arm at FU and 18/25 in lifestyle arm with 4 patients dropping out 'because unhappy with progress' 3.3/ 3.4 Quote from publication: "Data Analysis. We analyzed the weight change data according to the patient's randomly assigned program (intention‐to‐treat analysis) and used completer's analysis for the health and quality of life data. Longitudinal data analysis was performed using SAS software to estimate weight measures allowing for missing data" Comment: 24/25 in bariatric surgery arm at FU and 18/25 in lifestyle arm with 4 patients dropping out 'because unhappy with progress'" | ||||||||||||
| 4.1 Quote from publication: "The primary end point of the study was whether participants could lose 50% excess weight. We used the Centers for Disease Control and Prevention (CDC) growth charts and defined excess weight as the weight above the 85th percentile of BMI for age and sex. We calculated the total weight loss (kg), percentage of total weight lost, percentage of excess weight lost, change in BMI, and BMI z score. BMI z scores are the number of standard deviations that a patient's BMI deviates from the reference mean BMI for that age group. Anthropometric measures included neck, waist, and hip circumference. Adjustment for change in height was made when appropriate." Comment: appropriate assessment of the outcome. 4.2 Quote from publication: "The primary end point of the study was whether participants could lose 50% excess weight. We used the Centers for Disease Control and Prevention (CDC) growth charts and defined excess weight as the weight above the 85th percentile of BMI for age and sex. We calculated the total weight loss (kg), percentage of total weight lost, percentage of excess weight lost, change in BMI, and BMI z score. BMI z scores are the number of standard deviations that a patient's BMI deviates from the reference mean BMI for that age group. Anthropometric measures included neck, waist, and hip circumference. Adjustment for change in height was made when appropriate." Comment: probably no difference in objective outcome assessment between groups 4.3 Quote from publication: "the study was not blinded." Comment: Outcome assessors were aware of the intervention. 4.4 Comment: probably no difference in objective outcome assessment between groups | ||||||||||||
| 5.1 Comment: no changes in analysis plans detected. 5.2 Comment: all assessed outcomes were reported 5.3 Comment: all assessed outcomes were reported | ||||||||||||
| High RoB for at least one RoB domain | ||||||||||||
