NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Centre for Public Health Excellence at NICE (UK); National Collaborating Centre for Primary Care (UK). Obesity: The Prevention, Identification, Assessment and Management of Overweight and Obesity in Adults and Children [Internet]. London: National Institute for Health and Clinical Excellence (UK); 2006 Dec. (NICE Clinical Guidelines, No. 43.)

Cover of Obesity: The Prevention, Identification, Assessment and Management of Overweight and Obesity in Adults and Children

Obesity: The Prevention, Identification, Assessment and Management of Overweight and Obesity in Adults and Children [Internet].

Show details

15Management of obesity in clinical settings

The marketing authorisation for sibutramine has been suspended. See details.

15.1. General introduction to clinical management

This section presents the reviews used by the Guidance Development Group (GDG) to inform recommendations. The reviews were conducted to address the identified key clinical questions (see Appendix 2 for details).

We have presented the reviews for children and adults separately. Each review consists of a narrative summary (with a quantitative summary of appropriate information) with the associated evidence statements. Evidence tables and excluded studies can be found in Appendices 13 and 14 for children and Appendices 15 and 16 for adults, respectively.

15.2. Children

See also the adult reviews (section 15.3) for details of evidence on clinical settings, brief interventions, barriers and attitudes to management.

15.2.1. Factors to be considered in the clinical assessment of children and adolescents who are overweight or obese

[The aim of an initial assessment is to identify individuals who are at increased risk and who would benefit from intervention. This initial assessment should follow the classification of the degree of overweight or obesity as recommended by the GDG based on the earlier evidence reviews.

Therefore, the factors to be assessed at the initial presentation should be based on two evidence bases: one on the common comorbidities, and one on the effectiveness of weight loss in people with comorbidities and their expected health gain.

Further assessment(s) should aim to determine any determinants of energy imbalance.]

15.2.1.1. Evidence statements

(Table 15.1)

Table 15.1. Evidence statements and grading.

Table 15.1

Evidence statements and grading.

15.2.1.2. Evidence review on factors to be assessed in children and adolescents

In September 2005, the National Guideline Clearinghouse synthesised the recommendations on the assessment and treatment of obesity and overweight in children from six published guidelines.1

The different scopes, target populations, intended users, and practices covered can be seen in Appendices 13 and 14 (evidence tables and excluded studies, respectively).

The authors of the synthesis identified areas of agreement between the included guidelines surrounding assessment. They concluded that:

‘In addition to BMI, AAP, RNAO, SIGN, and Singapore MOH advise health care providers to evaluate other risk factors for obesity. Parental obesity is identified by these 4 groups as being a strong predictor that an obese child will become an obese adult. The evidence for other risk factors, with the exception of certain childhood syndromes (e.g., Praeder-Willi syndrome) or diseases (e.g., hypothyroidism) is less clear. In addition, these 4 groups cite physical inactivity and increased television viewing as probable risk factors for overweight and obesity. AHA, while not identifying specific risk factors to look for, notes that the identification of risk for overweight before adolescence is encouraged so that health habits can be improved at a stage of increased parental influence and control. USPSTF does not provide recommendations regarding other risk factors.1

The individual guideline recommendations on assessment are given in Table 15.2.

Table 15.2. Recommendations on assessment.

Table 15.2

Recommendations on assessment.

Based on evidence for the common comorbidities, the National Health and Medical Research Council (Australia) (NHMRC)2 recommended that an initial assessment in a child should include measurement of the waist circumference (although no cut-offs are defined) and blood pressure. However, the guideline stressed that these tests need to be performed and results interpreted in the context of greater degree of obesity, increasing age, history of comorbidities and a family history of metabolic disease related to obesity.

These and other assessments should be made as follows:

  • Waist circumference: Waist circumference appears to be associated with cardiovascular risk profile. However, cut-points for children and adolescents have not yet been established.
  • Blood pressure: Use an appropriate size cuff. The child may be hypertensive if systolic or diastolic blood pressure is greater than or equal to the 95th percentile for age, sex and height. The child is normotensive if blood pressure is below the 90th percentile.
  • Psychosocial distress: Clinicians should determine whether the child is being teased and bullied about their weight. If distress and low self-esteem cannot be managed by simple interventions, consider referral for expert counselling.
  • Fasting lipid profile: Should be considered in obese children and adolescents, particularly those who have a family history of cardiovascular risk factors.
  • Fasting insulin and glucose: Should be considered in obese children or adolescents, particularly those with a family history of type 2 diabetes, those with acanthosis nigricans, and those from certain ethnic backgrounds.
  • Liver function tests: May be necessary in greater degrees of obesity, followed by hepatic ultrasound if transaminases are elevated.
  • Endocrinology tests: Are not required unless there is other evidence of endocrine disease or short stature. Many overweight children and adolescents have cutaneous striae – do not investigate for Cushing’s disease unless the patient is hypertensive with growth delay, and obesity is of recent onset.

15.2.2. Energy imbalance in children and adolescents

The NHMRC guidelines2 identified a number of risk factors associated with the development of obesity in children. These were:

  • Genes: There is a significant genetic predisposition to obesity. Parental obesity is a risk factor for future, if not present, obesity.
  • Television viewing: The (mainly American) data on television viewing indicate a positive correlation between hours of viewing and overweight. The correlation is stronger in older children and adolescents and clearer at low or high (less than 2 or greater than 5) hours of viewing per day. Studies on other forms of small-screen entertainment are awaited.
  • Energy expenditure: (i) Just as measurement of obesity in children is limited by the lack of immediate morbidity and mortality data, so is the ability to develop physical activity – management guidelines that are based on evidence related to positive health outcomes. (ii) Measuring physical activity in the clinical setting is difficult. Until cheap, small and robust motion monitors become available, it will depend largely on self-reporting. (iii) Reduced physical activity energy expenditure may play a role in weight gain over time.
  • Artificial infant feeding: The majority of cohort studies support the finding that breastfeeding plays a small protective role against subsequent overweight.
  • Dietary fat, dietary carbohydrate and eating patterns: (i) The evidence that dietary fat intake is a significant risk for obesity in children and adolescents is minimal. (ii) There is no clear evidence that any particular dietary composition influences overweight or obesity in children or adolescents. (iii) There is minimal evidence that carbohydrate intake influences body weight in children and adolescents. (iv) Parents influence food choices and other eating behaviours in their children. Disordered eating in a parent may be associated with excess body weight in the child.
  • Single-gene defects and obesity syndromes: (i) There are a number of single-gene abnormalities in which obesity is the predominant feature. (iii) There are a number of rare congenital syndromes that have obesity as a component and in which intellectual impairment is a common feature.
  • Ethnicity: International and local data suggest that certain ethnic backgrounds entail a higher predisposition to obesity.
  • Early adiposity rebound: (i) There is evidence from population studies that early adiposity rebound is associated with higher adolescent and adult body mass index (BMI). (ii) No matter how overweight is defined in the individual study, there is a significant association between higher birth weight and higher weights in childhood. (iii) Additional risk is conferred by an average, rather than tall, birth length and by parental overweight. (iv) Small-for-gestational-age babies who exhibit catch-up growth are at risk of obesity in childhood.
  • Single-child, single-parent, rural versus urban, and socio-economic status: Evidence statements related specifically to Australia only.
  • Endocrine disease: Clinical observation confirms an association between obesity and a number of endocrine disorders. Height–growth failure is the feature that should alert the clinician.
  • Central nervous system pathology: Hypothalamic damage can result in a severe form of obesity in children and adolescents.
  • Acute lymphatic leukaemia therapy: There is general agreement that, at the end of therapy for acute lymphatic leukaemia, there is a higher prevalence of obesity among participants than at the commencement of therapy and that obesity persists.
  • Medications: The role of pharmacological agents in causing weight gain in children and adolescents has not been extensively studied.

Recommendations about what to assess included:

  • weight history of the child and first-degree relatives
  • medical history
  • family, school, and social environments (Australia data only)
  • ethnicity (Australia data only)
  • eating and physical activity behaviour of child and parents.

A recent cohort study3 aimed to identify risk factors in early life (up to 3 years of age) for obesity in children in the UK. Participants were 8234 children in a cohort aged 7 years and a sub-sample of 909 children with data on additional early growth-related risk factors. Data from 5493 children were available for the multivariate analysis. Risk factors in entire cohort were as follows: intrauterine and perinatal factors, infant feeding and weaning practice, family characteristics and demographics, lifestyle in early childhood, sedentary behaviour, and dietary patterns. From these, birth weight, parental obesity, sleep duration, and television viewing remained independently connected with the risk of obesity in the final model. Also, a further four factors were significant in the children in focus subsample: size in early life, weight gain in infancy, catch-up growth, and early adiposity or BMI rebound.3

15.2.3. Lifestyle interventions in weight management and other outcomes in children and adolescents

15.2.3.1. Evidence statements

(Table 15.3)

Table 15.3. Evidence statements and grading.

Table 15.3

Evidence statements and grading.

15.2.3.2. Evidence review on lifestyle interventions

There is scarce evidence on which are the required components or treatment phases for childhood obesity. However, throughout the literature it appears that a multidisciplinary approach is most commonly advocated. Programmes normally include one or several of the following components:

This review is partially based on the Cochrane review published in 2003.4 The aim of the Cochrane review was to systematically review the effects of a range of lifestyle interventions designed to treat obesity in childhood. The Cochrane review included interventions such as: diet, physical activity, and/or behaviour therapy with or without the participation of family members.

This section reviews evidence on the assessment of the effectiveness of lifestyle interventions such as dietary change, physical activity, behaviour therapy, or some combination of these components.

The inclusion criteria for this evidence review initially followed the steps of the Cochrane review, although in light of the scarcity of randomised evidence in regard to the treatment of obesity in adolescents and children, a revised inclusion criteria was defined which encompasses the following designs of studies.

Types of study
  • Randomized randomised controlled trials (RCTs)
  • Controlled clinical trials
  • Controlled before-and-after studies

Only studies with a minimum duration of 6 months or above (including follow-up) and published after 1985 were included, and also RCTs with a primary aim other than the treatment of childhood obesity. Studies based in a setting other than clinical were not included in this review.

Update searches have been undertaken, however no studies were included as none added further details or contradicted any of the recommendations.

Types of participant
  • Participants aged under 18 years at the start of the study, and exceptionally studies where the age cut-off was above 18 years and where the majority of the participants were below 18 years or presented age stratification.
Types of outcome
  • Primary outcomes to be measured (not self-reported) estimates of overweight in per cent, BMI, weight in kilograms, per cent weight loss, percentage of ideal body weight, BMI z-score and others.
  • Secondary outcomes to be behaviour change, participants’ views, measures of self-esteem, health status, well-being and quality of life.

Finally several other studies were retrieved by cross-referencing with the review of systematic reviews,5 the Australian Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents (NHMRC),2 The Scottish Intercollegiate Guidelines Network (SIGN),6 and other reviews.7–17

We considered the following as a working definition of what is behavioural treatment§ when reviewing and assessing the interventions described in the published papers.

  • Behavioural treatment draws on the principles of learning theory (stimulus–behaviour contingencies or behaviour–reward contingencies)
  • Assessment consists of identifying and specifying problem behaviours and the circumstances in which they are elicited (both antecedents and consequences).
  • Treatment starts with setting specific, measurable and modest goals that are continually revised as progress is achieved.
  • Target behaviours are monitored – usually by the child and/or parent – to obtain a record of behaviour change.
  • Behaviour change processes include stimulus control, graded exposure, extinction and reward.
  • The perspective is educational: teaching behaviour change skills to the client. The term problem-solving skills may be used, but this does not necessarily mean that the treatment contains the other elements of conventional behavioural treatment.
  • The term cognitive (as in ‘cognitive behaviour’ or CBT) may imply the inclusion of strategies designed to modify cognitions (thoughts) which can be identified as important stimuli for behaviour.

A treatment is behavioural if the published paper:

  • uses the terms behavioural treatment, cognitive behavioural treatment, behaviour therapy or CBT
  • mentions learning theory
  • refers to the use of the common components of behavioural treatment (self-monitoring, goal-setting, stimulus control).

Terms that do not, in themselves, denote behavioural treatment are:

  • motivational interviewing
  • counselling
  • learning
  • psychological
  • psychotherapy
  • problem solving
  • cognitive.
Settings

From 42 studies that were included in this review the majority consisted of specialist outpatient weight reduction programmes in university obesity research clinics in the USA.18–23,25,27–31,38,40,43–47,51,53,59–61 The other studies were as following:

  • outpatient research clinic in Hong Kong33
  • inpatient child obesity treatment programme in a medical centre in France35
  • outpatient paediatric clinic at the University of Leuven (Belgium) 34
  • outpatient child obesity treatment programme at Tel Aviv University24
  • outpatient child obesity research clinic at the University of Graz (Austria) 32
  • outpatient service at a paediatric hospital in Cuba36
  • family paediatrician office (Primary Care) in Italy37
  • referral from school after screening in Sweden26
  • university outpatient research clinic in Australia39
  • US outpatient paediatric primary care clinic41
  • university outpatient research clinic in Isreal49
  • appears to be a research clinic in Australia54
  • inpatient rehabilitation hospital in Germany55
  • outpatient and inpatient research programmes in Belgium.56–58
Dropout rates

Overall, the retrieved studies are of poor methodological quality, and the high dropout rates have an even greater impact on the robustness of the evidence for the treatment of obesity in obese/overweight adolescents and children, as seen in Table 15.4.

Table 15.4. Dropout rates in studies on treatment of obesity.

Table 15.4

Dropout rates in studies on treatment of obesity.

Guidelines summary

In September 2005, the National Guideline Clearinghouse synthesised the recommendations on the assessment and treatment of obesity and overweight in children from six published guidelines (Table 15.5).

Table 15.5. Summary of recommendations on the assessment and treatment of obesity in children.

Table 15.5

Summary of recommendations on the assessment and treatment of obesity in children.

Interventions emphasising a combination of diet and/or physical activity and behavioural modification component
Weight loss

According to the results from one study by Epstein and coworkers based on a specialist weight management programme, it appears that lifestyle exercise** can be more effective than aerobic exercise and calisthenics in maintaining weight loss.18 Children in the lifestyle group maintained their weight change, whereas children in the aerobic exercise group gained significant amounts of weight. After 24 months the lifestyle exercise group had significantly lower per cent overweight than the aerobic group. After 24 months the difference in per cent overweight between the lifestyle exercise group (large reduction in sedentary behaviours) and the calisthenics group was significant. At 24 months, the lifestyle group had maintained relative weight changes whereas the calisthenics group had returned to baseline levels.18

Two studies looked at the effect of decreasing sedentary behaviours on weight reduction. In one of these studies, Epstein and coworkers compared reinforcing decreased sedentary activity to reinforcing increased physical activity, and reinforcing decreased sedentary activity and increased physical activity.19 The results suggest that the decreased sedentary behaviours group (reduced time spent on watching television, playing computer games, imaginative play, talking on the phone and playing board games) had a more significant reduction in percentage overweight than the combined and exercise groups at year 1.19 In another study, Epstein and coworkers compared decreased sedentary behaviours with increased physical activity; high against low doses of increased physical activity; and high against low doses of reducing sedentary behaviours.20 Both the reduction of sedentary behaviours and increase of physical activity ranged from 10 to 20 hours per week. Results show a reduction in percentage overweight greater at 6 months than at 24 months for all groups, having a slightly higher reduction in both the high dose decreased sedentary behaviour and high dose increased physical activity groups. The results indeed suggest that targeting inactivity can be as useful as targeting increased physical activity. However, the results were only statistically significant for the first study.19

In one Epstein et al. study that included girls aged 8–12 years, percentage overweight in the diet plus exercise group showed greater decreases from baseline to 6 and 12 months, than in the diet-only group.21 Nevertheless, results were only statistically significant for the 6 months follow-up. One clinical controlled trial22 compared a protein-sparing modified fast diet (PSMF) (50% protein, 40% fat, 10% carbohydrate) with a hypocaloric balanced (HCB) (20% protein, 30% fat, and 50% carbohydrate) in children ranging from 7.5 to 16.9 years, associated with 20 minutes of daily aerobic activity and behavioural modification components. Results suggest that on a short-term basis (10 weeks), PSMF achieved a greater weight loss (statistically significant) than the HCB diet. At 6 and 14.5 months the results were not significant. However, percentage overweight significantly decreased in the PSMF compared with the HCB at 6 and 14.5 months.22

Epstein and coworkers23 also assessed the effects of weight change on serum lipids in overweight children aged 8–12 years old. This study compared participants who were on diet only to participants with diet plus lifestyle change exercise programme and a control group, although the two treatment groups (which appear to have been given behavioural intervention as described in the evidence table) were then combined as no difference was found between them. Up to 6 months, significant improvements in weight and percentage overweight were reported, although such changes were not maintained at 5 years follow-up.23

Eliakim and coworkers compared the effects of a weight management programme on body weight in obese children and adolescents aged 6 –16 years against a control group that was referred every 3 months to an outpatient nutritional consultation. Participants were prescribed a hypocaloric diet, a twice-weekly training programme (1 hour) and encouraged to reduce inactivity. The treatment group had a statistically significant decrease in body weight and BMI at 3 and 6 months compared with the controls.24

Sothern and coworkers25 evaluated the safety, feasibility and efficacy of a resistance training programme in obese children aged 7–12 years, with a PSMF diet, a moderate-intensity progressive exercise programme and behaviour modification. Total body weight significantly decreased at 10 weeks (p < 0.0003) and 1 year (p < 0.0003), and BMI also significantly decreased at 10 weeks (p < 0.0001) and 1 year (p < 0.0001). However, the difference in BMI between the 10 week and 1 year follow-up was not significant.25

Flodmark and coworkers26 evaluated the effect of family therapy on 44 obese children aged 10–11, by studying two treatment groups: one whereby children were given family therapy, and a second group that was given conventional treatment (for further detail on each treatment see evidence table [Appendix 13]). The results suggest that the increase of BMI was higher in the control group and conventional treatment group (p = 0.04 and p = 0.02, respectively), than in the family therapy group.26

In the SHAPEDOWN programme study,27 a range of cognitive, behavioural, affective, and interactional techniques was adapted to the needs of adolescents, encouraging adolescents to make continuous, sustainable, small modifications in diet, exercise, relationships, lifestyle, communications and attitudes. Relative weight decreased significantly during the first 3 months. For the subsequent 3 months, both groups decreased their relative weights in comparison with baseline values (p < 0.001), although by month 15 both groups’ relative weight had diverged significantly (p < 0.01).27

Israel and coworkers28 assessed two levels of parental involvement roles in the treatment of childhood obesity, and findings suggest that the group in which parents focused their efforts on the child (helper condition) was slightly superior than the group in which the parents also engaged in their own weight loss (weight loss condition), although there was no difference in the overall child’s weight status at 1-year follow-up.

Parent weight as predictor of child weight

Epstein and co-workers also examined, in two studies belonging to the same trial,29,30, the effect of parent weight on the weight loss of obese pre-adolescent children – beyond the effect of parent control versus child self-control. At year 1, children of non-obese parents were significantly lighter (p < 0.01) than baseline (% = −16.3) and lighter (p < 0.01) than children of obese parents (% = −7.7). Nevertheless these changes were not significant at 5 years. Thus, parent weight can be related to weight loss, but not weight maintenance in obese children.

Other outcomes

In one study, physical work capacity improved significantly, with increases of 33% from baseline to 6 months, and 55% from 6 to 24 months. Per cent time of being active also increased from baseline to 2 years, and targeted sedentary behaviours showed significant decrease from baseline at 6 and 24 months.31 Epstein and coworkers18 reported that only children in the aerobic exercise group maintained significant improvements in fitness at 1 years, whereas in the lifestyle exercise group, significant improvements were observed up to 6 months, although they returned to baseline levels of fitness at 1 year. No changes in fitness were observed for the calisthenics group. In another Epstein et al. study, child fitness improved significantly over time, with no differential changes by group.19 Physical fitness was also improved in the family therapy group after 1 year follow-up (p = 0.047).26

Epstein and coworkers18 also reported that eating behaviour in children improved significantly across all groups. The relative weight changes between parents and children increased across time, with p < 0.01 from 0 to 6 months, p < 0.01 from 6 to 12 months and p < 0.01 from 12 to 36 months.18

In one study, biochemical factors remained within normal values in every child throughout the duration of the study. Changes in both blood pressure and serum cholesterol and triglyceride levels were not significant. However, when the two groups were combined the initial mean serum cholesterol values decreased significantly at 10 weeks.22

Epstein and coworkers reported that high-density lipoprotein (HDL)-cholesterol levels significantly increased over the 6 months of the study. Serum cholesterol (p = 0.03) and serum triglycerides levels decreased (p = 0.01) significantly over 6 months. Moreover, fitness also improved significantly.23

In one clinical controlled trial, endurance time was significantly greater in the treatment group.24

Participants of the SHAPEDOWN trial showed significant improvements in weight-related behaviour, depression, self-esteem and knowledge of weight management concepts at post treatment and at 12 months follow-up compared with the control group.27

Children with non-obese parents were more compliant towards calorie limit (p = 0.01), exercise goal (p = 0.02), and self-monitoring (p = 0.01) components of treatment and showed better results in eating behaviour (p < 0.01), than those with obese parents.29,30

Reported harms

Almost half the children in two dietary groups reported decreased appetite. Hunger, fatigue, weakness, and muscle cramps were more common in the hypocaloric group. Of the children in the PSMF group, 11% reported bad breath, and 19% of the children in the HCB group reported headaches and abdominal pain.22

Other factors

No additional analysis was conducted on the effect of age, ethnicity, socioeconomic status, previous treatment for obesity, motivation, gender, degree of overweight/obesity, current medical conditions and setting and/or healthcare professional.

Methodological and context notes

It must be noted that in the study by Figueroa-Colon,22 both dietary groups were placed on a hypocaloric diet after 3 months. However, this study was very small. The authors contend that both PSMF and HCB diets should not be used without close medical supervision.22

Interventions emphasising diet and physical activity
Weight loss

The results from three studies32–34 suggest that combining physical activity with dietary interventions is more effective than diet alone. In one case, Schwingshandl and coworkers32 reported that after 12 weeks the children given physical training and dietary advice (mean age 11.0) had significantly greater mean change in fat-free mass than the children given dietary advice alone (mean age 12.2).

An additional primary study, which was published after the Cochrane review, reported that at 1 year, there were significant changes in body fat content from baseline in the combined group but not in the diet-only group. No change in weight was reported, although BMI showed no significant changes in either group.33

Reybrouck and coworkers34 compared a low-calorie diet (800–1000 kcal) combined with physical activity against diet only in children aged 3.9–16.4 years. The results also suggest that the mean decrease in overweight at 4 months was significantly greater for the children in the combined group than in those treated with diet only. At 8 months the mean decrease was much smaller and similar between the two groups.

Rolland-Cachera and coworkers35 compared two different diets in an inpatient setting in France: one composed of 15% protein and 54% carbohydrates (prot −), and the other composed of 19% protein and 50% carbohydrate (prot +) in children aged 11–16 years, with no statistically significant differences being reported between the two groups. Both groups showed a mean BMI decrease of 12.5 (statistically significant). Results suggest that the prot+ content did not induce any additional effectiveness in the treatment of childhood obesity, although weight loss was achieved with the combination of a moderately energy-restricted diet and normal fat content and physical activity (7 hours per week of vigorous sports and 7 hours per week of outdoor activities).

Amador and coworkers36 compared a non-restricted diet with a restricted one (up to 30% of energy requirements) combined with a physical activity programme. Results suggest that a non-restricted diet delivered a greater weight decrease (statistically significant) than the restricted diet at 6 and 12 months, although greater at 6 months. Nova and coworkers37 assessed the effect of having a greater level of involvement of the family paediatrician and family in the long-term management of obese children aged 3–12 years. The results suggest that a greater involvement from the family paediatrician and commitment from the family have a significantly greater reduction in percentage overweight at 6 and 12 months.

Other outcomes

In one trial, a significant decrease was seen in total cholesterol in both groups with low-density lipoprotein (LDL)-cholesterol decreasing only in the exercise group. Fasting glucose (p < 0.002) reduced slightly in the exercise group only. Between the first and second year, in both groups the energy intake increased by 171 kcal, physical activity decreased and time watching television increased.35

Other factors

No additional analysis was conducted on the effect of age, ethnicity, socioeconomic status, previous treatment for obesity, motivation, gender, degree of overweight/obesity, current medical conditions and setting and/or healthcare professional.

Methodological and context notes

Again, there is lack of robust evidence on this specific topic, thus, the validity and generalisability of the conclusions remain unconfident. Nevertheless, it is worth noting that due to the other health benefits, a healthy diet and physical activity are recommended for everyone regardless of their weight. In this sense, a healthy diet and the increase of the levels of physical activity in children should be promoted, regardless of their effect on weight reduction.

Interventions emphasising diet only
Weight loss

Few studies have examined the effects of dietary interventions alone on weight reduction in obese children, and to date we have not found any randomised clinical trial that assesses such interventions per se. Nevertheless, one retrospective cohort study38 has suggested that a low glycaemic index (GI) diet can be effective in the management of childhood obesity with a mean age of 10.6 years. The results indicate that the low-GI diet (with no restriction of total energy or specific macronutrient consumption) had a statistically significant BMI reduction of −1.47 kg/m2 whereas the standard reduced-fat diet had a reduction of −0.20 kg/m2 (statistically significant).38

Other outcomes

No other outcomes have been reported.

Other factors

The statistically significant difference (−1.15 kg/m2 for the low-GI diet vs −0.03 kg/m2 for the reduced-fat diet) remained the same after adjusting for age, sex, ethnicity, length of follow-up, baseline BMI and behaviour therapy referral.38 No further analysis was conducted on socioeconomic status, previous treatment for obesity, motivation, degree of overweight/obesity, current medical conditions and setting and/or healthcare professional.

Methodological and context notes

It must pointed out that the data concerning dietary interventions were based on a retrospective cohort study, with a significant dropout rate and possible biases.38

Interventions where the main focus was behavioural treatment in comparison with no treatment or usual care
Weight loss

Part of the evidence suggests that behaviour therapy can be more effective than conventional care, as three studies reported better results for behaviour therapy compared to usual care†† and/or controls.39–41 Senediak and Spence39 examined the effects of rapid (eight sessions in 4 weeks) or gradual (eight sessions over 15 weeks) behaviour therapy versus a non-specific control condition and a wait-list control group in obese children aged 6–13 years. Percentage overweight in the rapid behaviour group significantly decreased from baseline to 6 months, and in the gradual behaviour group, percentage overweight significantly decreased from baseline to 6 months. In the non-specific control group, mean percentage overweight significantly decreased from baseline to 6 months. No significant differences were found between the rapid and gradual groups over 6 months.39 In one of Epstein and coworkers’ trials, which assessed a family-based behaviour therapy on obese children aged 5–8 years, BMI significantly decreased from baseline to 12 months in the behaviour group compared with the control group (prescribed with the traffic light diet‡‡ and physical activity six times per week, but no behavioural intervention).40 Post-hoc analyses showed significant differences between groups for per cent overweight and BMI at 8 and 12 months. In Saelens and coworkers41 evaluation of a behavioural weight control programme for obese children aged 12–16 years, the results suggest that the BMI for the experimental group (healthy habits [HH] – multi-component behavioural weight control intervention) decreased (not significant) from baseline to follow-up. For the usual care group (typical care [TC] – single session of physician weight counselling), BMI significantly increased after the 4-month treatment and 3-month follow-up.41

Other outcomes

In one study, results suggest that treated children had improved eating habits compared with the control group, and a main effect of improved self-control was observed over time in children, although no changes in parent self-control were reported.40 According to Saelens and coworkers,41 the HH adolescents reported higher rates of total and eating specific behavioural skills use than the TC adolescents (p < 0.03). Parents of the HH adolescents also reported that their adolescents used more overall and specifically eating-related behavioural skills than did parents of TC adolescents (p < 0.04). HH adolescents continued to report higher overall and eating-related behaviour skills use at follow-up assessment compared with the TC adolescents (p < 0.01).

Other factors

No additional analysis was carried out on the effect of age, gender, ethnicity, socioeconomic status, previous treatment for obesity, degree of overweight/obesity, motivation, frequency or length of the intervention and current medical conditions.

Interventions comparing behavioural treatment at varying degrees of family involvement
Weight loss

Several pieces of research on obesity treatment in children and adolescents have assessed the influence of including family members in the therapy process.

There seems to be substantial evidence that behavioural treatment is more effective in children and adolescents aged 6–16 years if the parent(s) are given the main responsibility for the behaviour change.42–47 Golan and Crow48 published a 7-year follow-up of previous studies,42,49 and results demonstrated that the mean reduction in per cent overweight was superior in children of the parent-only group compared with those in the children-only group (p < 0.005). These results suggest that involving parents in the therapy process can be also more effective on a long-term weight loss and maintenance basis in obese children, as also seen in the next study. A 10-year follow-up study published by Epstein and coworkers50 with children aged 6–12 years aimed to compare three groups: child and parent target (group 1), child target (group 2), and non-specific target (group 3). At both 5 and 10 years, significant per cent overweight differences (p < 0.05) were shown between children in groups 1 and 3 with children in group 2 midway between the other groups.

Parent weight as predictor of child weight

One study which consisted of a secondary data analysis based on three RCTs51 (already included in this review), aimed to assess whether parent-standardised BMI (z-BMI) change influences child z-BMI (in children aged 8–12 years). Results suggested that parent z-BMI change can be a predictor of obese child z-BMI change in family-based treatment. Children of the parents in the greatest z-BMI change quartile had greater reductions in z-BMI changes over time (p = 0.01) than children of parents in the other three groups, who had smaller reductions or gains in z-BMI.51 Nevertheless, it seems that the results are also connected with the participation of the parents in the treatment process, and not solely due to their weight loss.

Other outcomes

Wadden and coworkers43 reported that total cholesterol concentration and HDL-cholesterol decreased significantly during treatment (p< 0.01 and p < 0.06, respectively). Furthermore, scores on the Pier–Harris scale§§ increased significantly (p < 0.05) during treatment, indicating possible improvement in self-esteem, and the child depression inventory decreased significantly (p < 0.01), which shows reductions in feelings of depression.43 Even before treatment, participants scored well within normal limits on both measures.43 Neither the participants mean initial risk of cardiovascular disease, nor mean triglyceride level or blood pressure registered any significant changes.43

Golan and coworkers42 pointed out that significant increase in the children asking permission to take or buy sweets was noted only in the experimental group – (p < 0.001) for taking and (p < 0.01) for buying – at termination of the programme. An overall reduction in the prevalence of poor eating habits was significantly greater in the experimental group. Moreover, a significant positive correlation was reported between the children’s reduction in overweight and the following: presence of food stimuli in the house, eating while standing, eating while doing another activity, eating following stress situations, eating between meals, place of eating and activity level.42 At the 7-year follow-up, Golan and coworkers reported that 6.6% of the girls from the child-only group reported eating disorder symptoms (binging and purging).48

Israel and coworkers44 reported that the analysis of the Eating and Activity Self-Control Scale (EASC) indicated an increase in children’s self-control and parental control regarding weight-related behaviours (p < 0.001 and p < 0.05), respectively). Parental opinion from the self-control rating scale also indicated significantly more self-controlled behaviours at week 26 than at week 1. Moreover, higher EASC self-control scores were significantly correlated with decreases in percentage overweight during treatment (p < 0.05).

Epstein and coworkers47 indicated that there was an overall decrease in food intake from pre to post treatment (8 months), and that there was a strong relation between changes in red food intake and weight loss (p < 0.02).

Other factors

No additional analysis was carried out on the effect of age, gender, ethnicity, socioeconomic status, previous treatment for obesity, degree of overweight/obesity, motivation, frequency or length of the intervention and current medical conditions.

Methodological and context notes

Another study examined the feasibility and generalisability of a family-based behavioural treatment for childhood obesity in a clinical setting in Britain, and assessed whether the results were comparable with the original studies in the USA. Although this study was excluded from our review, it is important to mention, considering the proportion of studies from the USA that are included in this study. The results support the applicability of the family-based behavioural treatment in a clinical setting in Britain.52

Interventions comparing problem solving, in addition to behaviour therapy with usual care or behaviour therapy
Weight loss

Based on the analysis of the results from two studies, there seems to be some contradictory evidence regarding the comparison of problem solving with behaviour therapy. Graves and coworkers53 aimed to examine the effects of incorporating parental problem solving training in a behavioural weight reduction programme in obese children aged 6–12 years. Results suggest that combining problem solving with behaviour therapy may be more effective than behaviour therapy alone, as percentage overweight in the problem solving group had a greater decrease (p < 0.01) from baseline to 6 months than in the behaviour treatment only group.53 Epstein and coworkers31 also compared including parent and child problem solving to a behavioural weight control programme, only child problem solving and standard treatment (behavioural intervention only) in obese children aged 8–12 years. Nevertheless, the results suggest that problem solving did not provide any additional benefits in terms of weight loss.31

Other outcomes

Improvements in problem solving for both parents and children was reported in one study for the problem solving group vs problem solving with family.31 In another study, parents in the problem solving group increased their problem solving ability from pre to post treatment, whereas behavioural and instruction-only parents did not.53

Children in both the problem solving and behavioural groups increased their consumption of green food and decreased their consumption of red foods significantly more than instruction-only children.53 Moreover, a positive correlation was found between pre-and post-treatment weight change, and change in the consumption of red foods (p < 0.05) and green foods (p < 0.01).

Other factors

No additional analysis was carried out on the effect of age, gender, ethnicity, socioeconomic status, previous treatment for obesity, degree of overweight/obesity, motivation, frequency or length of the intervention and current medical conditions.

Interventions focusing on CBT
Weight loss

Duffy and Spence54 did not find any additional effectiveness of cognitive therapy techniques such as targeting monitoring of negative thoughts, restructuring of maladaptive thoughts, problem solving, and self-reinforcement as adjuncts to behaviour therapy for 7–10-year-old and 10–13-year-old children. According to the results, percentage overweight significantly decreased (p < 0.001) in the relaxation control group from pre-treatment to the 6-month follow-up. In the CBT group percentage overweight also significantly decreased (p < 0.001) from pre-treatment to the 6-month follow-up. No significant differences were observed between groups.54 Similar results were reported in another study by Warschburger and coworkers,55 in which 6 months after intervention, 14.8% of the children and adolescents in the experimental group could be classified as non-obese, against 9.7% in the control group. However, these differences were not statistically significant.55

Braet and coworkers56,57*** examined adding a healthy eating lifestyle programme rather than a strict diet, combined with CBT, delivered through different therapeutic forms to children aged 7–16 years. Significant loss of weight was reported in all therapeutic groups, as early as 3 months up to 1-year follow-up, and the results suggest that group rather than individual approaches result in significantly better outcomes. No significant results were found when comparing the 1-year follow-up with the 4.6 year follow-up.

Braet and coworkers58††† also-assessed a 10-month inpatient cognitive behavioural weight loss programme, where the participants’ median age and BMI was 14 years and 33 kg/m2, respectively. During treatment, the children in the study group showed a decrease in the median adjusted BMI of −48% (range −4% to −102%). At 14-month follow-up, 13/27 children showed an increase in their overweight of less than 10% or continued to lose weight, compared with their post-treatment weight and 14/27 children had an increase of more than 10% overweight (up to +41%).

Other outcomes

Duffy and Spence54 reported a significant reduction in the consumption per day of red foods from pre-treatment to post-treatment in both groups (p < 0.001), with no significant differences between groups.

Warschburger and coworkers55 reported that both groups showed improvements in their quality of life over time (p < 0.01), and improvements in self-reported eating behaviours for the experimental group compared with the control group (p < 0.05).

Other factors

No additional analysis was carried out on the effect of age, gender, ethnicity, socioeconomic status, previous treatment for obesity, degree of overweight/obesity, motivation, frequency or length of the intervention and current medical conditions.

Interventions focusing on reinforcement and/or stimulus control of sedentary behaviours
Weight loss

From the analysis of the results of one study, there appears to be no significant improvement in applying mastery criteria and contingent reinforcement, as the experimental group decreased from 60.6 % (SD 25.3) at baseline to 30.5% at 6 months and 34.1% at 1 year, and in the control group mean per cent overweight decreased from 58.8% (SD 19.6) at baseline to 38.8% at 6 months and then increased to 42.1 % at 1 year.59

Based on one study, results suggest that stimulus control and reinforcing reduced sedentary behaviours are equally useful to reduce sedentary behaviours and consequently in reducing standardised BMI (z-BMI) figures. z-BMI values (kg/m2) for the stimulus control group were 3.3 ± 1.0, 2.3 ± 1.0 and 2.4 ± 1.0, at 0, 6 and 12 months, respectively, whereas the values for the reinforced reduction group at the same time were 3.2 ± 1.0, 2.2 ± 1.1 and 2.6 ± 1.0, respectively.60

Epstein and coworkers61 assessed how preferences for food used to reinforce behaviour change in young children can be applied to modify food preferences of older, obese children. Two groups were assigned: one treatment group where novel low-calorie foods were given contingent upon the behaviour changes established for weight loss; and a control group where low-calorie foods were provided for a daily snack and not contingent upon behaviour change. There was also no difference in the rate of change in percentage overweight between the two groups, although both had a significant decrease (p < 0.01) from 2 to 6 months (25.1%). The findings of this study do not back-up the hypothesis that by using unfamiliar foods as reinforcers, one can change the preference of children for those unfamiliar foods.

Other outcomes

Significant changes regarding consumption of red foods per week (p < 0.05) and days within the caloric range (p< 0.025) were reported. Moreover, parents showed a significant improvement in knowledge of behavioural principles across time (p < 0.001). Epstein and coworkers60 also reported significant reductions in the consumption of high-energy-density foods, and increases in physical activity and consumption fruits and vegetables were observed in both groups.

Other factors

There were differences between children who substituted physically active for sedentary behaviours and those who did not. There were a higher percentage of boys substituting physically active for sedentary behaviours than girls.60

No additional analysis was carried out on the effect of age, ethnicity, socioeconomic status, previous treatment for obesity, degree of overweight/obesity, motivation, frequency or length of the intervention and current medical conditions.

15.2.4. Pharmacological Interventions

Orlistat in weight loss and other outcomes in children and adolescents

15.2.4.1. Evidence statements

(Table 15.6)

Table 15.6. Evidence statements and grading.

Table 15.6

Evidence statements and grading.

15.2.4.2. Orlistat (120 mg three times daily) versus placebo
Weight loss

Two placebo controlled randomised trials were identified.62,63 One trial62 compared the use of orlistat in combination with a nutritionally balanced hypocaloric diet (designed to produce an initial weight loss of 0.5–1.0 kg per week), an exercise plan and a programme of behaviour modification (see evidence table for details [Appendix 13]). The study participants (n = 533) were adolescents aged 12–16 years who were overweight and obese (BMI ≥ 2 units above the 95th centile, excluded if BMI ≥ 44 kg/m2 or weighed over 130 kg). Both groups lost weight during the first 4 weeks of the intervention, with weight loss remaining stable in the placebo group from weeks 4 to 12 compared with continued weight loss in the orlistat group until week 12. Weight increased in both groups from week 12 onwards. At 12 months, the group assigned to orlistat gained less weight than the group assigned to placebo (+0.53 kg vs +3.14 kg, p < 0.001) with decrease in BMI compared with an increase in the placebo group (−0.55 vs +0.31, p = 0.001). Significantly more participants lost 5% or more and 10% or more of initial BMI in the orlistat group at 12 months. Overall, the trial was assessed as being of good quality.62 It is important to note that some of the children would have completed linear growth, although some would have been still growing during the study.62

Another RCT63 compared the effect of orlistat with placebo alone, without any additional dietary or activity component. This trial was assessed as of poorer quality than Chanoine’s 2005 study,62 and was considerably smaller. Participants (n = 32) were aged between 12 and 16 years, and were overweight or obese (BMI ≥ 85th centile adjusted for age and gender). At three weeks, participants in both groups had lost weight (7.0% orlistat vs 7.8% placebo of initial body weight), but the difference did not appear to be significant (no p value was reported). This trial, however, was not designed to measure weight loss, but the effects of orlistat on mineral balance.63

A quasi-randomised controlled study64 investigated the effect of orlistat in combination with diet (20% reduction in kcal per day) and activity (at least 30 minutes moderate daily physical activity). Participants (n = 42) were aged 10–16 years, and were obese (> 140% weight for height index). A significant weight change was seen in the orlistat-treated group compared with placebo (−6.27 kg vs +4.16 kg, p < 0.001) at approximately 10–11 months (follow-up varied). Similarly, the orlistat group showed a decrease in BMI compared with an increase in the control group, and weight loss (percentage of initial weight) compared to a weight increase in the control group. However, the participants in the orlistat group had a higher mean initial BMI than the placebo group (32.5 vs 31.2, p = 0.018). Again, this study64 was small and was assessed as of poorer quality than Chanoine’s study above.62

Two before-and-after studies investigated the effect of orlistat in children65 and adolescents.66 Norgren and colleagues65 conducted a pilot study of orlistat and dietary advice (sources of fat and recommended daily fat intake) in children (n = 11) aged 7–12 years (pre-pubertal) and who were overweight or obese (BMI >4 SD above normal). In the 12 weeks preceding treatment (no intervention), children tended to gain weight. However, at 12 weeks after initiation of orlistat treatment, the median weight change was −4.0 kg (range −12.7 kg to +2.5 kg, p = 0.016), with an corresponding decrease in BMI.65

Another study66–68 assessed the effect of orlistat as an adjunct to both a comprehensive behavioural programme and periods of inpatient evaluation (see details below). Participants (n = 20) were aged 12–17 years and were obese (BMI ≥ 95th centile for age, sex, race). Significant weight loss was seen at both 4 and 6 months (−4.4 kg and −5.4 kg, respectively), with a corresponding decrease in BMI. At 6 months, 30% of participants has lost 5% or more of initial weight and 15% had lost 10% or more.66–68

Both of these before-and-after studies were small, and due to the design used had increased potential for bias compared with studies using a controlled study design.

Other outcomes

A large good-quality RCT showed significant decreases in diastolic blood pressure (DBP), but not in other outcomes such as lipids, glucose levels, triglycerides or systolic blood pressure (SBP) at 12 months.62

The remainder of the studies reported a variety of different outcomes at different times.

Other factors
Age

Most studies included adolescents only (range 10–17 years), but one study65 included pre-pubertal children aged 7–12 years.

Current medical conditions

McDuffie and coworkers66 included adolescents with one of several obesity associated comorbidities (hypertension, type 2 diabetes or glucose intolerance, hyperinsulinaemia, hyperlipidaemia, hepatic steatosis, sleep apnoea). All other studies included otherwise healthy participants.

Degree of obesity

Most studies included children or adolescents who were overweight or obese, but Chanoine62 included only adolescents who were overweight (and excluded those who were very obese) and Ozkan and coworkers64 included only obese adolescents.

Setting

All studies were based in specialised treatment or research centres. Two studies included either partial66 or total63 inpatient treatment.

Country

No studies were based in the UK.

Methodological and context notes

Due to the inclusive searches and criteria, studies of varied design are included.

Sibutramine in weight loss and other outcomes in adolescents and children

The marketing authorisation for sibutramine has been suspended. See details.

15.2.4.3. Evidence statements

(Table 15.7)

Table 15.7. Evidence statements and grading.

Table 15.7

Evidence statements and grading.

15.2.4.4. Evidence review

This section reviews the evidence on the effectiveness of pharmacological interventions (sibutramine) combined with behaviour therapy and/or dietary interventions and physical activity in obese adolescents.

Types of study
  • RCTS
  • Controlled clinical trials
  • Controlled before-and-after studies
  • Cohort studies with a control group

Only studies with a minimum duration of 6 months or above (including follow-up) and published after 1985 were included, and also RCTs with a primary aim other than the treatment of childhood obesity. Studies based in a setting other than clinical and delivered by non-health-care professionals (for example, school teacher) were not included in this review.

We did not retrieve any studies in the update searches that would potentially add further details or contradict any of the recommendations.

Types of participant
  • Participants aged under 18 years at the start of the study, and exceptionally studies where the age cut-off was above 18 years and where the majority of the participants were below 18 years or presented age stratification.
Types of outcome
  • Primary outcomes to be measured (not self-reported) estimates of overweight in per cent and BMI.
  • Secondary outcomes to be behaviour change, participants’ views, measures of self-esteem, health status, well-being and quality of life.

Again, there is limited evidence in this field. The American Heart Association (AHA) pointed out that sibutramine has been studied in an RCT. It stated that sibutramine had been shown to be efficacious compared with behaviour therapy alone, but it may be associated with side effects including increases in heart rate and blood pressure.1 The NHMRC guidelines2 stated that there is no evidence that sibutramine has a role in the management of adolescent obesity. Similarly, the Singapore Ministry of Health guidelines referred to the non-existence of data on the long-term efficacy and safety of medication in childhood and adolescent obesity (grade C, level).1

Only two RCTs70,71 were retrieved from searches. One study consisted of 82 participants71 whilst the other trial included 6070. One of the studies was a 6-month placebo controlled trial,70 in which participants were given 10 mg/day. This 6-month period was preceded by a 4-week single-blind period where all participants were given a placebo capsule. Participants were also advised to achieve an energy deficit of 500 kcal/day and to undertake at least 30 minutes of moderate aerobic exercises per day, and to reproduce a ‘regular’ clinical setting, no behavioural counselling was given. Routine clinical advice to increase physical activity was given by the medical practitioners in the form of a leaflet and only one appointment was made with the dietitian. At 6 months the sibutramine group had a statistically greater reduction in weight and BMI (p < 0.001) compared to the placebo group.

The other study consisted of a 12-month trial,71 which comprised two phases: one placebo controlled period for 6 months and an open label extension for another 6 months. Participants were given a family-based behavioural weight loss programme including a 1200–1500 kcal diet and 120 minute/week of physical activity. At 6 months the placebo group had a significantly smaller reduction of weight and BMI than the sibutramine group (p = 0.001). The group that was randomised to sibutramine for the first 6 months and then continued on the medication for another 6 months gained weight during the second 6 months and ended up not significantly different from the placebo group. There was no statistically significant difference at month 12 between the two groups as the placebo group were able to switch to sibutramine in the open-label phase at 6 months to 12 months.

In one of the studies,70 there was a significant decrease (p < 0.05) in triglycerides and very-low-density lipoprotein at week 24 in the sibutramine group. In the other hand, in the other study,71 a significant increase in HDL-cholesterol (p = 0.001) and significant reductions in serum insulin (p < 0.001) were reported.

One of the studies70 reported statistically significant adverse events – constipation (p = 0.039) in the sibutramine group. The other study71 reported that during the 12-month study period, sibutramine was reduced to 10 mg in 16 participants and to 5 mg in 7 participants (42 participants in the sibutramine group). Ten participants discontinued treatment due to increases in blood pressure.

15.2.5. Surgery for weight loss and other outcomes in adolescents and children

15.2.5.1. Evidence statements

(Table 15.8)

Table 15.8. Evidence statements and grading.

Table 15.8

Evidence statements and grading.

15.2.5.2. Evidence review on surgery

This section reviews evidence on the assessment of the effectiveness of bariatric surgery in obese adolescents. Due to the absence of methodologically strong studies among the existing literature, more expanded inclusion criteria were adopted for this particular review. Thus, the following were included:

Types of study
  • RCTs
  • Controlled clinical trials
  • Controlled before-and-after studies
  • Cohort studies with a control group
  • Non-comparative studies (case series and case studies)

Only studies with a minimum duration of 6 months or above (including follow-up) were included.

Update searches have been undertaken, however no further studies were identified.

Types of participant
  • Participants aged under 18 years at the start of the study, and exceptionally studies where the age cut-off was above 18 years but where the majority of the participants were below 18 years and results were stratified by age.
Types of outcome
  • Primary outcome of measured (not self-reported) estimates of weight change in per cent and BMI.
  • Secondary outcomes of behaviour change, participants’ views, measures of self-esteem, health status, well-being and quality of life.

The National Institutes of Health (NIH) Bariatric Consensus Development Conference in 199172 set out the basis for the increase in adult bariatric surgery undertaken in the previous 6 years. This conference concluded that insufficient data existed to make recommendations for patients younger than 18 years of age. Fourteen years later, outcome data remain limited for adolescents, with no controlled assessment of bariatric surgery in this group.

Guidelines for surgical intervention

Only three guidelines have issued recommendations with regard to bariatric surgery in adolescents: NHMRC Australian guidelines for the management of overweight and obese children and adolescents,2 the Singapore Ministry of Health clinical guidelines,73 and the Institute for Clinical Systems Improvement (ICSI).74 The Scottish Intercollegiate Guidelines Network (SIGN)6 did not propose any recommendations.

The NHMRC Australian guidelines2 recommended (grade C) that bariatric surgery might be considered as a last resource in severely obese adolescents with obesity-related comorbidity. The surgery should be carried out only in an experienced surgical centre after meticulous consultation, education of the patient and family, and a full psychological assessment. Furthermore, postoperative care should be ensured in an experienced weight-management centre. The Singapore Ministry of Health guidelines73 stated that bariatric surgery cannot be recommended (grade B) for most adolescents, although with exceptions for those at the highest risk of mortality from obesity, and with both patient and parental understanding of the consequences of surgery. Finally, the ICSI74 recommended that bariatric surgery should be undertaken in carefully selected patients, those with a BMI greater than or equal to 40 kg/m2, or with a BMI of 35–39.9 kg/m2, and those who are at a very high absolute risk for increased morbidity or premature mortality. Patients should be motivated, well-informed in disease management, psychologically stable and accepting of operative risks.

Finally, the AHA guidelines stated that:

  • ‘surgical approaches to treat severe adolescent obesity are being undertaken by several centres. Indications used include a BMI > 40 kg/m2 and severe associated co-morbidities, such as obstructive sleep apnoea, type 2 diabetes mellitus, and pseudotumor cerebri
  • ‘more severe elevation of BMI (> 50 kg/m2) may be an indication for survival treatment in the presence of less severe co-morbidities such as hypertension and dyslipidemia, particularly if the degree of overweight hinders performing the activities of daily living
  • ‘an experienced team approach including comprehensive medical and psychological evaluation is critical both for selection of appropriate candidates and for postoperative care that is sophisticated and often intense
  • ‘weight loss goals and reduction of morbidity are often achieved with gastric bypass surgery. The rates of short-term mortality appear to be low but significant complications can occur. Intermediate and long-term outcomes, including information on malabsorption of critical nutrients, are unknown.
  • ‘overall, surgical therapy should be reserved for full-grown adolescents with the severest obesity-related morbidity, offered only by experienced multidisciplinary teams, and presented to families with appropriate informed consent procedures.’
Views and insights from surgical societies and associations and expert opinion

Inge and coworkers75–77 have thoroughly reviewed the role of surgery in the treatment of severe childhood and adolescent obesity. These authors contended that adolescents being considered for bariatric surgery should:

  • have been unsuccessful at managing weight for ≥ 6 months, as determined by the primary care provider
  • have achieved or nearly achieved physiological maturity
  • have a BMI ≥ 40 kg/m2 with serious obesity-related comorbidities or have a BMI of ≥ 50 kg/m2 with less severe comorbidities
  • be able to show adherence to comprehensive medical and psychological evaluations both before and after surgery
  • agree to avoid pregnancy for at least 1 year postoperatively
  • demonstrate responsibility in committing to nutritional guidelines postoperatively
  • provide informed assent to surgical treatment
  • possess decisional ability
  • be surrounded by a supportive family environment (ensure that both patients and families understand that bariatric surgery is not a cure for obesity but an effective weight loss tool if used in compliance with specific dietary and physical activity regimens; and to understand the known risks and possible side effects of bariatric surgery).

The authors asserted that the above suggested criteria could not be applied strictly to each patient but should be tailored to the individual’s needs, taking into account the level of maturity and severity of comorbid conditions. Strong emphasis was also placed on the requirement of having highly trained and skilled bariatric surgeons to perform safe and effective procedures, undertaken at appropriately equipped facilities capable of adolescent bariatric surgery. Furthermore, a multidisciplinary team with expertise in adolescent weight management and bariatric surgery is required to carefully manage all the aspects of the procedure for each individual.

Overall, the American Society for Bariatric Surgery78 agreed with the criteria proposed by Inge and co-workers, although they added:

  • A qualifier of a BMI of ≥ 35 kg/m2 in the presence of significant comorbidities.
  • A bariatric surgeon should have successfully performed at least 100 bariatric procedures or have completed a year-long bariatric surgery fellowship.
  • The preoperative presence of comorbidities as an indicator for surgery is not appropriate, as bariatric surgery can play an important role in preventing such comorbidities.
  • There is no real evidence that supports the argument that given any age and with a balanced nutrition, bariatric surgery will lead to impaired growth or early osteoporosis.
  • Recommendations for procedures should not be limited to the Roux gastric bypass and laparoscopic adjustable gastric banding (LAGB).
  • There is a need to discuss possible complications of gastric distension and possible rupture in the presence of a bowel obstruction when recommending the Roux gastric bypass; and an LAGB needs to encompass discussion of oesophageal dilation and the possible advent of functional and histological oesophageal problems in the future, alongside long-term risks of band erosion into stomach or balloon malfunction.
Primary studies synthesis‡‡‡

No RCTs or clinical controlled trials were retrieved from the searches. The bulk of the evidence consisted of case studies, case series and retrospective reviews that addressed bariatric surgery procedures as follows: LAGB, vertical banded gastroplasty (VBG),§§§ Roux-en-Y gastric bypass, jejunoileal bypass, biliopancreatic diversion and duodenal switch. All these come under the broader categories of restrictive, restrictive/malabsorptive, malabsorptive/restrictive surgical procedures.

  • Restrictive: the gold standard is LAGB. VBG was a forerunner to laparoscopic gastric banding, but had high rates of complications/failure. Both operations have similar, expected clinical outcomes. Laparoscopic gastric banding restricts intake (volume) of solid food. Older operations are VBG, horizontal gastroplasty, open adjustable banding.
  • Restrictive/malabsorptive: gold standard is gastric bypass (Roux-en-Y). Gastric bypass mainly restricts intake but also reduces absorption of nutrients.
  • Malabsorptive/restrictive: these are more similar to the older operations. The gold standard is duodenal switch (DS) and biliopancreatic diversion (BPD). Jejunoileal switch was the forerunner to the DS, but was abandoned in 1982. These reduce calorie absorption, with limited restriction.
Restrictive surgical techniques
LAGB

Three of the studies consisted of prospective case series 79–82 Abu-Abeid and coworkers79 looked at adolescents aged 12–19 years (who had been under the care of a dietitian for at least 1 year, and had failed to reduce weight with a 800 kcal/day diet) who underwent LAGB. Widhalm and coworkers82 followed eight patients with a mean age of 16.0 ± 1.3 years who underwent adjustable laparoscopic banding surgery. Dolan and coworkers80,81 also studied adolescents aged 12–19 years who underwent LAGB.

In one case study, an inpatient physical activity programme and dietary restriction to 800 kcal per day was given to a 12-year-old girl for 6 weeks. Since severe obesity persisted, she underwent LAGB. After the surgery, a daily caloric intake was still restricted to 1200 kcal per day and physical activity at least twice a week.83 Horgan and co-workers84 addressed patients aged 19 or younger who underwent LAGB between 2001 and 2003. Angrisani and coworkers85 also conducted a retrospective multicentre study in patients who underwent LAGB, aged 19 and younger.

Weight loss

In one study, BMI fell from 46.6 kg/m2 preoperatively to 32.1 kg/m2 at 23 months follow-up.79 In two studies belonging to the same trial that compared weight and BMI reduction in adolescents and adults, the median BMI of the adolescents decreased from 42.2 kg/m2 (preoperatively) to 32.4 kg/m2 at 12 months and further to 30.2 kg/m2 at 24 months.80,81 No differences were observed in weight reduction between adults and adolescents.80,81 Widhalm and coworkers82 reported that in all their eight patients there were no major problems after surgery, and the mean weight change after a mean follow-up period of 10.5 ± 6 months was − 25.0 ± 3.8 kg. In one case study, weight loss in a 13-year-old girl was 14 kg at 3 months post-operation.83

Angrisani and coworkers85 reported a mean percentage excess weight loss (%EWL) at 1, 3, 5 and 7 years follow-up of 45.6 ± 29.6; 39.7 ± 29.8; 43.7 ± 38.1; and 55.6 ± 29.2, respectively. Five of 25 (20%) patients had ≤ 25% EWL at 5 years’ follow-up, whereas none of the 10 patients subject to follow-up at 7 years had ≤ 25% EWL.

Other outcomes

Abu-Abeid and coworkers79 stated that all adolescents reported improved well-being; they were more physically active, more socially involved with their peers and reported feeling happier than before surgery.

Reported harms

Dolan and coworkers80,81 reported that two patients had complications: one had slippage of the band, and another patient required port replacement. Horgan and coworkers84 reported one patient who developed cholecystitis.

Angrisani and coworkers85 reported the following cases:

‘Laparoscopic conversion was necessary in 1 patient with gastric perforation on the anterior wall during perigastric band positioning. The overall postoperative complication rate was 6/58 (10.3%): Band slippage was observed in 1 patient and was treated by laparoscopic repositioning after 4 days, gastric pouch dilatation was observed in 2 patients and was treated by band repositioning, and intragastric migration was observed in 3 patients and was treated with band removal. The band also was removed in 2 patients for psychologic intolerance, and 1 patient was converted 2 years after surgery to laparotomic gastric bypass. The overall band removal rate was 6/58 (10.3%). Biliopancreatic diversion with gastric preservation and band left in situ was performed in 2 patients (3.4%).’

Regardless of the type of procedure, the authors stated that the postoperative course in the eight adolescents who underwent reoperation was uneventful.85

Older surgical techniques such as VBG, horizontal gastroplasty, open adjustable banding****

Capella and coworkers86 studied a form of gastric bypass which combined the VBG with a Roux-en-Y gastric bypass in adolescents aged 13–17 years of age. These adolescents had attempted several weight reducing regimens, such as medically supervised diets, physical activity, behaviour modification, commercial diets, psychological interventions and pharmacological agents. Postoperatively, patients were given advice from a dietitian on the benefits of a balanced diet and regular physical activity, and no attempts were made to refer them for another diet or behavioural modification programme.

Mason and coworkers87 retrospectively studied 47 severely obese individuals who were under 21 when they underwent VBG.

Weight loss

The mean BMI decreased from 49 kg/m2 preoperatively to 28 kg/m2 at 5.5 years post-operation.86

Mean BMI decreased from a mean 48.1 ± 7.01 kg/m2 at operation to 36.2 ± 5.99 kg/m2 at 5 years, and decreased from a mean 49.6 ± 7.73 kg/m2 at operation to 39.2 ± 7.15 kg/m2 at 10 years assessment. Average weight decreased from 138 kg at operation to 103.6 kg at 5 years, and from 135.8 kg at operation to 107.6 kg at 10 years.87

Other outcomes

In one study all serious comorbidities disappeared early in the weight loss process.86

Reported harms

No operation-related deaths among the 47 patients who underwent VBG were reported. No leaks, instances of peritonitis, wound infections or cases of pneumonia were reported. Three revisions were performed in female patients, two at 5 years, and the third at 12 years post-operation.87

Restrictive/malabsortive
Gastric bypass

The first studies on bariatric surgery date from the mid-1970s. In 1975, Soper and coworkers88 reported 18 severely obese adolescents younger than 20 years of age who underwent either gastric bypass or gastroplasty. Anderson and coworkers89 published a follow-up report on both procedures with 30 adolescents. In this study, careful dietetic counselling was provided to each patient and family. Both studies were case series.

Strauss90 reviewed records of adolescents aged 17 years or younger who underwent gastric bypass surgery, and who had made serious attempts at weight loss in diet and behaviour modification programmes; Stanford and coworkers91 also reviewed medical records of patients less than 20 years of age who underwent laparoscopic Roux-en-Y gastric bypass (RYGB). Breaux92 reviewed 22 patients (11 with sleep apnoea and 11 without sleep apnoea), whose ages ranged from 8 to 18 years, and who had undergone VBG, RYGB or BPD. BPD was only performed in super-obese patients with sleep apnoea.

One study consisted of interviews with patients who had undergone bariatric surgery;93 34 patients were interviewed an average of 6 years after the surgery and ranged from 11 to 19 years of age at the time of the surgery. Patients underwent RYGB or VGB.

Weight loss

The mean weight loss from two studies dating from 1975 and 1980 was approximately 40 kg at 3 years and 26 kg at 5 years post-operation.88,89

Strauss90 stated that maximum weight loss occurred by 12–15 months after the operation; 9 of 10 adolescents had weight loss in excess of 30 kg. The mean weight loss was 53.6 ± 25.6 kg for the nine adolescents who had persistent weight loss. Stanford and co-workers91 reported a BMI decrease from 55 kg/m2 (preoperatively) to 35 kg/m2 at 20 months’ follow-up. Rand and Macgregor93 reported that preoperatively, patients had an average BMI of 47 kg/m2, and at 6 years follow-up the average BMI had dropped to 32 kg/m2. The patients’ average excess body weight loss was 66%. Breaux92 pointed out that in the group without sleep apnoea BMI improved from a mean 56.4 kg/m2 to a mean 35.5 kg/m2 post-operation. In the group with sleep apnoea, mean BMI dropped from 70.3 kg/m2 to a mean BMI of 46.5 kg/m2 post-operation.

Other outcomes

Rand and Macgregor93 found that at follow-up, 82% of the patients considered themselves attractive, compared with a figure of 94% of those who felt unattractive before surgery.

Reported harms

Both Soper and coworkers88 and Andersen and coworkers89 reported postoperative complications. Soper reported: 3 wound infections; 3 patients with respiratory difficulty; 1 patient with thrombophlebitis; 1 patient with upper gastrointestinal bleeding; 1 patients with urinary tract infection; and 1 patient with protracted vomiting. Andersen reported: 3 patients with wound infections; 2 patients ‘slow to open’ due to stomal obstruction; 3 patients developed atelectasis; 2 developed pneumonia; and 1 developed a subphrenic abscess. Strauss 90 reported a serious complication in one patient with a distal gastric bypass who had protein-calorie malnutrition and micronutrient deficiency approximately 1 year after gastric bypass. Two other adolescents had symptomatic cholelithiasis requiring laparoscopic cholecystectomy. Small bowel obstructions occurred approximately 10 years after gastric bypass surgery in one patient.

Towbin and coworkers†††† found three cases of beriberi (thiamine or vitamin B-1 deficiency) in adolescents who underwent gastric bypass. Breaux92 reported nine complications including vitamin A and D deficiencies, folic acid deficiency, protein deficiency, gallstone development, kidney stone, postoperative laryngeal oedema and incisional hernia.

Malabsortive/restrictive
Duodenal switch and biliopancreatic diversion

Breaux92 reviewed 22 patients (11 with sleep apnoea and 11 without sleep apnoea), whose ages ranged from 8 to 18 years, and who had undergone VBG, RYGB or biliopancreatic diversion (BPD). BPD was only performed in super-obese patients with sleep apnoea.

Weight loss

Breaux92 found that in the group without sleep apnoea, BMI improved from a mean 56.4 kg/m2 to a mean 35.5 kg/m2 post-operation. In the group with sleep apnoea, mean BMI dropped from 70.3 kg/m2 to a mean BMI of 46.5 kg/m2 post-operation. Nevertheless, no figures were provided for the children who underwent BPD, as results for the three surgical procedures were all grouped together.

Other outcomes

All patients had resolution of sleep apnoea on the long-term follow-up.

Older surgical procedures such as jejunoileal bypass

Two retrospective reviews were retrieved that examined the long-term effects of jejunoileal bypass in obese adolescents. Organ and coworkers95 did a retrospective review on 16 patients aged 15–20 years at the time of the surgery (1970–1975). Silber and coworkers96 reviewed 11 patients (who had made repeated failures of medical-dietary treatments) aged between 11 and 22 years at the time of the surgery (between 1972 and 1974). Another study which consisted of a case series of four patients aged 11–16 years who had failed dietary treatment of at least 1 year was also retrieved.97

Weight loss

With regard to the two studies that addressed jejunoileal bypass, in one study the mean weight dropped from 121.99 kg (pre-operative) to 77.07 kg (p< 0.001) after a mean follow-up of 8.2 years.96 Mean BMI dropped from 43.21 kg/m2 to 27.26 kg/m2 (p< 0.001).95 Randolph and coworkers97 reported a mean percentage weight loss of 32.75% at 12 months.

Other outcomes

Organ and coworkers95 discovered a significant change in attitude which reflected a greater sense of pride in the patient’s own body. This also led to changes in their eating habits with greater psychosocial adjustment. Patients’ involvement in social activities increased and many found themselves more employable. Randolph and coworkers97 reported a gradual decrease in appetite, and that patients appeared brighter, more alert and more outgoing.

Reported harms

Randolph and coworkers97 reported significant effects on levels of absorption, evidenced by flattening of the glucose and xylose tolerance tests, an increase in faecal losses of fat and nitrogen, and lower levels of serum triglycerides and cholesterol.

Other factors
Current medical conditions

One or several comorbidities were present including dyslipidaemia, sleep apnoea, pulmonary hypertension, low back pain and severe arthalgias, hypertension, liver steatosis, hypertriglyceridaemia, hypercholesterolaemia, peptic oesophagitis, cholelithiasis, degenerative joint disease (DJD), bronchial asthma, type 2 diabetes mellitus, urinary urgency and stress incontinence, dependent oedema, and gastro-oesophageal reflux disease.84,86,90,91 Rand and coworkers also reported anaemia and thyroid problems in seven patients. Breaux also reported sleep apnoea and brain-stem tumour.

In one case study,83 a 12-year-old girl had arterial hypertension, uremic odour, end-stage renal failure, renal anaemia, hypercalcaemia, metabolic acidosis, preserved diuresis with isosthenuria and renal osteopathy.

Organ and coworkers95 reported that no patients with the Prader–Willi or Laurence–Moon–-Biedl syndromes or other endocrinopathies were included. Any existing metabolic defect was stabilised prior to surgical intervention. Silber and coworkers96 reported three patients who died within a year of procedure (one with known incipient heart failure, one with established diabetes mellitus, who died of perinephric abscess and sepsis, and the third with hepatic disease). Two of these had Prader–Willi syndrome. Patients with Prader–Willi syndrome were also present in Randolph et al.’s case series,97 and Widhalm et al.’s study.82

Setting

All the retrieved studies were undertaken in university surgery departments.

Country

The vast majority of the studies were from the USA, with the exception of: Australia,80,81 Israel,79 Germany83 and Austria.82

Methodological and context notes

All the retrieved studies were of poor quality, with no RCTs or other types of controlled study. Moreover, levels of statistical significance were either not applicable or not provided. Therefore the reliability and validity of the results remain uncertain, and robustness of the evidence statements and recommendations on the use of bariatric surgery in adolescents is extremely weak.

15.2.6. Referral to specialist care for children and adolescents

In September 2005, the National Guideline Clearinghouse synthesised the recommendations on the assessment and treatment of obesity and overweight in children from six published guidelines.1

Only the guidelines from SIGN6 made detailed recommendations on the process of referral. These can be seen in Table 15.9. However, the authors noted that formal trials of the impact of different referral criteria are not easily carried out’, and the subsequent recommendations were based on an expert committee statement.

Table 15.9. Referral recommendations from the Scottish Intercollegiate Guidelines Network.

Table 15.9

Referral recommendations from the Scottish Intercollegiate Guidelines Network.

The NHMRC guidelines recommended that:

  • if a child presents with obesity in association with intellectual disability and multiple physical abnormalities, the child should be assessed by a paediatrician, an endocrinologist and/or a geneticist (level C)
  • an obese child or adolescent with height–growth failure should be referred to a paediatrician or an endocrinologist or both (level B)
  • conditions that cause hypothalamic obesity are rare and should be managed in a tertiary institution (level B).

Other recommendations were made around the use of very-low-energy diets, drugs and surgery only in specialist settings.

15.2.7. Harms arising in children and adolescents who undergo weight management/maintenance programmes

15.2.7.1. Evidence statements

(Table 15.10)

Table 15.10. Evidence statements and grading.

Table 15.10

Evidence statements and grading.

15.2.7.2. Evidence review on harms

Types of study

We only included studies that specifically assessed the effects on eating behaviours and child/adolescent growth of professionally prescribed weight loss programmes, with the following designs:

  • RCTs
  • controlled clinical trials
  • before-and-after studies
  • cohort studies
  • non-comparative studies (case series and case studies).

Again, we decided to adopt more expanded inclusion criteria due to the nature of the existing evidence. Studies that examined the effects of unsupervised dieting were not included in this review.

Update searches have been undertaken, however no studies were identified.

Types of participant

Obese and/or overweight children and adolescents.

Types of outcome
  • Primary outcomes to be (not self-reported) reports of eating disorders and growth variations.
Reports of harm in other guidelines

According to the National Guideline Clearinghouse Guideline synthesis,1 none of the included guidelines reported any harm that may arise if children or adolescents undergo professionally delivered weight management/maintenance programmes. Only the United Stated Preventive Services Task Force (USPSTF)98 tackled the topic of eating disorders, reporting one good-quality RCT in a primary care setting, where no problematic eating was detected in adolescents participating in a behavioural intervention treatment. The NHMRC2 recommended that ‘abnormal eating behaviours in childhood and adolescent obesity should be addressed both before and during weight-management programs’.

Effects of professionally prescribed weight loss programmes on eating behaviour

We found one recent systematic review99 that examined the effect of dieting on eating behaviour and psychosocial status. The findings of the five studies that looked at the effects of dieting on eating behaviours, suggested that professionally delivered weight loss interventions did not contribute to the development of eating disorders in overweight children and adolescents.

Braet and coworkers58 assessed a 10-month inpatient cognitive behaviour weight loss programme, where the participants’ median age and BMI was 14 years and 33 kg/m2, respectively. At post-treatment, scores on the Drive for Thinness subscale of the Eating Disorder Inventory (EDI) decreased significantly, and the number of participants scoring at least one standard deviation above the norm on the subscale (showing an increased risk for developing an eating disorder) decreased from 7 to 2 during the treatment. On the Dutch Eating Behaviour Questionnaire (DEBQ), scores on the external eating subscale decreased, and no significant changes were found on the Emotional Eating or Restrained Eating scales.58

Levine and coworkers100 assessed a family-based behaviour change programme in children with a mean age of 10.2 years and weight of 79.7 kg. Symptoms of eating disorders were measured at pre-treatment and follow-up by the Children’s Eating Attitudes Test (ChEAT), which showed a statistically significant decreasing trend at follow-up, suggesting that concerns with dieting, unhealthy dieting behaviours and concerns regarding being overweight tended to decrease.100

Epstein and coworkers101 examined an intervention in which all participants were given the traffic light diet, and some also received training in problem solving techniques. Participants had a mean age of 10.3 years and a weight of 59.5 kg. The Kids’ Eating Disorder Survey (KEDS) which assesses weight dissatisfaction, purging/restricting and total symptoms of disordered eating did not show any significant changes.101

Braet and Van Winckel57 evaluated a cognitive behaviour change programme that delivered self-regulation, problem solving techniques and promotion of lifestyle change to children with a mean age of 11 years and weight of 62 kg. This programme was delivered either in a group or to individuals or as summer camp component. The DEBQ assessment showed that external eating decreased and restrained eating increased during the 4.6 years. On the other hand, emotional eating did not change. At follow-up, 9% of the participants scored more than one standard deviation above the norm on the bulimia subscale of the EDI.57

Another study of Epstein and coworkers102 consisted of a 10-year follow-up of an earlier trial where all interventions were family based and included the traffic light diet. Participants had a mean age of 10.4 years and a weight of 55.3 kg. On a medical history form, 4% of the participants reported that they had been treated for bulimia nervosa during the 10 year follow-up, and none reported treatment for anorexia nervosa.102 However, the authors pointed out that this prevalence of bulimia was not high, as self-reported studies have shown an average prevalence for eating disorders of 9% in girls.

Effects of professionally prescribed weight loss programmes on child/adolescent growth

No recommendations were found in the National Guideline Clearinghouse Guideline synthesis or the NHMRC guidelines regarding the effects of professionally prescribed weight loss programmes on child or adolescent growth.

One study103 examined whether a multidisciplinary weight loss programme in adolescents with severe obesity allowed adequate growth and avoided lean mass loss. Participants were aged 9 to 17 years and had a mean BMI of 38.4 ± 8.0 kg/m2 for girls and 34.5 ± 3.2 kg/m2 in boys. Total lean mass (LM) did not vary and was positively correlated to pubertal development in both sexes before and after weight loss.103

Another study104 examined height velocities prior to and during weight reduction in 14 girls and 5 boys with a mean ± SD age of 8.5 ± 2.7, achieved by restricting caloric intake in all cases to two-thirds of the usual daily intake. Protein intake was maintained at 1.5–2.0 g/kg of initial body weight per day. A significant correlation between the change in z-scores of height velocity prior to and during weight reduction and the change in weight was observed. However, it should be noted that this was a poor-quality study, and biases may have occurred, leading to some misinterpretation of results.104

15.3. Adults

15.3.1. Factors to be considered in the clinical assessment of adults who are overweight or obese

[The aim of an initial assessment is to identify individuals who are at increased risk and who would benefit from intervention. This initial assessment should follow the classification of the degree of overweight or obesity as recommended by the GDG based on the earlier evidence reviews.

Therefore, the factors to be assessed at the initial presentation should be based on two evidence bases: one on the common comorbidities, and one on the effectiveness of weight loss in people with comorbidities and their expected health gain.

Further assessment(s) should aim to determine any determinants of energy imbalance.]

15.3.1.1. Evidence statements

(Table 15.11)

Table 15.11. Evidence statements and grading.

Table 15.11

Evidence statements and grading.

15.3.1.2. Evidence review on factors to be assessed in adults and mature adolescents

In February 2005, the National Guideline Clearinghouse synthesised the recommendations on the assessment and treatment of obesity and overweight in adults from six published guidelines.1 The different scopes, target populations, intended users, and practices covered can be seen in Appendix 15 (evidence tables of included studies). The authors of the synthesis identified areas of agreement between the included guidelines surrounding assessment. They concluded that:

‘AGA, BWH, and Singapore MOH recommend screening for comorbid conditions, particularly obesity-related health risks, as part of the medical evaluation. The presence or absence of such conditions is helpful in determining the intensity of therapy. ACP refers to the assessment of comorbid conditions as part of an algorithm that is provided for the suggested management of obesity. AGA and Singapore MOH also recommend screening for psychiatric disorders, such as depression and binge eating, which may affect the success of therapy. BWH points to the presence of depression, disinhibition, and binge eating at baseline as factors that increase the likelihood of weight regain after an initial weight loss.’1

No area of disagreement was noted for assessment.

The individual guideline recommendations on assessment are given in Table 15.12.

Table 15.12. Existing recommendations on assessment for obesity.

Table 15.12

Existing recommendations on assessment for obesity.

The National Health and Medical Research Council (Australia) (NHMRC) guidelines2 on the management of overweight and obesity in adults recommended that after discussing weight with the individual and whether measurements should be taken, the next steps were to assess and treat associated comorbidities (specifically to measure blood pressure, plasma cholesterol, lipids and glucose) and to determine the individual’s need to lose weight. The decisions on which factors to assess were made on the identified common comorbidities (see Section 1 Chapter 3 section 3.1) and also on the evidence for the benefit of weight loss in individuals with these conditions. Diseases and conditions associated with obesity (see Table 15.13) were listed and further categorised into two groups: those with indirect links (as a result of metabolic consequences) and those with direct links (as a result of the excess weight).

Table 15.13. Diseases and conditions associated with obesity.

Table 15.13

Diseases and conditions associated with obesity.

The Agency for Healthcare Research and Quality (AHRQ) report on obesity in the elderly3 concluded that:

‘Those at risk for obesity-associated health problems stand to benefit most from intervention, if such intervention alters their weight-related risk. The strongest evidence for obesity intervention is for those with cardiovascular risk. Cardiovascular risk factors – including family history, diabetes, tobacco use, or dyslipidemia – can help identify this group.’3

Although no recommendations were made on this evidence, it seems logical that an initial assessment should focus on identifying those who have most to gain.

15.3.2. Energy imbalance in adults and mature adolescents

The NHMRC guidelines2 considered that the reasons (how and why) for energy imbalance should be assessed. The mechanism for energy imbalance is the imbalance between food intake (total energy and the energy per unit weight of food) and the energy expenditure, but the evidence suggested that food intake and levels of physical activity could only be estimated approximately in a clinical (or non-specialist) setting.

Reasons for why this imbalance should occur were categorised into six key areas: environment, genes, stress and psychological factors, current medication, life stage (early childhood and adolescence, pregnancy and childbirth, menopause) and life events (quitting smoking, marriage, giving up sport, holidays). The authors made evidence statements as follows:

  • The modern environment is a potent stimulus for obesity.
  • Some rare cases of single-gene mutations cause severe obesity disorders, which usually manifest early in life.
  • In general, cases of severe obesity are more likely to have a specific genetic basis than cases of overweight, which may result from environmental influences alone.
  • Psychological stress can have variable effects on a person’s body weight.
  • Several prescription medications can cause weight gain.
  • Obesity in childhood and adolescence is a risk factor for obesity later in life.
  • The tracking of childhood obesity into adult obesity is stronger for older children than for younger ones.
  • Pregnancy and menopause are critical periods for weight gain in women.
  • It appears that a change in weight at menopause can be prevented by lifestyle change.
  • Hormone replacement therapy after menopause can result in reduced body fat gain (particularly on the upper body) when compared with a placebo.
  • Certain life events – for example, marriage, holidays, and giving up sport – can have an influence on body fatness.
  • Quitting smoking can cause significant weight gain – on average 5–6 kg in the first year.
  • Lack of motivation and a history of failed attempts to lose weight may make it more difficult to maintain a low body weight.
  • Psychological factors – including early life experiences – can play an important part in the development of overweight or obesity.2

15.3.3. Lifestyle interventions (diet, behaviour therapy and physical activity) for weight loss and other outcomes in adults

All summary statistics, other than those presented in the figures, can be found in Appendix 17. Please note that all summary statistics have been checked by a consultant statistician.

15.3.3.1. Evidence statements – diet

(Table 15.14)

Table 15.14. Evidence statements and grading.

Table 15.14

Evidence statements and grading.

15.3.3.2. Evidence review on dietary interventions

This review is primarily based on a health technology appraisal (HTA) published in 2004.4 The aim of the HTA was to review systematically obesity treatments in adults to identify therapies that impact by achieving weight reduction, risk factor modification or improved clinical outcomes. All randomised controlled trials (RCTs) of dietary interventions in adults with a body mass index (BMI) of 28 or more were included. The duration of the trials had to be for 52 weeks or more. The main outcome was weight change in kilograms at 12 months’ follow-up.

The diets were classified as follows:

Due to reporting issues, healthy eating advice and 600 kcal/day deficit or low-fat diets were classified together, along with diets where the fat or calorie restriction was not stated or could not be estimated.

We used the definitions as above when classifying diets. Because of some concerns about the definitions, we have tried to be explicit (that is, include as much detail as possible about the dietary content) in both the evidence tables and the evidence statements.

Other outcomes included longer-term weight loss and blood pressure, lipids and fasting blood glucose. However, these outcomes were reported in only some of the included trials so the results are based on more limited evidence (often only the results of one trial) than that of weight change at 12 months so caution should be exercised when interpreting these.

Another review is cited which addressed the diagnosis and treatment of obesity in older people (defined as aged 60 years or more), a technology assessment published in 2003 by the US Agency for Healthcare Research.3

The aim of the review was to reassess the evidence for the diagnosis and treatment of obesity in older people. The key clinical question relevant to this evidence review is ‘Are there dietary or behavioural therapies that improve net health outcomes in obese older people?’ The defined inclusion criteria for dietary interventions were:

  • RCT of fair or good quality
  • weight loss or reduction in BMI, waist circumference, waist-to-hip ratio as a reported outcome
  • BMI of 25 kg/m2 or more
  • minimum 12-month follow-up
  • population generalisable to a typical US primary care population
  • sample mean age of 60 years or more.

Only one trial was reported as evaluating the effectiveness of a dietary intervention only (that is, not combined with any other intervention such as behavioural treatment) and satisfied the criteria above. However, when we examined this trial in detail, the intervention was not diet alone. See excluded studies (Appendix 16). No additional studies were identified from the Update searches.

600 kcal/day deficit or low-fat diet***** compared with usual care

A total of 13 RCTs were included in the Avenell HTA.4 No additional studies were identified for this comparison. One study5 was conducted in a workplace setting, but was included as the setting was not the focus of the study. One HTA included study was excluded6–8 in this review because not all participants were overweight.

Weight loss

Overall, there was an significant weight change of −5.32 kg at 12 months in this group (weighted mean difference [WMD] 95% confidence interval (CI) −5.88 to −4.75)††††† when compared with a control group, and although there was statistical heterogeneity (p ≤ 0.00001, I2 = 76.7%) between the 12 included studies the direction of effect was consistent.4 (Figure 15.1). One cluster randomised controlled trial (C-RCT) reported a mean ± SD weight change at 12 months of −0.88 ± 4.0 kg for the diet group and 1.3 ± 3.0 kg for the control group (not statistically significant).

Figure 15.1. Maintenance of weight loss over time for all adults on a low-fat or 600 kcal/day deficit diet compared with usual care.

Figure 15.1

Maintenance of weight loss over time for all adults on a low-fat or 600 kcal/day deficit diet compared with usual care.

Other outcomes

At 12 months, there were significant improvements in blood pressure (change in diastolic blood pressure [DBP] −3.44 mm Hg, WMD 95% CI −4.86 to −2.01 based on four trial results; change in systolic blood pressure [SPB] −3.78 mm Hg, WMD 95% CI −5.53 to −2.03 based on four trial results), lipids (change in total cholesterol −0.21, WMD 95% CI −0.34 to −0.08 based on four trial results), and fasting plasma glucose (change in fasting plasma glucose −0.28 mmol/l, WMD 95% CI −0.47 to −0.09 based on one trial) compared with usual care. But these were not maintained over time, even in obese populations with hypertension or type 2 diabetes. In an obese population with hypertension, changes in SBP and DBP in the diet group compared with the control group were not significant at 18 months (one study only). The data suggested that a low-fat diet or 600 kcal/day deficit diet prevented the development of diabetes and improved blood pressure control.4

Other factors

No additional analysis was carried out on the effect of age, ethnicity, socioeconomic status, previous treatment for obesity, motivation, frequency or length of the intervention in the HTA.4

Gender

Three studies recruited men only,5;9;10 and a further four studies11–14 had mainly male participants. No studies recruited women only.

The Dietary Intervention Study of Hypertension trial (DISH) showed no difference between men and women for weight loss (−11.0 kg vs −9.7 kg respectively, 45.8% and 46.8% lost 5% or more of initial body weight, respectively) at 56 weeks. Similarly, gender was not significantly associated with weight loss at 6 months in the Trial of Antihypertensive Interventions and Management (TAIM) study.14

Age

The age range of participants varied (overall, range 20 years minimum to 80 years maximum, where reported). Some studies recruited younger adults (for example, aged 25–49 years), whereas others recruited older people (for example, aged 50–80 years).

DISH showed no difference between people aged under 60 years and those aged 60 years or over for weight loss (−11.0 kg vs −8.5 kg, respectively, 46.7% and 45.4% lost 5% or more of initial body weight, respectively) at 56 weeks.

Other groups, including black and minority ethnic groups

Although the results of the DISH trial were analysed for groups of black and white participants, we have not reported the results here as the study was based in the USA. Therefore it may be less applicable to the UK population.

Current medical conditions

Of the included studies, four studies recruited only people with hypertension,14–17 one trial included people with ‘high normal’ blood pressure,12 one trial recruited people with glucose intolerance including some with diabetes, one trial recruited people with insulin resistance,13 one trial recruited people with one or more risk factor,5 and one trial recruited people who had had myocardial infarction (who also received exercise).11 The remainder recruited otherwise healthy participants.

Setting and/or healthcare professional

One study reported that there was no difference between the effect of diet counselling delivered by the doctor and the dietitian compared with the dietitian alone. The cost of 1 kg weight loss was less for the dietitian-alone group (Aus$7.30 vs Aus$9.76).18

Low-calorie diet ( 1000–1600 kcal/day) compared with usual care

Only one study was included for this comparison in the HTA review.19 No additional studies were identified.

Weight loss

One study (two trials based in Poland and the Netherlands) reported a weight change at 12 months of −6.25 kg (WMD 95% CI −9.05 to −3.45) compared with control. This weight loss was maintained over 36 months of follow-up.19

Figure 15.2. Maintenance of weight loss over time for women with breast cancer on a low-calorie diet (LCD) compared with usual care (results for Netherlands only).

Figure 15.2

Maintenance of weight loss over time for women with breast cancer on a low-calorie diet (LCD) compared with usual care (results for Netherlands only).

Other outcomes

No other data on change in risk factors were reported.

Other factors

This trial was a single feasibility study of 107 women who were obese (BMI of 27 kg/m2 or more), postmenopausal and who had undergone primary treatment for breast cancer with no signs of distant metastases.19

VLCD (< 1000 kcal/day) compared with usual care

Only one study was included for this comparison in the HTA review.20 No additional studies were identified.

Weight loss

This trial reported results of a VLCD in obese participants with asthma. At 12 months, the weight change was −13.40 kg (WMD 95% CI −18.43 to −8.37) compared with control.20

Figure 15.3. Maintenance of weight loss over time for people with asthma on a very-low-calorie diet (VLCD) compared with usual care.

Figure 15.3

Maintenance of weight loss over time for people with asthma on a very-low-calorie diet (VLCD) compared with usual care.

Other outcomes

Two of the participants found the liquid diet intolerable and followed an alternative low-energy diet. No other outcomes (other than those associated with asthma – see below) were reported.

Other factors
Current medical conditions

The trial reported improvements in forced expiratory volume, forced vital capacity, exacerbations of asthma, the use of rescue medications and steroids.20

Low-calorie diet compared with 600 kcal/day deficit or low-fat diet

Only one study was included for this comparison in the HTA review.21;22 One additional study was identified.23

Weight loss

At 12 months, the low-calorie diet was associated with a summary estimate of weight change of 0.78 kg (WMD 95% CI −1.06 to 2.63, based on two trials).

Other outcomes and factors

None were reported in Shah and coworkers’ trial.21 Mean metabolic cardiac risk factor levels and blood pressure measurements improved from baseline in the low-calorie diet group (significant changes were seen for triglycerides, low-density lipoprotein [LDL], high-density lipoprotein [HDL] levels, p ≤ 0.05), but the changes were not statistically significant at 12 months compared with the low-fat group.23

Age and gender

Shah and coworkers21 recruited only women aged 25–45 years, who were otherwise healthy. Dansinger and coworkers23 recruited both men and women of any age, who had at least one identified risk factor.

VLCD compared with 600 kcal/day deficit or low-fat diet

Only one study was included for this comparison.24 No additional studies were identified.

Weight loss

No data were reported for weight change at 12 months.

Simonen and coworkers24 compared a VLCD (PSMF VLCD) to a low-fat diet in people with type 2 diabetes. At 24 months, the VLCD produced a weight change of −4.70 kg compared with the low-fat diet (WMD 95% CI −11.79 to 2.39).

Other outcomes and factors

Although this trial was undertaken in a population of people with type 2 diabetes and other clinical outcomes were reported, the authors of the review did not present these results due to concern over significant baseline differences between the groups in rates of treatment of diabetes and HbA1c levels.

VLCD compared with low-calorie diet

Three studies were included for this comparison in the HTA review.25–27 No additional studies were identified. One study27 was conducted in a workplace setting, but was included as the setting was not the focus of the study.

Also, two trials25;26 compared the use of a VLCD for a short period of time, in conjunction with a low-calorie diet, to continuous use of a low-calorie diet.

Weight loss

At 12 months, based on three trials, VLCD was associated with a weight change of −0.15 kg compared with the low-calorie diet (WMD 95%CI −2.73 to 2.43). No significant effect was seen at 18 months (change of −1.13 kg, 95% CI −5.32 to 3.06, based on results from one trial only).27

Figure 15.4. Maintenance of weight loss over time for people on a very low-calorie diet (VCLD) compared with a low-calorie diet (LCD).

Figure 15.4

Maintenance of weight loss over time for people on a very low-calorie diet (VCLD) compared with a low-calorie diet (LCD).

Other outcomes

No other outcomes were reported.

Other factors

No additional analysis was carried out on the effect of age, ethnicity, socioeconomic status, degree of obesity, current medical conditions, previous treatment for obesity, motivation or the setting or delivery of care.

Gender

One study recruited men (aged 26–52 years) only27 and one recruited women (aged 21–59 years) only.25

Low-fat diet compared with another weight reducing diet

The HTA did not examine the comparisons of low-fat and low-calorie diets, where the aim of the trials was to evaluate the different types of diet using the same calorie value. The authors cited a Cochrane review28 published in 2002. The results of trials which met our defined inclusion criteria are presented below). No additional studies were identified.

One study29 was conducted in diet clubs and workplaces, but was included as the setting was not the focus of the study.

Weight loss

At 12 months, the overall weight change was 0.49 kg (WMD 95% CI −1.14 to 2.11 based on five trial results), again with significant heterogeneity (p = 0.03, I2 = 62.6%).

Figure 15.5. Maintenance of weight loss over time for people on a low-fat diet compared with another weight reducing diet.

Figure 15.5

Maintenance of weight loss over time for people on a low-fat diet compared with another weight reducing diet.

Other outcomes

Only one trial30 showed any significant effect on total cholesterol for women with a family history of diabetes. The other trials showing no significant differences on serum lipids, blood pressure or fasting blood glucose.29;31;32

Other factors
Gender and current medical conditions

One trial recruited women with a family history of diabetes or a diagnosis of non-insulin-dependent diabetes mellitus (NIDDM);30 the remaining trials recruited mainly women, but who were otherwise healthy.29;31;32

A subgroup analysis found that a low-carbohydrate/low-fibre diet tended to be more successful for weight loss among women than a higher-carbohydrate/higher-fibre diet.29

Age

A subgroup analysis found that a low-carbohydrate/low-fibre diet tended to be more successful for weight loss among younger people (aged 40 years or less) than a higher-carbohydrate/higher-fibre diet.29

Social class

A subgroup analysis found that a low-carbohydrate/low-fibre diet tended to be more successful for weight loss among people in a lower social class (classes III–IV) than a higher-carbohydrate/higher-fibre diet.29

PSFM‡‡‡‡ compared with 600 kcal/deficit or low-fat diet

Three trials have been published since the HTA4 that evaluated the use of PSMF (caloried content approx 1400–1900kcal/day, food based)compared with a 600 kcal/deficit or low-fat diet in people who were overweight.23;33–35

Weight loss

At 12 months, the overall summary estimate of weight change was −0.56 kg (WMD 95% CI −2.17 to 1.04 based on three trial results) for the PSMF (not a VLCD) diet compared with a low-fat diet.

Figure 15.6. Maintenance of weight loss over time for people on a protein-sparing modified fast (PSMF) compared with a 600 kcal/deficit or low-fat diet.

Figure 15.6

Maintenance of weight loss over time for people on a protein-sparing modified fast (PSMF) compared with a 600 kcal/deficit or low-fat diet.

Other outcomes

Summary estimates at 12 months showed no significant difference between a PSMF diet and low-fat diet, apart from a significant increase in HDL levels (0.08 mmol/l, p = 0.001) on the PSMF. However, lower levels of adherence to the low-carbohydrate diet suggested that a low-fat diet may be more sustainable in the long term.23

Fasting plasma glucose was reported in both the Dansinger23 and Stern studies.33 The Stern paper did not report results overall but split into people who had diabetes and those who did not. Therefore, we were not able to estimate a summary effect, but results from the Dansinger trial23 and subgroups of people with or without diabetes in the Stern trial33 suggested that there was no significant difference between the diets.

Other factors

No additional analysis was done.

PSMF§§§§ compared with low-calorie diet

Seven RCTs were included in the Avenell HTA.4 One additional study was identified for this comparison.23

Weight loss

The overall weight change associated with a PSMF compared with a low-calorie diet at 12 months was −0.62 kg (WMD 95% CI −2.35 to 1.11 based on four trial results). No further significant effect was seen at 18******, 24, 36 and 60 months. However, there was considerable range in the calorie content of each of the diets; three studies used approx 400–500kcal/day,36–38 compared with one assumed to be much higher.23

Figure 15.7. Maintenance of weight loss over time for people on a protein-sparing modified fast (PSMF) compared with a low-calorie diet (LCD).

Figure 15.7

Maintenance of weight loss over time for people on a protein-sparing modified fast (PSMF) compared with a low-calorie diet (LCD).

Other outcomes

Dansinger and coworkers23 showed no significant differences between PSMF and a low-calorie diet for changes in cholesterol, LDL or HDL levels, triglycerides, or blood pressure measurements at 12 months.

Other factors
Gender

Of the included trials, two recruited men only27;39 and one recruited women only.36 One trial recruited men, but then excluded them from the analysis due to small numbers.37

Torgerson and coworkers40;41 did an intention to treat (ITT) analysis of results by men and women. At 24 months, the mean weight change in men was −13.0 ± 15.6 kg ) in the PSMF group and −5.8 ± 9.7 kg in the low-calorie diet group (p = 0.1, non-significant). At 24 months, the mean weight change in women was −6.8 ± 10.6 kg) in the PSMF group and −6.4 ± 8.4 kg) in the low-calorie diet group (non-significant). Similarly at 48 months, there was no statistically significant difference between the weight change in men and women who completed the study.

Wing and coworkers38 found that women in the PSMF group lost significantly more weight at 12 months than did the low-calorie diet participants (14.1 kg vs 8.6 kg, p < 0.023). Men had comparable losses between the groups (15.4 kg and 15.5 kg). The percentage of women losing 15 kg or more or at least 5 BMI units was significantly greater in the PSMF group (p < 0.01 for both comparisons). The proportion of men achieving these two goals was similar in each group.

Current medical conditions

Most studies recruited otherwise healthy participants, but Dansinger and coworkers recruited people with at least one risk factor,23 Wing and coworkerspeople with NIDDM,38 and Wing and coworkers39 people with type 2 diabetes.

PSMF†††††† compared with VLCD

Both of the diets being compared in this study27 could be classed as PSMF, but with different daily caloric values (420kcal/day vs 1000kcal/day).

Weight loss

No data for weight loss at 12 months were reported.

The summary estimate weight change at 18 months was 1.56 kg for the PSMF compared with a very-low-calorie diet (WMD 95% CI −1.57 to 4.69, based on results from one trial with multiple treatments).27

Other outcomes and factors

None were reported.

Low-fat diet compared with very-low-fat diet

One study was identified that compared a low-fat diet (Zone) and a very-low-fat diet (Ornish).23

Weight loss

There was an overall weight change at 12 months of 0.10 kg in the low-fat group compared with the very-low-fat group (95% CI −2.83 to 3.03).23

Figure 15.8. Maintenance of weight loss over time for people on a low-fat diet compared with a very-low-fat diet.

Figure 15.8

Maintenance of weight loss over time for people on a low-fat diet compared with a very-low-fat diet.

Other outcomes and factors

There were no significant changes in lipids and blood pressure between the groups. Although HDL levels increased in the low-fat group and decreased in the very-low-fat group (0.08 mmol/l vs −0.01 mmol/l), and DBP decreased in the lowfat group compared with an increase in the very-low-fat group (−1.20 mm Hg vs 0.20 mm Hg), these changes were not significantly different between the groups (p = 0.07 and p = 0.40, respectively).

Low-calorie diet compared with very-low-fat diet
Weight loss

One study published since the HTA found an overall weight change at 12 months of 0.30 kg in the low-calorie diet group compared with very-low-fat group (95% CI −2.42 to 3.02).23

Figure 15.9. Maintenance of weight loss over time for people on a low-calorie diet compared with a very-low-fat diet.

Figure 15.9

Maintenance of weight loss over time for people on a low-calorie diet compared with a very-low-fat diet.

Other outcomes and factors

Changes in lipids and blood pressure were not significantly different between the groups, other than HDL levels which increased in the low-calorie diet group and decreased in the very-low-fat group (0.09 mmol/l vs −0.01 mmol/l, p = 0.04).

PSMF‡‡‡‡‡ compared with very-low-fat diet
Weight loss

One study published since the HTA found an overall weight change at 12 months of −1.20 kg in the PSMF (food based, low carb) group compared with the very-low-fat group (95% CI −1.51 to 3.91).23

Figure 15.10. Maintenance of weight loss over time for people on a protein-sparing modified fast (PSMF) compared with a very-low-fat diet.

Figure 15.10

Maintenance of weight loss over time for people on a protein-sparing modified fast (PSMF) compared with a very-low-fat diet.

Other outcomes and factors

Changes in lipids and blood pressure were not significantly different between the groups, other than HDL levels which increased in the PSMF group and decreased in the very-low-fat group (0.09 mmol/l vs −0.01 mmol/l, p = 0.01).

Low glycaemic index (high-protein) diet compared with high glycaemic index (standard-protein) diet

One Cochrane review was identified that evaluated the effect of low glycaemic diets on coronary heart disease and other risk factors (including weight). It was not clear whether participants in the included trials were overweight, so the review was excluded.42

Two trials were found that compared the effectiveness of a diet high in protein compared with a diet high in carbohydrate and lower in protein.43;44

Weight loss

The summary estimate weight change at 12 months was −1.90 kg for the high-protein (low glycaemic [GI]) diet compared with a standard/medium-protein (high GI) diet (WMD 95% CI −6.45 to 2.65, based on results from one trial).44

Figure 15.11. Maintenance of weight loss over time for people on a high-protein (HP) diet compared with a standard/medium-protein (S-MP) diet.

Figure 15.11

Maintenance of weight loss over time for people on a high-protein (HP) diet compared with a standard/medium-protein (S-MP) diet.

Other outcomes and factors

None were reported.

Brinkworth and coworkers43;45 recruited mainly women aged 20–65 years with hyperinsulinaemia. Due and coworkers44 again recruited mainly women, but those who were otherwise healthy and aged 18–56 years of age.

15.3.3.3. Evidence statements – behaviour therapy (with or with diet)

(Table 15.15)

Table 15.15. Evidence statements and grading.

Table 15.15

Evidence statements and grading.

15.3.3.4. Evidence review on behaviour therapy (with or without diet)

This review was primarily based on four key reviews.4;46–48 A comparison of the reviews can be seen in Table 15.16. Additional searching was also done to identify any other RCTs published since these key reviews were published. Reference lists of other reviews were also cross-referenced.49–51

Table 15.16. Comparison of systematic reviews on behaviour therapy for weight loss in adults.

Table 15.16

Comparison of systematic reviews on behaviour therapy for weight loss in adults.

For this evidence review, only RCTs with a duration (including follow-up) of 12 months or more were included. Also, mean BMI of participants had to be 28 or over.

On the advice of our co-opted expert, we only included specific techniques as being behavioural treatment. Such techniques were defined as:

  • drawing on the principles of learning theory: stimulus–behaviour contingencies or behaviour–reward contingencies
  • assessment consisting of identifying and specifying problem behaviours and the circumstances in which they are elicited (both antecedents and consequences)
  • treatment starting with setting specific, measurable and modest goals that are continually revised as progress is achieved
  • target behaviours being monitored – usually by self-monitoring – to obtain a record of behaviour change
  • behaviour change processes including stimulus control, graded exposure, extinction and reward
  • having a perspective of educational: teaching behaviour change skills to the client. The term problem solving skills may be used, but this does not necessarily mean that the treatment contains the other elements of conventional behavioural treatment
  • the use of the term cognitive (as in ‘cognitive behaviour’ therapy or CBT) may imply the inclusion of strategies designed to modify cognitions (thoughts) which can be identified as important stimuli for behaviour.

We identified a treatment as being behavioural if the study paper:

  • used the terms behavioural treatment, cognitive behavioural treatment or behaviour therapy, CBT
  • mentioned learning theory
  • referred to the use of the common components of behavioural treatment (self-monitoring, goal-setting, stimulus control).

Study reports that used the following terms alone were excluded as not meeting our criteria for behavioural treatment

  • motivational interviewing
  • counselling
  • learning
  • psychological
  • psychotherapy
  • problem solving
  • cognitive.
Diet and behaviour therapy versus usual care

One trial included in the HTA was excluded in this evidence review because there was no requirement that participants in the study were overweight or obese.52 Hakalaand Karvetti53 and Karvetti and Hakala54 were also excluded as although the participants received counselling or lectures from a psychologist, no details of the behavioural techniques used were reported. Wing and coworkers55 compared an active intervention on diet and behaviour therapy with passive information on diet and behaviour therapy, so was moved to the appropriate section.

One additional RCT was found comparing diet and defined (or named) behaviour therapy with usual care.56

Update searches identified one relevant study,57 comparing diet and a cognitive behavioural programme with a waiting list control. The long term results (18 months after completion of the programme) support the evidence already reviewed in detail.

Weight loss

At 12 months, a combination of diet (no specific details other than ‘balanced, fat-reduced nutrition’) and behaviour therapy (self-monitoring, strategies to control eating behaviour, problem analysis and self-observation, alteration of cognitive patterns, social competence, relapse prevention) was associated with a summary estimate of weight change of −3.10 kg (WMD 95% CI −7.17 to 0.97, based on one comparison) compared with usual care.

Figure 15.12. Maintenance of weight loss over time for diet and behaviour therapy (BT) compared with usual care.

Figure 15.12

Maintenance of weight loss over time for diet and behaviour therapy (BT) compared with usual care.

Other outcomes

No outcomes (such as lipids or blood pressure) were reported. But improvements in psychological outcomes (such as control and attractiveness) were seen in the diet and behaviour therapy group compared with the control group.

Other factors
Gender and setting

Both men and women (mainly women) were recruited from referrals and from adverts. The trial was conducted in a primary care setting, by general practitioners (GPs) and tutors who had undergone specific training to deliver the programme of diet and behaviour therapy.

Diet and behaviour therapy versus information

One RCT was found comparing diet and defined (or named) behaviour therapy with a healthy lifestyle intervention.58

Weight loss

At 12 months, a combination of diet (classified as calorie deficit) and behaviour therapy (self-monitoring, goal setting, cognitive restructuring, problem solving, and environmental management) was associated with a summary estimate of weight change of −3.51 kg (WMD 95% CI −5.60 to−1.42, based on one comparison) compared with information.

Other outcomes

No outcomes (such as lipids or blood pressure) were reported. No physiological outcomes (related to arthritis) were found to be significantly different between the groups.

Other factors
Participants and current medical conditions

Both older men and women (mainly women) with ostearthritis were recruited from adverts and publicity.

Active diet, and behaviour therapy versus passive (information based) diet and behaviour therapy

Two trials were included in this comparison.55;59 Cousins and coworkers58 compared the use of behaviour therapy and a low-calorie diet with information on diet and behaviour, and Wing and coworkers55 used an VLCD in conjunction with behavioural techniques compared to self-help (diet and behaviour therapy).

Weight loss

At 12 months, the active diet and behaviour therapy (problem solving, relapse prevention, stimulus control, dealing with problem situations, assertion, behaviour chain analysis) was associated with a summary estimate of weight change of −3.73 kg (WMD 95% CI −5.77 to −1.70, based on three comparisons) compared with passive information. However, if the different diets were considered separately, only the VLCD (800–1000kcal/day) remained significantly more effective (weight change of −5.20 kg, 95% CI WMD 95% CI −8.07 to −2.33).

Figure 15.13. Maintenance of weight loss over time for active diet and behaviour therapy (BT) compared with passive diet and BT.

Figure 15.13

Maintenance of weight loss over time for active diet and behaviour therapy (BT) compared with passive diet and BT.

Other outcomes

Only Wing and coworkers reported any outcomes other than weight. No significant changes were seen in any of the reported outcomes (lipids, blood pressure, triglycerides, fasting plasma glucose, %HbA1C) at 12 or 24 months, other than for total cholesterol (a change of −0.30 mmol/l WMD 95% CI −0.58 to −0.02) at 24 months (one trial).55

Other factors

Cousins and coworkers recruited women participants aged 18–45 years only,59 and Wing and coworkers recruited mainly older women (inclusion criteria 40–55 years).55

Context and methodological notes

The Cousins study was included in the diet, activity, and behaviour therapy versus control section of the HTA. No details of the physical activity were reported (other than ‘group exercise’), and the control group received a manual of diet and behaviour therapy strategies, so this study has been reclassified. Wing and coworkers55 compared an active diet and behaviour therapy intervention with passive information on diet and behaviour therapy, so was moved to the appropriate section.

Family versus individual treatment

Seven trials were included in this comparison.59–65

Weight loss

At 12 months, the family-based intervention was associated with a summary estimate of weight change of −2.96 kg (WMD 95% CI −5.32 to −0.60, based on five comparisons) compared with individual intervention. However, if the different interventions were considered separately, only the intervention involving spouses (with low-calorie diet, behaviour therapy and physical activity if no weight loss) remained significantly more effective (weight change of −6.09 kg, 95% CI WMD 95% CI −10.64 to −1.54).

Figure 15.14. Maintenance of weight loss over time for family compared with individual intervention.

Figure 15.14

Maintenance of weight loss over time for family compared with individual intervention.

Other outcomes

Only Wing and coworkers64 reported clinical outcomes other than weight. No changes were significant between interventions at any time points.

Other factors
Gender

Four of the trials recruited women only,59;60;62;63 and overall the majority of participants were women.

Delivery

One of the trials recruited support from people other than family.64 However, no details of the members and relationships of the support group were reported. All interventions were delivered to family and individual in the same sessions.59–65

One trial64 noted a significant interaction of gender and treatment, such that women did better when treated with a spouse and men did better when treated individually.

Methodological and context notes

Slightly different figures from the HTA have been used for Black and Lantz59 due to conversion/calculations rounding.

Also, there were some differences between 12 month follow-up. For example, Cousins and coworkers58 assessed at 12 months from beginning of intervention, Black and Lantz59 and Murphy and coworkers60 assessed at 12 months after the end of the intervention which lasted 10 weeks. Also Wing and Jeffery64 used 18 months’ figures.

Group versus individual treatment

Four trials were included in this comparison.66–69

Weight loss

At 12 months, the group-based intervention was associated with a summary estimate of weight change of 1.59 kg (WMD 95% CI −1.81 to 5.00, based on four comparisons) compared to individual intervention. However, at 24 months, one trial66 showed a significantly greater weight loss in the individual group compared with the group intervention.

Figure 15.15. Maintenance of weight loss over time for group compared with individual intervention.

Figure 15.15

Maintenance of weight loss over time for group compared with individual intervention.

Other outcomes

No clinical outcomes were reported. One study66 reported higher attendance rates in the individual groups than in the group sessions.

Other factors
Gender

Three of the four trials recruited women only.67–69 Again, the majority of participants were women.

Degree of obesity

Hakala and coworkers recruited men and women who were severely overweight (at least 50% overweight).66 The other trials recruited women who were overweight (different criteria used) but not severely obese specifically.

Setting

Two of the studies were set in the NHS.66,67 None of the results were significant.

Diet and behaviour therapy versus diet

Phenix70 was excluded from this review, as this was a PhD thesis and no subsequent published papers were identified.

Three trials were included in this comparison.36;67;68 All compared the use of diets (low-calorie diet,66,67 PSMF37) in conjunction with behaviour therapy compared to the use of the diet alone.

Weight loss

At 12 months, a combination of diet and behaviour therapy (cue avoidance, self-monitoring, stimulus control, slowing rate of eating, social support, planning, problem solving, assertiveness, cognitive restructuring, modifying thoughts, reinforcement of changes, relapse prevention, strategies for dealing with weight gain) was associated with a summary estimate of weight change of −7.66 kg (WMD 95% CI −11.96 to −3.36, based on two comparisons) compared with diet alone. However, if the different diets were considered separately, only the PSMF (food based 400–500 kcal/day) remained significantly more effective (weight change of −8.19 kg, 95% CI WMD 95% CI −13.64 to −2.74).

Figure 15.16. Maintenance of weight loss over time for diet and behaviour therapy (BT) compared with diet alone.

Figure 15.16

Maintenance of weight loss over time for diet and behaviour therapy (BT) compared with diet alone.

Other outcomes

No other outcomes were reported.

Other factors
Gender

All trials in this section recruited women only.

Setting

Jones and coworkers67 and Long and coworkers68 were conducted in the NHS. Mixed results were seen, but a combination of group behaviour therapy and diet seemed to work better than individual behaviour therapy and diet at 6 months, although this was not statistically significant.

Different levels of intensity of behaviour therapy and diet

One study was identified that compared different levels of intensity of the same behavioural treatment programme in conjunction with periods of a VLCD.

Update searches identified a further study, showing that weight loss could be achieved even with relatively minimal follow-up contact (twice a year).71

Weight loss

At 12 months, a combination of VLCD and intensive behaviour therapy (self-monitoring, relapse situations, eating behaviour) was associated with a summary estimate of weight change of −1.18 kg (WMD 95% CI −4.16 to 1.80, based on one comparison) compared to a combination of VLCD and less intensive behaviour therapy.

Figure 15.17. Maintenance of weight loss over time for diet and intensive behaviour therapy (BT) compared with diet and less intensive BT.

Figure 15.17

Maintenance of weight loss over time for diet and intensive behaviour therapy (BT) compared with diet and less intensive BT.

Other outcomes

Only significant differences were seen for blood pressure (both SBP and DBP) at 24 months, with the more intensive programme appearing to be more effective.

Other factors
Gender and setting

Both men and women (mainly women) were recruited from referrals to a secondary care obesity clinic.

Comparison of different behavioural treatments

One study was identified that compared different types of behaviour therapy.72

Update searches identified one pilot study comparing different approaches to behavioural treatments.73 Weight changes were similar, and therefore would not change either evidence statements or recommendations.

Weight loss

At 17 months, a combination of 20 weeks’ behaviour therapy (self-monitoring, goal setting, stimulus control) with a low-calorie diet and physical activity followed by 12 months of relapse prevention training was associated with a summary estimate of weight change of +4.97 kg (WMD 95% CI 0.46 to 9.48, based on one comparison) compared with a combination of the 20 weeks’ programme followed by 12 months of group problem solving. However, both groups did lose weight from baseline.

Figure 15.18. Maintenance of weight loss over time for behaviour therapy and relapse prevention maintenance compared with behaviour therapy and problem solving.

Figure 15.18

Maintenance of weight loss over time for behaviour therapy and relapse prevention maintenance compared with behaviour therapy and problem solving.

Other outcomes

No other outcomes were significant.

Other factors
Gender and setting

In both trials, men and women (mainly women) were recruited from adverts. The setting was unclear.

15.3.3.5. Evidence statements – physical activity (alone or in combination with diet or behaviour therapy)

(Table 15.17)

Table 15.17. Evidence statements and grading.

Table 15.17

Evidence statements and grading.

15.3.3.6. Evidence review on physical activity (alone or in combination with diet or behaviour therapy)

This review was primarily based on three key reviews.4;48;74 A comparison of the reviews can be seen in Table 15.18. Although the AHRQ report did not specifically evaluate physical activity, some of the included RCTs included a physical activity component. Additional searching was also done to identify any other RCTs published since these key reviews were published. Reference lists of other reviews were also cross-referenced.

Table 15.18. Comparison of systematic reviews on physical activity (alone or in combination with diet or behaviour therapy) for weight loss in adults.

Table 15.18

Comparison of systematic reviews on physical activity (alone or in combination with diet or behaviour therapy) for weight loss in adults.

For this evidence review, only trials with a duration (including follow-up) of 12 months or more were included. Also, mean BMI of participants had to be 28 kg/m2 or over.

Physical activity versus control (no treatment)

Three trials were included in this comparison.5;10;13

Weight loss

At 12 months, physical activity (minimum of 30 minutes three times a week) was associated with a summary estimate of weight change of −3.09 kg (WMD 95% CI −4.00 to −2.18, based on three comparisons) compared with no treatment.

Figure 15.19. Maintenance of weight loss over time for physical activity compared with no treatment.

Figure 15.19

Maintenance of weight loss over time for physical activity compared with no treatment.

Other outcomes

One trial10 showed significant decreases in levels of triglycerides at 12 months (−0.30 mmol/l, WMD 95% CI −0.49 to −0.10) in the activity group, but other clinical outcomes were not significantly different between the groups.

Other factors
Gender

Two of the three included trials recruited men only.5;10;75;76

Current medical conditions

The Oslo Diet and Exercise Study (ODES) study recruited men and women who had athero-thrombogenic syndrome (or insulin resistance).

Age

All included studies recruited middle-aged people (age range 30–59 years).

Setting and delivery

Of the three trials, one was community based, one workplace based, and one based in a university clinic. None were based in the UK.

The physical activity was supervised in two studies10;13 by ‘highly qualified instructors’ and ‘training staff’, respectively.

Context and methodological notes

The included studies in the Cochrane review were re-examined and either excluded (or were reassigned to alternative categories). The ODES study was also added.

Physical activity versus information

One trial was found that compared physical activity with information on a healthy lifestyle.58;77

Weight loss

At 18 months, mean ± SD weight change in kilograms was −3.46 ± 6.89 kg in the activity group, and −1.10 ± 6.23 kg in the information group. Mean weight change in the intervention group compared with behaviour therapy was −2.36 kg (95% CI −4.41 to −0.31).

Other outcomes

A significant improvement in the 6-minute walk distance (p < 0.05) was observed in the activity group.

Other factors
Current medical conditions

The included study recruited only participants with osteoarthritis of the knee.

Age

Messier58 recruited older people only (aged 60 years or older).

Setting and delivery

The trial was based in an older people’s independence centre in the USA. The physical activity could be done either in the centre or at home. Details of whether the facility-based training was supervised was not reported.

Physical activity versus diet

Three trials were included in this comparison.5;10;13 All three diets were classified as 600 kcal/deficit or low-fat diets.

Weight loss

At 12 months, physical activity (minimum of 30 minutes three times a week) was associated with a summary estimate of weight change of +3.32 kg (WMD 95% CI 2.28 to 4.35, based on three comparisons) compared with a 600 kcal/deficit or low-fat diet.

Figure 15.20. Maintenance of weight loss over time for physical activity compared with diet.

Figure 15.20

Maintenance of weight loss over time for physical activity compared with diet.

Other outcomes

No reported outcomes were significantly different overall at any time points.

Other factors
Gender

Two of the three included trials recruited men only.5;10;75;76

Current medical conditions

The ODES study recruited men and women who had athero-thrombogenic syndrome (or insulin resistance).

Age

All included studies recruited middle-aged people (age range 30–59 years).

Setting and delivery

Of the three trials, one was community based, one workplace based, and one based in a university clinic. None were based in the UK.

The physical activity was supervised in two studies10;13 by ‘highly qualified instructors’ and ‘training staff’, respectively.

The dietary advice was delivered to individuals5;10 or with the spouse initially.13

Context and methodological notes

The included studies in the Cochrane review were re-examined and either excluded (see Appendix 16) or were reassigned to alternative categories. The ODES study was also added.

Although the Wood10 study did mention that behavioural strategies were used, no details were reported so the intervention was classed as diet only.

Physical activity versus diet and behaviour therapy

One trial was found that compared physical activity with diet and behaviour therapy.58;77

Weight loss

At 18 months, mean ± SD weight change in kilograms was −3.46 ± 6.89 kg in the activity group, and −4.61 ± 7.22 kg in the diet and behaviour therapy group. Mean weight change in the intervention group compared with behaviour therapy was 1.15 (95% CI −1.02 to 3.32).

Other outcomes

A significant improvement in the 6-minute walk distance (p < 0.05) was observed in the activity group.

Other factors
Current medical conditions

The included study recruited only participants with osteoarthritis of the knee.

Age

Messier58 recruited older people only (aged 60 years or older).

Setting and delivery

The trial was based in an older people’s independence centre in the USA. The physical activity could be done either in the centre or at home. Details of whether the facility-based training was supervised was not reported. No details of who delivered the behaviour therapy were reported, but the majority of sessions were group sessions.

Context and methodological notes

Two papers on the same study were identified which appeared to give different results. For this review, the Messier58 paper was used, but Nicklas77 was checked for apparent differences (possibly due to different populations used).

Physical activity and diet versus control (no treatment)

Three trials were included in this comparison.13;78–87

Weight loss

At 12 months, physical activity (minimum of 45 minutes three times a week) and diet (600 kcal/deficit or low fat) was associated with a summary estimate of weight change of −6.87 kg (WMD 95% CI −7.88 to −5.87, based on five comparisons) compared with a no treatment control.

Figure 15.21. Maintenance of weight loss over time for physical activity and diet compared with no treatment.

Figure 15.21

Maintenance of weight loss over time for physical activity and diet compared with no treatment.

Other outcomes

Overall, significant improvements in levels of total cholesterol (−0.27 mmol/l, WMD 95% CI −0.42 to −0.12, three comparisons), LDL-cholesterol (-0.20 mmol/l, WMD 95% CI −0.33 to −0.06, three comparisons), HDL-cholesterol (0.12 mmol/l, WMD 95% CI 0.09 to 0.16, three comparisons), triglycerides (−0.29 mmol/l, WMD 95% CI −0.41 to −0.17, three comparisons), fasting plasma glucose (−0.33 mmol/l, WMD 95% CI −0.54 to −0.12, three comparisons), DBP (−4.64 mm Hg, WMD 95% CI −6.04 to −3.25, three comparisons), and SBP (−4.60mm Hg, WMD 95% CI −6.61 to −2.58, three comparisons) at 12 months were seen in the activity and diet group, but other clinical outcomes were not significantly different between the groups.

Other factors
Gender

The authors of one trial78 reported that ‘exercise prevented weight gain in women and produced weight loss in men’.

Current medical conditions

The ODES study recruited men and women who had athero-thrombogenic syndrome (or insulin resistance).

Setting and delivery

Of the three trials, one was community based, and two were based in university clinics. None were based in the UK.

The physical activity was supervised in two studies (ODES and Midwest Exercise Trial) by ‘highly qualified instructors’13 and ‘research assistants’,78 respectively.

The dietary advice was delivered to individuals (ODES13) or with the spouse initially (ODES13). In the Midwest Exercise trial, participants had to eat in the university cafeteria, but no specific counselling was reported as being given (other than told to follow an ad-libitum diet of 30–35% fat, 45–55% carbohydrates, 10–25% protein).78–80

Context and methodological notes

The Finnish Diabetes Prevention Study (FDPS) (HTA) was moved to another section, as the control group did receive some specific dietary advice.

Physical activity and diet versus information

One RCT was found comparing physical activity and diet (600 kcal/deficit or low fat) with information.88–90

Weight loss

At 12 months, a combination of physical activity (30 minutes of moderate exercise daily plus supervised resistance training twice a week) and diet (classified as calorie deficit) was associated with a summary estimate of weight change of −3.50 kg (WMD 95% CI −4.27 to −2.73, based on one comparison) compared with information.

Figure 15.22. Maintenance of weight loss over time for physical activity and diet compared with information.

Figure 15.22

Maintenance of weight loss over time for physical activity and diet compared with information.

Other outcomes

At 12 months, participants in the activity and diet group had greater improvements in HDL, triglyceride and fasting plasma glucose at 12 months than the information alone group. However, these were not maintained in the longer term (up to 36 months). In addition, participants in the activity and diet group were less likely to develop diabetes (p = 0.0001) during the study period.

Other factors
Participants and condition

People (majority women) were included if they had impaired glucose tolerance.

Delivery

A nutritionist gave individualised and group information on diet. The study physician and the nutritionist informed participants of general risk factors for diabetes. Physical activity was delivered either by an exercise instructor or physiotherapist if part of the study team, or if not, by commercial services (the delivery of the intervention varied across centres, depending on local availability of services and personnel).

Physical activity and diet versus diet alone

Three RCTs13;27;81–87 and one pilot study27 were included in this comparison.

Weight loss

At 12 months, a combination of physical activity (min 45 minutes three times a week) and diet (classified as 600 kcal/day deficit or low fat) was associated with a summary estimate of weight change of −1.95 kg (WMD 95% CI −3.22 to −0.68, based on three comparisons) compared to diet alone.

A significant difference was maintained at 18 months (−7.63 kg (WMD 95% CI −10.33 to −4.92, based on six comparisons), with dietary interventions including a low-calorie diet (1000 kcal/day) and PSMF (420 kcal/day).

Figure 15.23. Maintenance of weight loss over time for physical activity and diet compared with diet alone.

Figure 15.23

Maintenance of weight loss over time for physical activity and diet compared with diet alone.

Other outcomes

Overall, three comparisons from two trials13;86 showed significant improvements in triglyceride levels at 12 months (−0.18 mmol/l, WMD 95%CI −0.31 to −0.06). Also blood pressure was improved (DBP −12.10 mm Hg, WMD 95%CI −15.20 to −9.00, and SBP −8.90 mm Hg, WMD 95%CI −13.65 to −4.15) at 18 months, but not at 12 months (one comparison).27

Other factors
Gender

Two studies (Pavlou 1989 1 and 1989 2) recruited men only.27

Current medical conditions

The ODES study recruited men and women who had athero-thrombogenic syndrome (or insulin resistance).

Setting and delivery

One study was community based,13 two were workplace based,27 and one was based in a university clinic.86 None were based in the UK.

The physical activity was supervised in all studies. No details of who supervised the exercise were reported in three studies, but in one trial the exercise was supervised by ‘highly qualified instructors’.86

Dietary interventions were delivered by nutritionists,88 or by a registered dietitian.86 There was a variety of methods of delivering the dietary advice: from counselling with the spouse13 to weekly education sessions27 (assumed to be group sessions).

Context and methodological notes

The included studies in the Cochrane review were re-examined and either excluded (see Appendix 16) or were reassigned to alternative categories. HTA studies were also included. However, the Phenix study70 was excluded as was not a published paper, but a PhD thesis (no subsequent articles published, checked April 2005).

Physical activity and behaviour therapy versus passive (information or self-help) behaviour therapy
Weight loss

At 12 months, a combination of physical activity (approximately 45 minutes five times a week) and behaviour therapy (stimulus control, problem solving, reducing barriers, exercising in different weather conditions) was associated with a summary estimate of weight change of −0.10 kg (WMD 95% CI −2.52 to 2.32, based on one comparison) compared to passive (information, self-help) behaviour therapy (self-monitoring, goal setting, social support, relapse prevention, problem solving) alone.

Figure 15.24. Maintenance of weight loss over time for physical activity (PA) and behaviour therapy (BT) compared with information.

Figure 15.24

Maintenance of weight loss over time for physical activity (PA) and behaviour therapy (BT) compared with information.

Other outcomes

No other reported outcomes were significant.

Other factors
Age

Wing 1998 recruited people aged 40–55 years.55

Current medical conditions

Wing and coworkers recruited people who did not have diabetes, but who had one or both biological parents with diabetes.55

Setting and delivery

Wing and coworkers’ study was assumed to be based in a university hospital in the USA.55 Behavioural treatments and information on exercise were delivered in lectures given by a team led by a behaviour therapist and an exercise physiologist, who also supervised the exercise.

Physical activity, diet and behaviour therapy versus control (no treatment)

Six trials were included in this comparison.91–102

Weight loss

At 12 months, a combination of physical activity (varying in level from three sessions to 30–45 minutes 4 to 5 times week), behaviour therapy (situational control, including cue avoidance, self-monitoring of calorie intake, eating behaviours and pulse rate, management of eating behaviours, relapse prevention, goal setting, cognitive reframing and coping imagery, stimulus control, social assertion, reinforcement techniques for enhancing motivation, cognitive strategies for replacing negative thinking with more positive statements and constructive self-statements), and diet (either calorie deficit or a low-calorie diet) was associated with a summary estimate of weight change of −4.22 kg (WMD % CI −4.80 to −3.64, based on five comparisons) compared with control.

Figure 15.25. Maintenance of weight loss over time for physical activity, diet and behaviour therapy compared with control.

Figure 15.25

Maintenance of weight loss over time for physical activity, diet and behaviour therapy compared with control.

Other outcomes

From the reported outcomes, no significant differences were seen other than for blood pressure. Both DBP and SBP were statistically significantly improved at all time points (other than DBP at 36 months) in populations with ‘high normal’ blood pressure or with diagnosed hypertension.

Other factors
Age

The included studies covered a wide range of ages (25–45 years,102 60–80 years TONE92).

Current medical conditions

Jalkanen’s99 and the TOHP studies96–98 included participants with hypertension.

Gender

The TOHP I study96 suggested although there were sex differences in the results for blood pressure, these were largely (but not exclusively) due to the differential weight loss for men and women.

Ethnicity

Subanalysis from the TONE study suggested that white participants lost more weight than black (African American) participants without, but not with, a concurrent focus on sodium reduction.92–95

Setting and delivery

One study was based in primary hypertension clinics,99 three were based in university clinics or academic health centres,92–98 and two gave no details of the setting. None of the studies were based in the UK.

Where reported, dietary components were delivered by professionals with expertise in nutrition.

Where reported, physical activity components were delivered by physiotherapists,99 or exercise physiologists.96 Activity was supervised in one study99 (Jalkanen) and took place in the weekly sessions in the TOHP studies.96–98

Behaviour therapy was delivered by psychologists,96;99 trained interventionists,92;102 undergraduates with a thorough knowledge91 or health educator.97 The majority of studies used a group format. TONE92 used a mixture of individual and group sessions, and the Ost study91 was unclear but appeared to be individual sessions.

Context and methodological notes

Several studies55;59;103;104 were moved to various other sections as the control group was considered to have received some intervention, even if the intervention was passive (for example, provision of a manual alone without additional support).

Physical activity, diet and behaviour therapy versus information

Seven trials were included in this comparison.55;58;77;103–108

Weight loss

At 12 months, a combination of physical activity (minimum 150 minutes per week), behaviour therapy (behaviour change goals and problem solving, goal setting, menu planning, self-efficacy, consideration of body image, social support, social eating, removing road blocks, positive thinking, dealing with high-risk situations and slips, cue elimination, stress management and relapse prevention, self-monitoring, problem solving, managing cues, stimulus control, positive assertion, positive thinking, holiday eating, social support, motivation, role playing, modelling food tasting and grocery store tours), and diet (either calorie deficit or a VLCD) was associated with a summary estimate of weight change of −3.82 kg (WMD 95% CI −4.63 to −3.02, based on six comparisons) compared with information alone.

Narayan and coworkers’ study104 which included only Pima Indians as the study population, appeared to be an outlier at 12 months.

Figure 15.26. Maintenance of weight loss over time for physical activity, diet and behaviour therapy compared with information.

Figure 15.26

Maintenance of weight loss over time for physical activity, diet and behaviour therapy compared with information.

Other outcomes

The majority of clinical outcomes were not significantly different between groups and were not maintained over time (where reported). However, Wing and coworkers showed improved levels of triglycerides at 6 and 24 months (not at 12 months.55

Also both DBP and SBP were improved significantly at 12 months (DBP −1.74 mm Hg, WMD 95%CI −3.43 to −0.04, and SBP −3.59 mm Hg, WMD 95%CI −6.31 to −0.86, based on three comparison) but this was not maintained at 24 months.

Wolf and coworkers108 also reported that participants in the combined group lowered their use of medications, primarily diabetes medications, by 0.8 medications per day more than participants treated with usual care (p = 0.03). In seven of nine quality-of-life domains, the combined group improved compared with usual care (p < 0.05).

Other factors
Gender

Djuric and coworkers included women only, and the rest of the studies had a majority of women participating.105

Current medical conditions

Mayer-Davis and coworkers107 included only people with diabetes, Lindhahl and coworkers103 those with abnormal glucose tolerance, Messier58 older people with arthritis, and Djuric and coworkers105 survivors of breast cancer.

Setting and delivery

Studies were conducted in a variety of settings, including primary care. No studies were based in the UK.

Dietary components were delivered by dietitians55;105 and nutritionists,107 where reported. Physical activity components were delivered by exercise physiologists55 and nutritionists.107 Activity was supervised in two studies (during the inpatient stage103 and walk with the therapist at weekly meetings55). Behaviour therapy was delivered by behaviour therapists and health educators.

Lindahl and coworkers delivered all interventions in a 1-month stay at a wellness centre, repeated in a 4-day stay at 12 months.103

Context and methodological notes

Some studies were excluded from this section for the following reasons: weight change in the control (education) group was not reported;52 no requirement for participants to be overweight;109 and because the intervention was not delivered in a clinical setting.105,110

Physical activity, diet and behaviour therapy versus diet

Only one study compared physical activity, diet, and behaviour therapy with diet.111

Weight loss

At 12 months, a combination of physical activity (supervised training plus home-based activity), behaviour therapy (self-monitoring, stimulus control, self-reinforcement, cognitive restructuring and relapse prevention training), and diet (calorie deficit) was associated with a summary estimate of weight change of −0.67 kg (WMD 95% CI −4.22 to 2.88, based on one comparison) compared to diet alone.

Figure 15.27. Maintenance of weight loss over time for physical activity diet and behaviour therapy compared with diet.

Figure 15.27

Maintenance of weight loss over time for physical activity diet and behaviour therapy compared with diet.

Other outcomes

No reported outcomes were significant.

Other factors
Current medical conditions

Blonk and coworkers recruited only people with type 2 diabetes.111

Setting and delivery

The study was conducted in the Netherlands, and no detail of setting was reported.

A dietitian provided dietary education. Group sessions on behaviour therapy were led by a psychologist experienced in eating disorders. Group exercise training was led by two physiotherapists, so assumed to be supervised.

Context and methodological notes

Phenix cluster RCT70 (HTA study) excluded as no published papers could be identified (unpublished PhD thesis only).

Physical activity, diet and behaviour therapy versus behaviour therapy

Only one study compared physical activity, diet and behaviour therapy with behaviour therapy.112

Weight loss

At 12 months, a combination of physical activity (individualised level), behaviour therapy (self-monitoring, stimulus control, reinforcement, cognitive change), and diet (calorie deficit) was associated with a summary estimate of weight change of −5.80 kg (WMD 95% CI −8.91 to −2.69, based on one comparison) compared with behaviour therapy (enhancing body acceptance, disentangling self-worth from weight, barriers transformation, increased support and assertion, self-monitoring) alone.

Figure 15.28. Maintenance of weight loss over time for physical activity, diet and behaviour therapy (BT) compared with BT.

Figure 15.28

Maintenance of weight loss over time for physical activity, diet and behaviour therapy (BT) compared with BT.

Other outcomes

No reported outcomes were significant.

Other factors
Age

Only women aged 30–45 years were recruited.

Gender

Bacon and coworkers112 recruited women only. The participants also had to be ‘chronic dieters’ (defined as a Restraint score greater than 15).

Setting and delivery

Bacon’s study was conducted in the USA and assumed to be based in university clinics.

The diet programme was taught by an experienced registered dietitian, based on the LEARN manual. The non-diet programme was facilitated by a counsellor with experience of using that approach previously. Both interventions were delivered in group formats. Activity was not supervised.

Physical activity, diet and behaviour therapy versus physical activity

One trial was found that compared physical activity, diet and behaviour therapy with physical activity.58;77

Weight loss

At 18 months, mean ± SD weight change in kilograms was −5.20 ± 6.89 kg in the combined group, and −3.46 ± 6.89 kg in the activity group. Mean weight change in the intervention group compared with activity was −1.74 (95% CI −3.98 to 0.50).

Other outcomes

At 18 months, no significant differences in improvement were seen for self-reported physical function, motility or pain between groups.

Other factors
Current medical conditions

The included study recruited only participants with osteoarthritis of the knee.58

Age

Messier and coworkers recruited older people only (aged 60 years or older).58

Setting and delivery

The trial was based in an older people’s independence centre in the USA. The physical activity could be done either in the centre or at home. No details of whether the facility-based training was supervised were reported. No details of who delivered the behaviour therapy were reported, but the majority of sessions were group sessions.

Context and methodological notes

Two papers on the same study were identified and which appeared to give different results. For this review, the Messier paper was used58, but Nicklas77 was checked for apparent differences (possibly due to different populations used).

Physical activity, diet and behaviour therapy versus diet and behaviour therapy

Six studies were included in this comparison.55;58;77;113–117

Weight loss

At 12 months, a combination of physical activity (min 45 minutes three times a week), behaviour therapy (contracts to reward behaviour change, stress management, stimulus control and goal setting, cognitive behaviour therapy, slowing down rate of eating, reducing eating signals in the home, social pressures, pre-planning and relapse prevention techniques, problem solving, reducing barriers, exercising in different weather conditions), and diet (calorie deficit or VLCD) was associated with a summary estimate of weight change of −1.59 kg (WMD 95% CI −3.67 to 0.49, based on six comparisons) compared with diet and behaviour therapy.

No significant difference was seen between the two approaches at any reported time point.

Figure 15.29. Maintenance of weight loss over time for physical activity, diet and behaviour therapy (BT) compared with diet and BT.

Figure 15.29

Maintenance of weight loss over time for physical activity, diet and behaviour therapy (BT) compared with diet and BT.

Other outcomes

No reported outcomes were significant.

Other factors
Gender

Sikand and coworkers115 and Wadden and coworkers116 recruited women only.

Current medical conditions

Wing and coworkers117 recruited people with type 2 diabetes. Wing and coworkers55 recruited people who did not have diabetes, but who had one or both biological parents with diabetes. Messier and coworkers recruited older people with arthritis.58

Setting and delivery

All studies were based in the USA, and where reported, were conducted in university clinics.

Dietary components were delivered by dietitians where reported. Similarly, activity was delivered by exercise physiologists, where reported. The activity was supervised in two studies (initially),113;117 but individuals were also encouraged to exercise on their own. Also supervised activity was undertaken in two studies (walk with the therapist at weekly meetings).116;117 It was unclear in one study.111

Behaviour therapy was delivered by behaviour therapists,58 dietitians with experience in behaviour modification,113 and clinical psychologists.116 All studies used a group format for delivery.

Context and methodological notes

The Phenix cluster RCT70 (HTA study) was excluded as no published papers could be identified (unpublished PhD thesis only). Also the Kaplan52 (HTA study) was excluded as all participants received the Exchange diet and an exercise prescription based on the results of the graded exercise test.

Wadden116 was added to the meta-analysis as follows: 26 week results added to 6 months, 48 week results added to 12 months, and 100 weeks added to 24 months.

Wing117 was moved to the section comparing physical activity, diet and behaviour therapy with physical activity and behaviour therapy.

Physical activity, diet and behaviour therapy versus physical activity and behaviour therapy

One study compared a combined approach with physical activity and behaviour therapy.55

Weight loss

At 12 months, a combination of physical activity (approximately 45 minutes five times a week maximum), behaviour therapy (stimulus control, problem solving, reducing barriers, exercising in different weather conditions), and diet (VLCD) was associated with a summary estimate of weight change of −7.00 kg (WMD 95% CI −10.90 to −3.10, based on one comparison) compared with physical activity and behaviour therapy.

Figure 15.30. Maintenance of weight loss over time for physical activity, diet and behaviour therapy (BT) compared with activity and BT.

Figure 15.30

Maintenance of weight loss over time for physical activity, diet and behaviour therapy (BT) compared with activity and BT.

Other outcomes

Some clinical outcomes (such as cholesterol, triglycerides, and SBP) improved in the short term (6 months), but the differences were not maintained beyond that time.

Other factors
Age

Wing and coworkers included only people aged 40–55 years.117

Current medical conditions

Wing recruited people who did not have diabetes, but who had one or both biological parents with diabetes.117

Setting and delivery

The study was based in the USA and assumed to be in a university hospital.

Group meetings were led by a multidisciplinary team, with primary therapists for activity being a behaviour therapist and an exercise physiologist, and for diet a behaviour therapist and a registered dietitian. The activity was supervised (walk with the therapist at weekly meetings).

Intensity of physical activity

This review was based upon six RCTs118–123 which varied either exercise intensity (kilocalories expended per week) or duration (length of exercise time) or both. Weight loss was not the primary outcome measure in all of these studies but each trial did vary exercise prescriptions and report weight loss between groups. For this evidence review, only trials with a duration (including follow-up) of 12 months or more were included. Also, mean BMI of participants had to be 28 or over. Due to the nature of the studies (weight loss not being a primary outcome, heterogeneity of the interventions) no summary statistics were calculated. Additional studies comparing exercise with stretching124;125, walking vs resistance training126, and intensive vs a control exercise programme127 were identified in the Update searches. However, because the GDG considered that there was a lack convincing evidence overall, recommendations would be based on accepted national guidance,128 so these have not been added to this review.

Weight loss

Between group weight losses were observed in only two of the seven studies reviewed. Jakicic and coworkers reported no significant difference between the long and short bout exercise groups but weight loss was significantly different between the short bout group which used exercise equipment (treadmill) and the short bout group which did not use a treadmill.120 In Jeffery and coworkers’ study the between-group weight loss in the high physical activity group (2500 kcal/week) and the standard behaviour group (1000 kcal/week) was significant at 18 months (p = 0.04).123

Other outcomes

There were no significant cardiovascular differences reported.

Other factors
Gender

Four of the included trials recruited women only.118–121 Overall, the majority of participants were women.

Age

All included studies recruited middle-aged people (age range 25–58 years).

Setting and delivery

All studies were US based. Exercise settings varied between home and exercise centre and between supervised and unsupervised. Reporting of home-based, unsupervised exercise was by self-report.

15.3.4. Pharmacological interventions

15.3.4.1. Evidence statements – orlistat

(Table 15.19)

Table 15.19. Evidence statements and grading.

Table 15.19

Evidence statements and grading.

15.3.4.2. Evidence review on orlistat

This review was primarily based on two key reviews.4;129 A comparison of the reviews can be seen in Table 15.20. Additional searching was also done to identify any other RCTs published since these key reviews were published. Reference lists of other relevant reviews were also cross-referenced. Update searches did not identify any additional relevant studies.

Table 15.20. Comparison of systematic reviews on orlistat for weight loss in adults.

Table 15.20

Comparison of systematic reviews on orlistat for weight loss in adults.

For this evidence review, only trials with a duration (including follow-up) of 12 months or more were included. Also, mean BMI of participants had to be 28 or over.

Orlistat and diet versus placebo and diet

16 studies were included in this comparison.130–152

Weight loss

See summary statistics (Tables 15.2115.23) for results for the individual studies.

Table 15.21. Orlistat and diet vs placebo and diet: weight loss (12 months).

Table 15.21

Orlistat and diet vs placebo and diet: weight loss (12 months).

Table 15.22. Orlistat and diet vs placebo and diet: weight loss (24 months).

Table 15.22

Orlistat and diet vs placebo and diet: weight loss (24 months).

Table 15.23. Orlistat and diet vs placebo and diet: weight loss (48 months).

Table 15.23

Orlistat and diet vs placebo and diet: weight loss (48 months).

Weight maintenance

See summary statistics (Table 15.24 and Appendix 17).

Table 15.24. Orlistat and diet vs placebo and diet: weight maintenance.

Table 15.24

Orlistat and diet vs placebo and diet: weight maintenance.

Other outcomes

See summary statistics (Tables 15.2515.32 and Appendix 17).

Table 15.25. Orlistat and diet vs placebo and diet: total cholesterol.

Table 15.25

Orlistat and diet vs placebo and diet: total cholesterol.

Table 15.26. Orlistat and diet vs placebo and diet: LDL-cholesterol.

Table 15.26

Orlistat and diet vs placebo and diet: LDL-cholesterol.

Table 15.27. Orlistat and diet vs placebo and diet: HDL-cholesterol.

Table 15.27

Orlistat and diet vs placebo and diet: HDL-cholesterol.

Table 15.28. Orlistat and diet vs placebo and diet: triglycerides.

Table 15.28

Orlistat and diet vs placebo and diet: triglycerides.

Table 15.29. Orlistat and diet vs placebo and diet: %Hb1Ac.

Table 15.29

Orlistat and diet vs placebo and diet: %Hb1Ac.

Table 15.30. Orlistat and diet vs placebo and diet: fasting plasma glucose.

Table 15.30

Orlistat and diet vs placebo and diet: fasting plasma glucose.

Table 15.31. Orlistat and diet vs placebo and diet: diastolic blood pressure.

Table 15.31

Orlistat and diet vs placebo and diet: diastolic blood pressure.

Table 15.32. Orlistat and diet vs placebo and diet: systolic blood pressure.

Table 15.32

Orlistat and diet vs placebo and diet: systolic blood pressure.

Other factors
Age

Derosa and coworkers133 included people aged over 40 years only, Kelley and coworkers138 and Miles and coworkers143 people aged 40–65 years, Swinburn and coworkers150 people aged 40–70 years, and Torgerson and coworkers151 people aged between 30 and 60 years.

Current medical conditions

Most studies included people who were otherwise healthy, or had one or more risk factors.131,140,150,151 Hollander and coworkers,137 Kelley and coworkers,138 and Miles and coworkers143 included people with type 2 diabetes only. Only people with raised cholesterol levels were included by Broom and coworkers131 and Derosa and coworkers.133 Bakris and coworkers130 included people with hypertension and who were taking antihypertensive medication.

Setting

Of those studies that reported details of the setting, most were university research clinics or outpatient clinics.

Two were based in primary care alone135,140 and one was based in both primary and secondary care.131

Country

Two studies only were conducted in the UK.131,134

Recruitment

Recruitment was by referral or from clinical databases in Finer and coworkers’ study (partially),134 and in two other studies.133,145 However, most studiesdid not report details of recruitment.

Methodological and context notes

The summary statistics were only calculated by subgroup (that is, weight reduction and weight maintenance) as comparing different outcomes, so we considered it was not appropriate to combine for an overall effect size.

The comparison of failures to achieve 5% or 10% weight loss was only done at 12 months, where the data were available.

Note that many studies calculated weight loss from after a trial run-in period. Therefore, the absolute weight loss may be an under-estimate. Details can be seen in the Evidence tables.

Orlistat and diet versus placebo and then orlistat and diet

One study was identified for this comparison.153

Weight loss

See summary statistics (Appendix 17).

Other outcomes

See summary statistics (Appendix 17).

Other factors

Participants had raised cholesterol.

Orlistat and lifestyle modification versus control (waiting list)

Only one study compared orlistat and lifestyle modification with a waiting list control.154

Weight loss

See summary statistics (Appendix 17).

Other outcomes

Women in the intervention group had significant improvements in total cholesterol and LDL-cholesterol. No other (usual) outcomes were significant.

Other factors
Age

Included women were aged 21–65 years.

Current medical conditions

Women with borderline hypertension or diabetes treated with oral medications were permitted to participate if their physician also consented.

Setting

The study was based in a community setting in the USA, and included only women of Mexican American descent.154

Orlistat and 1000 kcal/day deficit diet versus orlistat and 500 kcal/day deficit diet

One study was identified and participants only continued if they achieved a 5% or more weight loss from baseline at 3 and 6 months.155

Weight loss

See summary statistics (Appendix 17).

Other outcomes

See summary statistics (Appendix 17).

Other factors

No details reported (see Table 15.19 [Generalisability]).

15.3.4.3. Evidence statements – sibutramine

(Table 15.33)

Table 15.33. Evidence statements and grading.

Table 15.33

Evidence statements and grading.

15.3.4.4. Evidence review on sibutramine

The marketing authorisation for sibutramine has been suspended. See details.

Sibutramine and diet versus placebo and diet

Seven studies were included in this comparison.156–164

Weight loss

See summary statistics for results for the individual studies (Table 15.34 and Appendix 17).

Table 15.34. Sibutramine and diet versus placebo and diet: weight loss.

Table 15.34

Sibutramine and diet versus placebo and diet: weight loss.

Weight loss was maintained at 15 months (−3.70 kg, 95% CI −5.71 to −1.69) on a weight reduction programme, and also when used with a weight maintenance programme (−3.40 kg, 95% CI −4.45 to −2.35).

Other outcomes

No significant differences were seen for levels of total cholesterol, LDL-cholesterol, %HbA1c, fasting plasma glucose or SBP overall at 12 months. Levels of HDL-cholesterol (+0.10 mmol/l), and triglycerides (−0.18 mmol/l) improved at 12 months. DBP increased significantly at 12 months (+2.16 mm Hg, 95% CI 1.20 to 3.11) in the sibutramine group compared with placebo.

For people with type 2 diabetes, significant improvements were seen in the sibutramine groups for levels of fasting plasma glucose (−0.40 mmol/l) and triglycerides (−0.30 mmol/l) compared with placebo at 12 months.

For people with hypertension, SBP did not differ significantly between groups, but at 12 months DBP increased in the treatment group compared with placebo (+3.20, 95% CI 1.53 to 4.87).

Other factors
Age and gender

The ages of participants ranged overall from 17 to (maximum defined) 70 years of age. The majority of the participants were women.

Current medical conditions

Three trials included people who were otherwise healthy,156;158;159 two studies people with type 2 diabetes,162;164 and two studies people with well-controlled hypertension.157;163

Setting

One study158 was based in a primary care setting and another study164 included 21 primary and secondary care centres in several countries. The remaining studies were based in either university research clinics or outpatient clinics. The Sibutramine Trial of Obesity Reduction and Maintenance (STORM)159 study was conducted in eight European health centres (appeared to be primary care).

Country

Only Smith and Goulder’s study158 was conducted exclusively in the UK. One Study156 was conducted in France. McNulty and coworkers’ study164 was conducted in England, Canada, France, Belgium. STORM159 was conducted in eight European countries, including the UK.

Recruitment

In Smith and Goulder’s study158 the participants were recruited by by general practitioners and family doctor internists. Apfelbaum and coworkers156 recruited participants from 12 medical centres, although no details were given. In STORM,159 participants were recruited from local health centres. All other studies did not provide details of recruitment.

Sibutramine and diet with activity versus placebo and diet with activity

One trial compared the use of sibutramine or placebo with diet (50% carbohydrates, 30% fats, 20% proteins) and activity (30 minutes of walking per day) in people with type 2 diabetes.165

Weight loss

No significant difference in weight loss was seen at 12 months, although both groups had lost weight (−4.10 kg sibutramine vs −1.40kg placebo).

Other outcomes

Significant improvements were seen in the sibutramine group at 12 months for levels of LDL-cholesterol (−0.37 mmol/l) and %HbA1c (−-0.70).

Other factors

Participants had type 2 diabetes, were taking glibenclamide, and were mainly women. The study was based in secondary care in Mexico.

Sibutramine and diet with activity and behaviour therapy versus placebo and diet with activity and behaviour therapy

Two trials compared the use of sibutramine or placebo with a combination lifestyle intervention.166;167

Weight loss

A significant difference in weight loss was seen at 12 months (−3.48 kg, 95% CI −4.45 to −2.51) between the two groups.

Other outcomes

No significant changes were seen at 12 months.

Other factors

Participants in both studies were otherwise healthy, and the majority were women.

Sibutramine and lifestyle versus lifestyle intervention alone then sibutramine and lifestyle for all participants

One trial compared the use of sibutramine and lifestyle with lifestyle intervention alone in people with type 2 diabetes.168;169 After 12 months, the standard lifestyle group was prescribed sibutramine.

Weight loss

A significant difference in weight loss was seen at 12 months (−6.50 kg sibutramine compared to lifestyle alone), but at 24 months, the standard lifestyle (followed with 12 months sibutramine) group lost more weight than the sibutramine (continuous for 24 months) group (−8.10 kg vs −4.60 kg, respectively).

Other outcomes

No significant changes were seen at any time points.

Other factors

Participants had type 2 diabetes and were mainly women. The study was based in a research centre in the USA.

Sibutramine with a combination lifestyle intervention versus sibutramine, low-calorie diet and activity

One trial was included in this comparison.170;171

Update searches identified one additional study by the same authors, using a similar, but extended protocol.172 This study concluded that the use of sibutramine in combination with lifestyle was more effective than either sibutramine or lifestyle alone. Therefore, as the study supported the detailed evidence review, no further details are reported here.

Weight loss

A significant difference in weight loss was seen at 12 months (−10.61 kg, 95% CI −14.64 to −6.58).

Other outcomes

Authors reported significant reductions at 12 months in triglyceride and low-density lipoprotein cholesterol levels but systolic and diastolic blood pressure both increased significantly (p<0.05 for all).

Other factors

Participants were women who were otherwise healthy. The study was based in a research clinic in the USA.

Sibutramine and diet versus placebo and diet (for 4 months) then open label sibutramine for all participants

One trial was included in this comparison.173

Weight loss

A significant difference in weight loss was seen at 4 months (−4.10 kg, 95% CI −7.58 to −0.62), but this was not maintained at 12 months, although both groups did lose weight (−12.90 kg sibutramine for 12 months, −11.90 kg placebo then diet).

Other outcomes

No other outcomes were reported.

Other factors

Participants were women who were otherwise healthy. The study was based in an obesity management centre in the Czech Republic.

Sibutramine in combination with orlistat

Wadden and coworkers171 (continuation trial of Wadden and coworkers170) evaluated whether adding orlistat to sibutramine would induce further weight loss in participants who previously had lost weight while taking sibutramine alone. The initial period of the trial (that is, sibutramine alone) lasted 12 months, with the period using both drugs lasting only 16 weeks. Because the period of evaluation of the combination was only 16 weeks, this trial did not meet our criteria of a 12 month follow-up (for the combination), so it was excluded from the review.

15.3.4.5. Summary of previous NICE Technology Appraisals - drugs

Orlistat

The Technological Appraisal Guidance- No.22: Guidance on the use of Orlistat for the treatment of obesity in adults (2004) stated that orlistat should be prescribed for people who have lost at least 2.5kg by dietary control and increased physical activity alone in the month prior to first prescription. This guideline recommends that pharmacological treatment should be initiated only after dietary, exercise and, additionally, behavioural approaches have been started (NICE 1.2.5.1, full guideline 1.7.5.1).

The Technological Appraisal Guidance- No.22: Guidance on the use of Orlistat for the treatment of obesity in adults (2004) did not recommend the continuation of therapy beyond 12 months. This guideline recommends that the decision to use drug therapy for greater than 12 months (usually for weight maintenance) should be made after discussing potential benefits and limitations with the patient (NICE 1.2.5.20, full guideline 1.7.5.20).

See also Section 6 for the detailed health economic modelling.

Sibutramine

The Technological Appraisal Guidance- No.31: Guidance on the use of Sibutramine for the treatment of obesity in adults (2001) recommended a starting dose of 10mg/day. This guideline does not recommend a starting dose but suggests that prescription should be in accordance to the drug’s summary of product characteristics (NICE 1.2.5.4, full guideline 1.7.5.4).

See also Section 6 for the detailed health economic modelling.

15.3.5. Surgery and referral to specialist services

15.3.5.1. Evidence statements

(Table 15.35)

Table 15.35. Evidence statements and grading.

Table 15.35

Evidence statements and grading.

15.3.5.2. Evidence review on different surgical procedures

This review was primarily based on two key reviews.174;175 Additional searching was also done to identify any other trials published since these key reviews were published. Reference lists of other reviews were also cross-referenced. 176;177

On expert advice, we widened our inclusion criteria for study design, and included study designs as follows:

  • RCTs
  • controlled clinical trials
  • controlled before-and-after studies
  • case series.

The focus of the study had to be the effectiveness of the surgical procedure, with a key (required) outcome of weight loss. Other criteria for inclusion were a minimum period of follow-up (mean or median) of at least 12 months for RCTs, and (for studies other than RCTs) a minimum follow-up of 24 months and minimum number of 150 participants.

The evidence review only considered procedures that were currently being performed. These are of three types: restrictive, restrictive/malabsorptive, malabsorptive/restrictive.

  • Restrictive: the gold standard is laparoscopic adjustable gastric banding (LAGB). Vertical gastric banding (VBG) was a forerunner to laparoscopic gastric banding, but had high rates of complications/failure. Both operations have similar, expected clinical outcomes. Laparoscopic gastric banding restricts intake (volume) of solid food. Older operations are VBG, horizontal gastroplasty, open adjustable banding.
  • Restrictive/malabsorptive: gold standard is gastric bypass (Roux-en-Y), which mainly restricts dietary intake but also reduces absorption.
  • Malabsorptive/restrictive: these are more similar to the older operations. The gold standard is duodenal switch (DS) and biliopancreatic diversion (BPD). These reduce calorie absorption, with limited restriction.

Gastric balloons are a short-term option, so were not considered as an appropriate surgical intervention for this review. Therefore, we did not review the evidence for this procedure.

The areas reviewed were as follows:

  • Surgery versus non-surgical interventions
  • LAGB versus gastric bypass (comparative studies)
  • LAGB versus DS-BPD (comparative studies)
  • DS-BPD versus gastric bypass (comparative studies)
  • Laparoscopic gastric bypass versus open gastric bypass (comparative studies)
  • Gastric bypass (open or laparoscopic) (single-arm studies)
  • LAGB (single-arm studies)
  • DS-BPD (single-arm studies)

Although we have reported details of complication rates associated with each procedure, caution should be used when interpreting these. Many of the series were undertaken over a long period of time, and initial rates of complications were often higher than would be anticipated for an established procedure or service (see Learning curve review below for further details). Also, concern was raised about the effect of different procedures on food intake (and therefore nutrient intake) and appetite. Where studies reported such outcomes, these have been reported, but it should be noted that this is an area of active research.

Also, because of the heterogeneous nature of the populations included in the series and the different levels of reporting, the rates differed considerably across studies.

An evidence review was also undertaken, using the same criteria as above, to evaluate the effectiveness of staged surgery for people with a BMI of 55 kg/m2 or more.

Surgery versus non-surgical interventions

Some of the studies originally included in the HTA have been excluded as they considered the use of procedures no longer performed.178 Also, two cohort studies included in the Cochrane review174 were excluded as they did not have a minimum of 150 participants.179;180 No additional studies were identified.

Three studies met our inclusion criteria,181–183 and details can be seen in Appendix 15. Two of these were RCTs,181;182 and one was a matched cohort study.183–198

No additional studies were identified in the Update searches (see Methods chapter for details).

Weight loss (Figure 15.31)

See evidence tables and statements (Table 15.35). No summary statistics were calculated due to differences in study design, and lack of data to include all relevant studies and time points.

Other outcomes

See evidence tables (Appendix 15).

Restrictive surgery

This section was primarily based on the Technology Evaluation Centre (TEC) review of 2005175 which evaluated laparoscopic gastric banding compared with various other procedures. Of the four included studies comparing the effectiveness of laparoscopic gastric banding with other procedures, two case–control studies met our inclusion criteria.199;200 as Hell and coworkers study201 had fewer than 150 participants, and Morino and coworkers202 compared laparoscopic gastric banding with VBG, which is no longer performed.

Of the 46 single-arm studies (excluding 6 with fewer than 150 participants, a further 16 with a follow-up of less than 24 months, and 5 because no weight loss data were reported), 20 met our inclusion criteria (see Appendix 15 for details).203–222 We also identified one additional paper223 published since the TEC review.175

Update searches identified a further four relevant new or updated single-arm studies.224–227 Three of these studies did not add any additional data to the evidence already reviewed in detail, with weight loss and revision rates being within the ranges reported in the evidence statements above. One study224 however, examined the effectiveness of LAGB in people with a BMI less than, or equal to 35kg/m2. This was a retrospective study using sub-group data from the Angrisani study204 already included in the evidence review. Because of the retrospective nature of the study, and lack of other supporting evidence on this (the authors themselves noted in the Conclusions that surgical indications for BMI ≤35 were questionable), we have not reported the data in this review.

Weight loss

See evidence tables Appendix 15

Other outcomes

See evidence tables Appendix 15 and evidence statements Table 15.35.

From those studies that reported revision rates, removal of the band occurred in approximately 2.3% (median) of participants (range 0.6%–15%). Approximately 6.5% (median) of patients were reoperated on (range 0.5–24%).

In addition:

  • Dargent213 reported that comorbidities improved in all people who lost 25% or more of excess weight (no further details).
  • Favretti and colleagues214 noted that 20% of patients who had excess weight loss greater than 30% lost compliance to dietetic, psychological and surgical advice.
  • In the Frigg study,215 the rate of remission and improvement in comorbiditites were: hypertension 58% and 42%, diabetes 75% and 8%, dyspnea 85% and 12%, arthralgia 52% and 24%, reflux 79% and 11%, self-esteem 45% and 39%, and general physical performance 58% and 33%. An improvement in stress incontinence, sleep apnoea, peripheral oedema, and regulation of menstruation was also reported. Greater weight loss was associated with greater reduction in dyspnoea, arthralgia, self-esteem, and physical performance. Hypertension, diabetes, reflux, and oedema improved independent of the amount of weight loss.
  • US Food and Drug Administration (FDA) data210 showed improvements in quality of life measures, such as depression, appearance, physical and emotional function, pain, general, mental, and physical health (no statistical significance reported).
  • O’Brian and coworkers219 reported improvements for people with diabetes (54% asymptomatic, 43% better control), asthma (Asthma Severity Score from 44.5 to 14.3, p < 0.001, 100% reduction in medication), dyslipidaemia (reduction from 34% to 9% of people with triglyceride levels > 2.0 mmol/l), hypertension (complete resolution in 55%, improvement in 31%), sleep apnoea (33% to 2%), gastro-oesophageal reflux disease (GORD) (76% complete resolution, 14% significantly improved), and quality of life (Beck Depression Index 18.0 to 7.8, p < 0.001, Medical Outcome Survey Short Form 36 subscales all returned to normal scores).
  • Spivak and coworkers220 reported improvements in hypertension (recovered 40%, improved 23%), diabetes (recovered 29%, improved 36%), hypercholesterolaemia (recovered 67%, improved 0%), sleep apnoea (recovered 38%, improved 13%), and GORD (recovered 82%, improved 5%).
  • Weiner and coworkers222 reported that 92% of people were satisfied with the general success, and this was associated with an improvement in quality of life. Improvements were also seen for people with hypertension (recovered 38%, improved 57%, diabetes (recovered 38%, improved 63%), and asthma (recovered 6%, improved 9%).
Restrictive/malabsorptive surgery

This section was primarily based on the TEC review of 2005175 which evaluated gastric bypass compared with various other procedures. Eight of the nine included studies either compared the effectiveness of gastric bypass with procedures that are no longer performed (see other sections for comparisons with LAGB and DS-BPD) or did not meet the inclusion criteria for non-RCTs. In addition, Westling and Gustavsson228 compared open versus laparoscopic GB, so was reviewed elsewhere.

Open gastric bypass

An additional search for single-arm studies of open GB found seven studies which met our inclusion criteria.229–235 We focused the review on Roux-en-Y gastric bypass, as that is the commonest method performed in the UK. Details of the studies can be seen in Appendix 15. Any comparative studies comparing open gastric bypass with other included procedures were reviewed in the appropriate section.

Update searches identified one additional study.236 Whilst the mean weight loss was within the range as reported above in the evidence statements, reoperation rates were considerably higher overall (27%). However, two techniques were used (silastic ring and, in the later cases, Fobi pouch) and the reoperation rates were 70% and 7% respectively. The reoperation rate for the newer technique is therefore within the range expected.

Weight loss

See evidence tables Appendix 15 and evidence statements Table 15.35.

Other outcomes

See evidence tables Appendix 15 and evidence statements Table 15.35.

Three studies reported details of reoperation rates:

  • Reoperations were required because of suspicion of anastomotic leak in two patients (only one proved to have a leak), inadvertent removal of a gastrostomy tube (one patient) and small bowel obstruction (five patients).230
  • Pories and coworkers233 reported a reoperation rate of 2.8% (no details).
  • Reoperation due to early and late complications was done in 12% of patients.235

In addition, Avinoah and coworkers229 reported meat intolerance in over half of people and significant, but gradual, decreases in mean iron saturation, haemoglobin, and mean corpuscular volume. Mean vitamin B-12 levels declined significantly during the first four years but showed signs of improvement in the longer term. No significant change was seen in serum folic acid, and levels of serum albumin were normal throughout. People who took vitamin and mineral supplements had significantly higher mean values of vitamin B-12, folic acid and iron saturation than those who did not.

Balsiger and coworkers230 reported a decrease in use of antihypertensive medication (36% to 16%), insulin (12% to < 1%), and anti-inflammatory medication (33% to 9%). Also, 93% of people were satisfied at 3 years after the gastric bypass, with early postprandial satiety in 84% of people and decreased appetite in 82%. Gastrointestinal events (such as constipation, heartburn, vomiting) were rare, but 22% of people reported diarrhoea at least once a week.

Csendes and coworkers231 reported improvements in diabetes (resolved 100%), hyperlipidaemia (resolved 92.5%, improved 7.4%), hypertension (resolved 63.6%, improved 36.3%) and osteoarticular problems (resolved 73.3%, improved 26.6%). Quality of life was also much better with respect to self-esteem (81.6%), work capacity (63.2%), sociability (52.9%), and physical capacity (83.9%), but for the majority, sexual activity remained as before the operation (42.5%). The final Bariatric Analysis and Reporting Outcome System (BAROS) index in 96.6% of people was very good or excellent.

Pories and coworkers233 reported improvements in non-insulin-dependent diabetes mellitus or impaired glucose tolerance (IGT) (91% maintained normal fasting plasma glucose and HbA1c levels) and hypertension (58.1% to 14%). Also, improvements were almost always seen in cardiopulmonary function, sleep apnoea, snoring, asthma, peptic reflux, arthritis, fertility and mental health (including mood).

Schoepel and coworkers234 reported improvements in diabetes (75%), hypertension (81%), GORD (86%), back pain (66%), arthritis (52%), sleeping problems (59%), skin rashes (67%), urinary incontinence (74%) and shortness of breath (92%). Over 90% of people showed an improvement in self-esteem, over 85% in physical activity and over 80% in social activity. Two-thirds of people increased their workload. Almost two-thirds reported an increase in sexual activity.

We were not able to get a complete full-text copy of the Torres paper,235 so have reported only those outcomes as published in the TEC review175.

Laparoscopic gastric bypass

An additional search for single-arm studies of laparoscopic gastric bypass found five studies which met our inclusion criteria.237–241 We focused the review on Roux-en-Y gastric bypass, as that is the commonest method performed in the UK. Details of the studies can be seen in Appendix 15. Any comparative studies comparing laparoscopic gastric bypass with other included procedures are reviewed in the appropriate section.

No additional studies were identified in the Update searches.

Weight loss

See evidence tables (Appendix 15) and evidence statements (Table 15.35).

Other outcomes

See evidence tables (Appendix 15) and evidence statements (Table 15.35).

Reversal of the laparoscopic gastric bypass was reported in only one paper, and was done in only one patient (out of 1497 operations).238 Revision for a leak at the jejunojejunostomy was also reported in one patient (out of 275).239 Reoperation rates for indications including cholecystectomy and leaks also varied (median 1.8%, range 0.03–9.8%).

In addition, Schauer and coworkers239 reported that 97% of people available to follow-up would choose laparoscopic gastric bypass again, if given the opportunity, and 95% of people reported an improvement in their quality of life. Also, improvements were seen in comorbidities such as diabetes (resolved 82%, improved 18%), hypercholesterolaemia (resolved 63%, improved 33%), osteoarthritis or degenerative joint disease (resolved 41%, improved 47%), GORD (resolved 72%, improved 24%), hypertension (resolved 70%, improved 18%), sleep apnoea (resolved 74%, improved 19%), hypertriglyceridaemia (resolved 57%, improved 29%), depression (resolved 8%, improved 47%), urinary incontinence (resolved 44%, improved 39%) and asthma (resolved 13%, improved 69%). Other conditions that either improved or resolved for the majority of people (although the numbers were small for some conditions) were peripheral oedema, migraine headaches, gout, coronary heart disease, chronic obstructive pulmonary disease, congestive heart failure and obesity hyperventilation syndrome. Venous insufficiency remained unchanged in 71% of people but improved in 29% of the seven people with this condition.

Schauer and coworkers240 evaluated the effect of laparoscopic gastric bypass in people with impaired fasting glucose or type 2 diabetes. Improvements were seen in fasting plasma glucose and Hb1Ac (returned to normal levels 83%, markedly improved 17%). Also, a significant reduction in use of oral antidiabetic agents (80%) and insulin (79%) was reported. People with the shortest duration of IGT or diabetes (< 5 years), the mildest form of type 2 diabetes (diet controlled), and the greatest weight loss after surgery were most likely to achieve complete resolution of their type 2 diabetes. Other improvements were seen in hypertension (resolved 36%, improved 53%), hypercholesterolaemia (resolved 37%, improved 41%), sleep apnoea (resolved 33%, improved 47%), symptoms of diabetic neuropathy (improved 50%) and erectile dysfunction (18%, although 82% were unchanged).

Wittgrove and Clark241 reported improvements in GORD (resolved 98%), diabetes (resolved 98%, improved 2%) and hypertension (resolved 92%).

Laparoscopic versus open gastric bypass

This was primarily based on Colquitt and coworkers’ Cochrane review.174 Four RCTs were included.228;242–244 Because there was a body of RCT evidence comparing these two procedures, we did not review lower level evidence, such as case series.

No additional studies were identified in the Update searches.

Weight loss

See evidence tables (Appendix 15) and evidence statements (Table 15.35)..

Other outcomes

See evidence tables (Appendix 15) and evidence statements (Table 15.35).

Malabsorptive/restrictive surgery

This review considered the evidence for the effectiveness of DS (with BPD) and was based primarily on the 2005 TEC review.175

One study comparing DS-BPD and Roux-en-Y gastric bypass was included245 and seven single-arm studies (all included in the TEC review)246–252. One further observational study has been published,223 but did not meet the criterion of 150 participants: only 40 people underwent DS-BPD and 90 people Roux-en-Y gastric bypass.

No additional studies were identified in the Update searches.

Weight loss

See evidence tables (Appendix 15) and evidence statements (Table 15.35).

Other outcomes

See evidence tables (Appendix 15) and evidence statements (Table 15.35).

Revisions were reported in:

  • 5.7% of patients to increase the length of the common channel.246
  • 2.7% of patients to increase the length of the common limb due to malnutrition.248
  • 3.8% of patients due to low protein and excess weight loss, excess diarrhoea, or poor weight loss. Also one reversal was reported due to patient demand.249
  • 6.8% of patients for recurrent protein malnutrition.250

Overall reoperation rates also varied (median 3.9%, range 2.7–6.3%).

In addition, Anthone and coworkers246 reported that no clinical sequelae occurred from hypocalcaemia or anaemia in patients who did not require revision of the length of the common channel. No evidence of hepatic dysfunction or liver failure was seen. No specific food intolerances were seen, and reported mean energy intake was 1600 calories (approximately 63% of preoperative intake).

Biron and coworkers247 showed that, for people with an initial BMI less than 50 kg/m2, a residual BMI of 35 kg/m2 caused a significant drop in the degree of satisfaction from 90% to 40%. For super-obese people (BMI ≥ 50 kg/m2), the same critical point was found at a BMI of 40 kg/m2 where satisfaction dropped from 91% to 57%.

Guedea and coworkers248 reported that glycaemia, cholesterolaemia and triglyceridaemia resolved in 100% of people, and 82.4% stopped antihypertensive medication at 12 months. Sleep apnoea also resolved in 100% and osteoarthritis and difficulty in walking improved in 84%. Menstruation became regular in 100% of women. The results of the operation (using the BAROS classification) were assessed as either excellent or very good at 5 years by 66% of people.

Hess and Hess249 reported normal blood sugar levels for 100% of people with diabetes prior to surgery. In a sample of 100, 9% required iron supplementation or surgery (for excessive uterine bleeding). Although alkaline phosphatase was elevated in a sample of 100 people, vitamin D levels were within the normal range although low. The authors stressed the need for adequate vitamin D and calcium supplementation.

Marinari and coworkers250 reported that 3.5% of operations were classified as a failure, 11% were fair results, 22.8% good, 39.5% very good and 23.2% as excellent results using the BAROS classification.

Slater and coworkers251 focused specifically on vitamin and calcium deficiencies following BPD. Generally, the incidence of deficiencies increased over time. By year 4, 48% of the participants were found to have low calcium levels (from 15% at year 1) and 63% had low levels of vitamin D (57% at year 1). Low vitamin A was found in 69% of participants at 4 years (52% at 1 year) and low vitamin K in 68% (51% at 1 year). However, the incidence of low levels of vitamin E or zinc did not increase over time (vitamin E 0% at 1 year, 4% at 4 years; zinc 51% at 1 year, 50% at 4 years). The authors recommended high levels of supplementation after such surgery, and highlighted the need for ‘long-term nutritional monitoring’.

Totte and coworkers252 reported improvements in hypercholesterolaemia (resolved 100%), type 2 diabetes (resolved 100%), type 1 diabetes (improved 100%), hypertension (resolved 83.6%, improved 17.4%), hypertension and cardiomyopathy (improved 100%), Pickwickian syndrome (resolved 100%), respiratory insufficiency (resolved 82%, improved 18%), sleep apnoea (resolved 66.6%, improved 33.3%), severe arthritic pain (improved 100%) and depression (improved 85.6%).

Staged surgery

This review considered the evidence for the effectiveness of staged surgery in people with a BMI of 50 kg/m2 or greater. Due to the limited evidence on this technique, no limitations on study duration or number of participants were imposed. Four relevant studies were identified.253–256

Arteaga and coworkers253 examined the morbidity and mortality of a two-step approach to surgery (jejunoileal bypass, converted to a Roux-en-Y gastric bypass at 6–24 months). Regan and coworkers256 described their experience of a two-stage gastric bypass (laparoscopic sleeve gastrectomy, followed by laparoscopic Roux-en-Y gastric bypass). Milone and coworkers254 compared laparoscopic sleeve gastrectomy and intragastric gastric balloon as a first stage procedure, before conversion to DS-BPD. However, these authors only reported results for the initial stage, so this study is not considered further in this review. Similarly, although Nguyen and coworkers255 described the use of a staged procedure (modified Roux-en-Y gastric bypass, followed by completion of the sleeve gastrectomy), no results were reported so the study was excluded.

No additional studies were identified in the Update searches.

Weight loss

Arteaga and coworkers253 reported weight loss of approximately 53.4 kg (n = 20/24) after the first stage of surgery (mean follow-up of 14.1 months). Mean BMI decreased from 63.0 to 46.9, and excess weight loss was 44.3%. Of the eight people who went on to have the second stage surgery, mean total weight loss was 80.0 kg and excess weight loss 62%.

Of the seven patients who underwent surgery in Regan and coworkers’ study,256 mean weight loss was 36 kg after stage one (mean follow-up 11 months), and 55 kg in total after stage two (mean follow-up 2.5 months, n = 6/7). Mean BMI decreased from 63 kg/m2 to 50 kg/m2, and then to 44 kg/m2 at each stage. Excess weight loss was 33% and 46%, respectively.

Other outcomes

Complication rates were:

  • for stage one – 8.3% for major complications and no deaths. At stage two, one major complication (12.5%) and no deaths. Overall, the complication rate was 9.4%253
  • 35.7% for overall complications, and no deaths.256

No quality of life or other outcomes were reported in either study.

15.3.5.3. Evidence review on competencies and training for bariatric surgery

A recently published evidence-based guideline was identified that made recommendations on the competencies and skills required of surgeons undertaking bariatric surgery.257 Graded recommendations were that:

  • ‘All surgeons performing obesity surgery should have an adequate technical expertise (based on high quality evidence).
  • S/he should be a qualified and certified general or gastrointestinal surgeon with additional training in obesity surgery (based on medium quality evidence).
  • Technical expertise in laparoscopic surgery alone is insufficient to start a bariatric surgery programme (based on medium quality evidence).’
Learning curve and the effect on operative outcomes

We identified several papers on the effect of the learning curve for our procedures of interest. Inclusion criteria for this review were as for the other surgical reviews (see above), except that no study duration was defined. Also the effect of the learning curve had to be the focus of the study, or a comparison of initial and the later operations had to be reported.

Seven studies were identified: five on laparoscopic gastric bypass,237;258–261 one on laparoscopic and open gastric bypass,262 and one on laparoscopic adjustable gastric banding.263 No additional studies were identified in the Update searches. Results from the five studies on laparoscopic gastric bypass were as follows.

Ballesta-Lopez and coworkers237 analysed a consecutive series of 600 patients to determine problems that arise during the learning curve. The rate plateau of morbidity and mortality (no details reported) was reached after the first 18 patients when the surgical technique was revised and fully standardised. The authors concluded that although the complication rate plateau is often cited as 75–100 operations, this could be lower if adequate training was provided.

Kligman and coworkers258 reported results of the initial 160 consecutive patients undergoing laparoscopic gastric bypass by a single surgeon over a 24-month period. Patients were divided into quartiles for data analysis. Duration of surgery decreased significantly between quartiles (p < 0.01). However, the conversion rate (3.1%) and mean hospital length of stay (2.1 ± 2.4 days) were unaffected by surgeon experience. In addition, the complication rates did not change statistically between quartiles. The authors suggested that throughout the learning curve laparoscopic gastric bypass could be accomplished with acceptable complication rates, conversion rates and hospital length of stay, and that the duration of surgery decreases with experience.

Oliak and coworkers259 aimed to determine the length of the learning curve for a skilled laparoscopic surgeon. The study population consisted of the first 225 consecutive laparoscopic gastric banding procedures attempted by one laparoscopic surgeon. The average operative time decreased (from 189 minutes in the first 75 patients to 125 minutes in the last 75 patients). Most of the improvement in operative time occurred over the first 75 patients. The perioperative complication rate decreased (from 32% in the first 75 patients to 15% in the second and third groups of 75 patients). Complication rates did not significantly decrease after the first 75 patients. Low mortality and conversion rates were achieved early in the series. The authors concluded that low mortality rates and low conversion rates could be achieved early in the learning curve. Complication rates plateaued after approximately 75 operations, and operative times decreased substantially over the initial 75 cases. Operative times continued to decrease at a slower rate beyond the initial 75 cases.

Schauer and coworkers260 determined the effect of operative experience on outcomes in the first 150 consecutive patients. The patients were divided into three groups (1, 2 and 3) of 50 consecutive patients, and outcomes for each group were compared. The patients in group 3 had a larger BMI (p < 0.05), were more likely to have had prior abdominal surgery, and were more likely to have secondary operations at the time of gastric bypass. The operating time decreased (from a mean of 311 minutes in group 1 to 237 minutes in group 3), and technical complications were reduced by 50% after 100 cases. The authors concluded that operative time and technically related complications decreased with operative experience even though heavier patients and higher-risk patients were more predominant in the third group.

Shikora and coworkers261 evaluated technical experience and patient volume on complication rates in the first 750 consecutive patients. For the first 100 cases, the overall complication rate was 26% with a mortality of 1%. This complication rate decreased to approximately 13% and was stable for the next 650 patients (11% to 13.4%). Complications possibly related to technique decreased (trocar site wound infection 8% to 2%, splenic injury 1% to 0%, bowel obstruction 5% to 0%, GI track leak 3% to 0%), while others (such as intraluminal bleeding, thromboembolism, anastomotic stricture) remained approximately the same. The overall mean operating time was 138 minutes (range, 65–310 minutes). It decreased from 212 minutes for the first 100 cases to 132 minutes for the next 650 and 105 minutes (range, 65–200 minutes) for the last 100 cases. The authors concluded that morbidity and mortality could be reduced by up to 50% with experience.

Ballantyne 2005262 compared learning curves for three surgeons in terms of surgical time for laparoscopic gastric bypass and open gastric bypass. Median time for the first surgeon reduced for each subsequent 100 operations: (175 minutes, 125 minutes, 110 minutes, and 100 minutes resp). Median time for second surgeon was 120 minutes overall, and for the third surgeon 173 minutes. The length of surgery significantly correlated with surgical experience in terms of numbers of operations and the BMI of patient. The authors concluded that the length of surgery continued to shorten beyond 400 operations for the first surgeon. Previous fellowship training in laparoscopic gastric bypass shortened surgical times during the initial clinical experience as an attending for the second surgeon.

Weiner and coworkers263 evaluated the outcomes of the first 100 patients and the total number of 984 patients undergoing LAGB. All complications were seen during the first 100 procedures. During the learning curve, more band removals were performed (6/100 compared with 7/884), migration was higher (1/100 compared with 2/884), slippages were higher (12/100 compared with 32/884) and port revisions were higher (3/100 compared with 11/884).

Training and the effect on operative outcomes

We identified only one paper on the effect of the training on our procedures of interest. Inclusion criteria for this review were as for the other surgical reviews (see above), except that no study duration was defined. No additional studies were identified in the Update searches.

Kothari and coworkers264 reported on the outcomes of laparoscopic gastric bypass following completion of an advance laparoscopic fellowship. Outcomes were measured prospectively and analysed by different time quartiles to assess the effect of the training. Upon quartile analysis, there was no difference in complication rates, and the complication rates were comparable to published outcomes in the literature. The authors concluded that fellowships in advanced laparoscopy with emphasis on laparoscopic gastric bypass provided the optimal training environment for the acquisition of necessary skills. With fellowship training, complication rates were comparable to published outcomes in the literature without a period of higher complications (the learning curve).

15.3.5.4. Evidence review on hospital and surgeon volume and surgical outcomes

Inclusion criteria for this review were as for the other surgical reviews (see above), except that no study duration was defined. Also the effect of the volume of operations had to be the focus of the study. Three relevant studies were identified.265–267 No additional studies were identified in the Update searches.

Courcoulas and coworkers265 explored the volume–outcome relation for gastric bypass surgery. They analysed 4685 cases of gastric bypass surgery for obesity, undertaken between 1999 and 2001. Statistical modelling techniques were used to determine whether mortality or the adverse outcome rate was significantly related to hospital and surgeon volume. Outcomes were adjusted for risk factors such as age, gender, comorbidities and others. There was a significant risk-adjusted relation between surgeon volume and adverse outcome (postoperative complications, non-routine hospital transfers), and the same trend was observed for deaths. Surgeons who performed fewer than 10 procedures per year had a 28% risk of adverse outcome and a 5% risk of death, compared with 14% (p < 0.05) and 0.3% (p = 0.06), respectively, for high-volume surgeons. The effect of hospital volume did not reach statistical significance, but there was an interaction between surgeon and hospital volume. Surgeons who performed 10–50 cases per year operating in low-volume hospitals had a 55% risk of adverse outcome (p < 0.01). The authors concluded that risk-adjusted in-hospital adverse outcome was significantly lower when gastric bypass is performed by higher-volume surgeons.

Flum and coworkers266 evaluated the risk of early mortality of patients undergoing bariatric surgery. A retrospective cohort study of 16,155 Medicare patients undergoing bariatric procedures (mean age 47.7 years, 75.8% women) was conducted. The rate of 30-day, 90-day and 1-year mortality was 2.0%, 2.8% and 4.6%, respectively. The odds of death at 90 days were 1.6 times higher (95% CI 1.3 to 2.0) for patients of surgeons with less than the median surgical volume of bariatric procedures (among Medicare beneficiaries during the study period) after adjusting for age, gender and comorbidity index. The authors concluded that the risk of early death after bariatric surgery was considerably higher overall than previously suggested and associated with lower surgeon volume of bariatric procedures. Other associations with early death were advancing age and male sex.

Nguyen and coworkers267 examined the effect of hospital volume on morbidity, mortality and costs in academic centres. All patients who underwent Roux-en-Y gastric bypass were included (n = 24,166). There were 22 high-volume (n = 13,810), 27 medium-volume (n = 7634), and 44 low-volume (n = 722) hospitals included in the study. Compared with low-volume hospitals, patients who underwent gastric bypass at high-volume hospitals had a shorter length of hospital stay (3.8 vs 5.1 days, p < 0.01), lower overall complications (10.2% vs 14.5%, p < 0.01), lower complications of medical care (7.8% vs 10.8%, p < 0.01), and lower costs ($10,292 vs $13,908, p < 0.01). The expected mortality rate (adjusted for severity) was similar between high- and low-volume hospitals (0.6% vs 0.6%,), demonstrating similarities in characteristics and severity of illness between groups. The observed mortality, however, was significantly lower at high-volume hospitals (0.3% vs 1.2%, p < 0.01). In a subset of patients older than 55 years, the observed mortality was 0.9% at high-volume centres compared with 3.1% at low-volume centres (p < 0.01). The authors concluded that bariatric surgery performed at hospitals with more than 100 cases annually was associated with a shorter length of stay, lower morbidity and mortality and decreased costs. This volume–outcome relation was even more pronounced for a subset of patients older than 55 years, for whom in-hospital mortality was threefold higher at low-volume compared with high-volume hospitals. High-volume hospitals also had a lower rate of overall postoperative and medical care complications, which may have been related in part, to the formalisation of the structures and processes of care.

15.3.5.5. Summary of previous NICE Technology Appraisal – surgery

Surgical Interventions

The Technological Appraisal Guidance- No.46: Guidance on the use of surgery to aid weight reduction for people with morbid obesity (2002) defined people as being morbidly obese if they have a BMI either equal or greater than 40kg/m2, or between 35kg/m2 and 40kg/m2 in the presence of significant co-morbid conditions that could be improved by weight loss. This guideline also recommends surgical intervention as a first line option for people with BMI equal to or greater than 50kg/m2 (NICE 1.2.6.7).

The Technological Appraisal Guidance- No.46: Guidance on the use of surgery to aid weight reduction for people with morbid obesity (2002) recommends that choice of intervention should be based upon best available evidence, facilities and equipment available, and experience of the surgeon who would perform the operation. This guideline concurs with this but additionally would include the degree of obesity and the presence of any comorbidities. (NICE 1.2.6.3, full guideline 1.7.6.3).

The Technological Appraisal Guidance- No.46: Guidance on the use of surgery to aid weight reduction for people with morbid obesity (2002) recommended that databases should be established by hospitals wanting to develop their service, to enable outcomes to be monitored in both the short term and the long term. This guideline recommends making arrangements for prospective audits to monitor outcomes both in the short term and the long term (NICE 1.2.6.5, full guideline 1.7.6.5).

See also Section 6 for the detailed health economic modelling.

15.3.6. Evidence review on referral to specialist care for adults and mature adolescents

In February 2005, the National Guideline Clearinghouse synthesised the recommendations on the assessment and treatment of obesity and overweight in adults from six published guidelines.1

There was only one specific set of recommendations on referral (Table 15.36).

Table 15.36. Recommendation on referral.

Table 15.36

Recommendation on referral.

The NHMRC2 recommended that when single-gene mutation obesity is confirmed, the patient should be referred to a specialist who deals with these problems (level B). Reasons for referral from primary to secondary care may include:

  • the view that morbid obesity is a condition that cannot be managed effectively in primary care
  • failure of conventional treatment or
  • for assessment for the suitability for pharmacological treatment.

Most of these referrals would be to registered dietitians, private sector slimming organisations, physicians, community-based programmes/self-help groups, trained exercise specialists or bariatric surgeons. People would normally be referred from secondary to tertiary care if long-term conventional and pharmacological treatment had failed and if they met the eligibility criteria for surgery.268

Due to the lack of evidence in this area, the GDG discussed the role of secondary and tertiary care for people with obesity. Key roles for specialist services were considered to be:

  • the assessment of possible causes of severe obesity, including genetic causes, medication for other conditions, rare neurologic/metabolic conditions
  • the assessment of people with complex disease states and/or complex needs.

Specialist clinics or services were also considered to be the most appropriate setting for:

  • trialling new drugs or using older drugs in novel ways (such as off licence or new combinations)
  • pushing the boundaries of management, such as specialised diets
  • deciding on treatment (drug or surgery) for people with a BMI greater than 40 kg/m2
  • providing a higher level of care than available in primary or general secondary care
  • providing leadership, education and being a source of reference and information for primary or general secondary care.

In the experience of the GDG, people were most often referred to specialist services if:

  • BMI was greater than 40 kg/m2
  • treatment in primary care or general secondary care had failed
  • the obesity or causes were considered ‘unusual’, and therefore may be of research interest.

No additional studies were identified in the Update searches.

15.3.7. Interventions in a UK clinical setting

The following review was presented to the GDG to inform any recommendations about the delivery of care. The GDG considered that no recommendations could be made based on the available evidence. Therefore, the review is presented here for completeness only.

15.3.7.1. Evidence statements

(Table 15.37)

Table 15.37. Evidence statements and grading.

Table 15.37

Evidence statements and grading.

15.3.7.2. Evidence review on interventions delivered in a UK clinical setting

This review aims to provide corroborative evidence from the UK on the effectiveness of any intervention designed to address the management of overweight or obesity in adults and children. The inclusion criteria were as follows:

  • Interventions: Any intervention which targeted providers’ management of obesity and aimed to improve provider practice or patient outcomes or target the individual (such as diet, physical activity). Only studies conducted in the UK were included.
  • Setting: Only those interventions conducted in a clinical setting are included. Other settings are covered in other evidence reviews.
  • Participants: All qualified healthcare professionals involved in the management of obesity and/or all individuals classified as overweight or obese (mean initial BMI ≥ 28 kg/m2 for adults).
  • Outcomes: Studies reporting weight, diet or physical activity outcomes were included provided that baseline and follow-up data were provided.
  • Length of follow-up: At least 12 weeks duration. Length of follow-up was measured from commencement of the intervention.
  • Study design: Only studies with a control or comparison group were included.

Excluded studies are listed in Appendix 16.

The Centre for Reviews and Dissemination (CRD) review on non-clinical settings was cross-referenced, as were other relevant reviews. Excluded references from the intervention reviews were also reassessed for inclusion.

Interventions for the management of obesity and overweight in children

We identified nine studies that could be potentially relevant for this review, although only one of these presented a control/comparison group. However, this study was excluded on the basis of the date of its publication (1971) and therefore being of little relevance to services as they exist currently.269 We also identified two informative papers that evaluated the WATCH IT programme270 in Leeds, and the role of physiotherapists in combating childhood obesity in England.271 However, none of the identified papers met our inclusion criteria (see Appendix 16).

Interventions targeted at individuals (adults)

Two studies were identified that assessed the effect of an intervention targeted at the individual patient in a clinical setting.272;273 One study compared the effect of a low-fat (reduced fat intake by 10%) high complex carbohydrate diet, low-fat high simple carbohydrate diet and control (maintenance of fat intake at habitual amounts of approximately 35–40% of energy) in people with at least three identified risk factors for metabolic syndrome (note: however, overweight or obesity was not a requirement). At least 60% of total energy intake was provided free from the study grocery shop. Participants collected their food once or twice a week from the shop, where they could also discuss their energy and macronutrient intakes with the dietitian.272 This intervention did not appear to be a pragmatic and implementable approach to weight loss, so the results were not reported in this review.

One small (under-powered), quasi-randomised controlled trial273 compared a 500 kcal deficit diet and a health eating diet (based on the ‘Balance of good health’) in adults referred to a dietetic department for weight loss. Participants were invited to four sessions with a dietitian. held at 4-weekly intervals. Both groups lost, on average, clinically significant amounts of body weight by 12 weeks. However, the dropout rate for both groups was high (69% overall at 12 weeks).

Complex interventions targeted at healthcare professionals

Two trials were identified the evaluated the effectiveness of implementing programmes for the management of obesity and overweight. Both trials were conducted in primary care and involved elements of training, process and organisational change, and combined lifestyle interventions.

Moore and colleagues evaluated a training programme designed to improve the management of obesity delivered by primary care teams.274 The programme included several components:

  • A nutrition training programme: This was delivered in three 90-minute sessions, at intervals of no less that 1 week, and no more than 2 weeks apart. All GPs and practice nurses were invited. The training was delivered by four dietitians trained in the standardised delivery of the programme. A model approach was promoted, which included best practice, and was brief enough to be delivered in primary care. Primary care teams then devised individualised weight management protocols to implement in their practice. Practitioners were given a ‘ready reckoner’ to calculate the appropriate caloric intake for individuals.
  • Model of management: Practitioners were expected to see patients approximately every 2 weeks until they lost 10% of initial body weight and then approximately every 1–2 months to support weight maintenance. Current and target weight and dietary and activity targets were to be recorded in the patient records. This was to facilitate the continuity of support from the practice team.
  • Lifestyle information for patients: This included information on the clinical benefit of weight loss, and effective interventions, including reducing energy intake, increasing physical activity, and the use of obesity drugs. A 500 kcal/day deficit diet was recommended, and diet sheets and other written resources were provided.

Control practices were asked to provide usual care. Consecutively attending adults, aged 16–64 with a BMI of 30 kg/m2 or over, were asked to participate.

At 12 months, there was no difference between the intervention and control group patients with regard to weight (mean weight change +0.3 kg intervention vs −0.7 kg control, p = 0.5), but some improvements were seen in practitioner knowledge, the likelihood of weight being discussed in the consultation, and the recording of weight, target weight and dietary targets.

The authors noted limitations of the study: mainly female participants, skewed towards extreme obesity (mean BMI approximately 37 kg/m2 overall), high dropout rate (although the study remained 80% powered).

Another study aimed to evaluate the effectiveness of the Counterweight Programme in improving the management of obesity in primary care using a structured approach.275;276 The programme is based on the evidence-based quality assessment cycle and is in four phases.

  • Phase 1 Audit and project development – setting priorities: The aim of the audit was to determine current approaches. Data were collected on weight screening rates, availability of equipment and patient education materials, and the current organisation of care. Also, baseline attitudes, knowledge, confidence and willingness to treat of the GPs and practice nurses were assessed. The health burden of obesity for each practice was also measured.
  • Phase 2 Practice training and support – setting guidelines: Audit results were fed back in a 1-hour workshop with the GPs and practice nurses, and the treatment pathway and priorities for implementation were also discussed. The role of the GP was identify suitable patients for management during routine clinical practice, and to refer on to the practice nurses. GPs were expected to raise weight as an issue as appropriate and to possibly discuss the benefit of a 5–10% weight loss. A desk-top flip chart was provided to facilitate patient screening and the assessment of motivation. The GP intervention was designed to be opportunistic and of 1–5 minutes duration. A 6–8-hour training programme was designed to teach core competencies to practice nurses. A structured approach was used to cover topics such as screening and assessment, principles of healthy eating and energy balance, dietary approaches to weight management, physical activity guidelines, behaviour change strategies, pharmacotherapy, patient monitoring and ethical considerations. Training manuals were provided to support formal workshops. Guidance was also provided on the use of Counterweight Programme patient education materials. A variety of teaching methods were used, including problem-based learning through case studies, group discussion and practical exercises in line with adult learning theory. A weight management adviser (a state registered dietitian with specialist experience) worked with the practice nurse once or twice a month to facilitate clinics and patient groups. After 6 months, the weight management adviser was mainly involved in data collection and training new practice nurses.
  • Phase 3 Patient intervention – measuring performance: Weight loss targets were set at 5–10% of initial body weight. Patient screening and treatment pathways were developed. Theoretical approaches were used for both changing clinical and patient behaviour. For example, the screening pathway encouraged clinicians to consider stages of change of the individual patient. The screening pathway was designed to target those at highest risk (BMI ≥ 30 kg/m2 or BMI ≥ 28 kg/m2 with comorbidities). The treatment pathway suggested a 3-month minimum lifestyle intervention (either in group or individual formats) as the first line approach. Practice nurses were encouraged to see individual patients for six appointments (10–30 minutes each) over the 3-month period, or for six group sessions of 1 hour. Quarterly follow-up appointments were recommended, where treatment and weight loss were reviewed. At 3 months, people who had lost ≥ 5% were recommended to continue with the lifestyle approach, whereas those had not were considered for alternative lifestyle interventions, pharmacotherapy or referral to a dietitian. Additional options were dependent on local obesity policies and services. If 10% of initial body weight was lost, weight maintenance was advised. Relapse prevention was discussed and weight check appointments offered at least quarterly.
  • Phase 4 Evaluation – improving performance (back to Phase 1): Evaluation (in the form of an RCT) is currently ongoing.

Theoretical approaches to changing practice have been employed to design multifaceted interventions targeting barriers to change (see Table 15.38). Control practices were audited but received no further intervention other than feedback of the audit results.

Table 15.38. Theoretical approaches to changing clinical practice and Counterweight strategies - see paper for full details.

Table 15.38

Theoretical approaches to changing clinical practice and Counterweight strategies - see paper for full details.

The trial is currently ongoing and is due to report full results soon. Interim results show however that for those participants who have data at 12 months, that the mean weight change was −3.2 kg (n = 445), and that 32.6% reached greater than 5% weight loss from baseline. One in six people entering the programme achieved a 5% weight loss or more (ITT analysis) at 12 months.247

The different interventions used can be seen in Figure 15.32.

Figure 15.32. Proportion of individuals reaching 12 months allocated treatment types over the 12 months (adapted).

Figure 15.32

Proportion of individuals reaching 12 months allocated treatment types over the 12 months (adapted).

15.3.8. Patients’ and healthcare professionals’ views and attitudes to the management of overweight or obesity

To inform the development of recommendations relating to patient-centred care, we reviewed evidence on barriers to change and attitudes in the clinical setting (specifically the consultation) reported by healthcare professionals and/or individuals who were overweight or obese.

  • Types of study: Those that assessed barriers and attitudes (individual/HCP/family/carer/other) to the management of weight in the clinical setting. In addition, qualitative studies identifying barriers/motivation to management of overweight/obesity (focus groups, interviews, surveys). Barriers and attitudes had to be identified by participants themselves, not presupposed by researchers, and were restricted to studies from the UK.
  • Types of participant: (i) Healthcare professionals such as GPs, practice nurses, dietitians and health visitors; (ii) family/carers of obese and/or overweight adults, adolescents, and children; and (iii) obese and/or overweight adults, adolescents, children, and their parents.

We identified two published studies that explored this topic in children and adolescents.

  • Types of outcome: participants’ views on the management of weight in the clinical setting.

General practices are important in the management of overweight and obese people as they are frequently the first access point to care. Approximately 75% of the population see their GP during a year, and about 90% in 5 years.277 Patients are then seen by a GP directly because of overweight or obesity, associated comorbidities or some condition that is not connected to weight. Thus, GPs, practice nurses, dietitians and health visitors can play a major role in helping people with weight problems.

This review aims to outline and clarify barriers and attitudes to the management of weight in a clinical setting that are felt to exist by healthcare professionals, patient and family/carers.

15.3.8.1. Healthcare professionals’ views in primary care

(See also Section 1, Chapter 3 section 3.5.)

Owen278 conducted a study via focus groups interviews (based on semi-structured questions) among five groups of healthcare professionals (GPs, dietitians, practice nurses, health visitors and school nurses). The findings echoed those of the National Audit Office,268 calling for a multidisciplinary training and activity programme to be implemented across Wales to ensure healthcare professionals know what advice should be given for weight management.

Despite this call for training, it is frequently unfeasible, as practitioners have expressed their uncertainty of being able to devote so much time for indepth training.274

Hankey and coworkers279 conducted a study on the attitudes, beliefs and eating habits of health professionals with respect to obesity, nutrition and weight management. This study consisted of a postal questionnaire survey of 1400 GPs, 613 practice nurses and 360 practice dietitians who were members of the British Dietetic Association in Scotland. The overall response rate was 65%.

This survey showed that there were discrepancies between professional groups with regard to knowledge of obesity, nutrition and weight management. One example was the disagreement around the role of fat and sugar in increasing obesity. Nearly half of the GPs and 38% of the practice nurses answered incorrectly, whereas 75% of the dietitians correctly answered that fat was the most enhancing factor. Also, almost 65% of all GPs, practice nurses and dietitians believed that people can adhere to an 800–1200 kcal/day diet without weight loss. Knowledge on the measurement of waist circumference appeared to be weakly understood.279

Overall this study demonstrated that health professionals had gaps in their knowledge of nutrition and obesity management, which could create misleading advice. Also, the majority of the respondents felt that weight management advice should be given by professionals with specific training. Practice nurses reported feeling unskilled in regard to giving weight management advice, but many mentioned not having available time to undertake specific training.279

In a similar study, Mercer and Tessier280 carried out interviews with GPs and practice nurses within the Greater Glasgow Healthboard area. Of 30 GPs and 30 practice nurses who were contacted, 10 GPs and 10 practice nurses agreed to be interviewed. The majority of interviews showed that weight management was an unpopular task, with the GPs preferring to delegate it to the practice nurse. This thus created the feeling in practice nurses of patients being ‘off-loaded’ on them. Another issue raised by GPs and practice nurses was a sense of frustration derived from patients’ lack of motivation or lack of success of attempted interventions. Guidelines were also revealed not to be frequently used among GPs, and GPs felt that the Scottish Intercollegiate Guidelines Network (SIGN) was over-promoting the use of slimming pills.280

This issue of GPs asserting that treating obesity was not within their professional domain, and that management obesity should chiefly be the responsibility of the patient, arose clearly in a study conducted by Epstein and Ogden281 and also in a study by Ogden and coworkers282 Epstein and Ogden281 conducted semi-structured interviews with 21 GPs (130 GPs were invited to participate) working in an inner London primary care trust. The GPs claimed that patients tended to see obesity as medical problem, which should be managed by the doctor. According to the authors, this contradiction was then hard to balance with the GPs’ lack of faith in the effectiveness of existing treatments, and the urge to ensure good patient–doctor relationships. The second study282 consisted of a cross-sectional survey with comparisons of models of obesity made between GPs and patients. Questionnaires were completed by 89 GPs and 599 patients. Findings showed that patients pointed out internal uncontrollable causes for the onset of obesity, although would rely on external factors to treat it. On the other hand, GPs blended in cause and solution to internal factors that resided in the patient.

The studies therefore show consistently that GPs and patients have different views. GPs would prefer patient-based interventions in primary care with regard to obesity management, and patients would opt for a professionally led approach.

To briefly summarise the debate on what can hinder or influence the practice of healthcare professionals in primary care, Maryon-Davis277 enumerated the most commonly reported barriers to effective treatment in this particular clinical setting:

  • psychological complexities of cases
  • high rate of relapse
  • perceived lack of effective interventions
  • lack of time
  • lack of resources
  • lack of onward referral options.

15.3.8.2. Practice nurses’ and health visitors’ views of managing overweight and/or obesity

Green and coworkers283 aimed to examine health visitors and practice nurses knowledge regarding the assessment and management of obesity through a postal questionnaire sent to 35 health visitors and 49 practice nurses based at 24 practices within one regional health authority. The questionnaire assessed knowledge concerning diet, behavioural techniques and physical activity regarding obesity management, and was completed and returned by 17 health visitors and 28 practice nurses.

Several respondents were unclear on how to calculate or interpret BMI, and were unaware that a high central fat distribution was linked with greater risk of health problems associated with obesity. In most cases, dietary advice did follow current recommendations, as recommending a low-fat diet, although in some cases this did not occur, as advice would be to follow a reduced carbohydrate diet. Advice on physical activity also varied in quality and quantity. Few respondents mentioned tailoring either dietary or physical activity to cultural or socioeconomic needs.283

Similarly in another study,284 many respondents felt that the advice they gave was useful but was only followed sometimes by patients, and that the responsibility lay with the patients.

Ogden and Hoppe284 studied ways of improving practice nurses’ management of obesity. This was attempted by having practice nurses complete a questionnaire regarding their obesity-related beliefs and behaviours, before and 1 month after being randomly assigned to a ‘learner-centred’ group (which received a leaflet and was asked to attend an interactive seminar), or ‘expert’ group (which received a leaflet), or a control group.

The intervention itself did not have any effect on practice nurses’ beliefs about obesity, and no effects on the patients’ weight. Nevertheless, practice nurses in the ‘learner’ group reported spending more time on their consultations and being more patient centred, with the result of having patients more satisfied with the consultation. On the other hand, practice nurses in the ‘expert’ group reported that they gave advice more frequently and were less patient centred, resulting in an increase in patients’ confidence in achieving weight loss.284

15.3.8.3. Dietitians’ views of managing overweight and/or obesity

Harvey and coworkers285 explored the role of the level of severity (obese vs overweight) on the perceptions of dietitians towards overweight and obese people, and how their views interacted with their practice. For this, they randomly selected 210 dietitian members of the British Dietetic Association, with a further 68 questionnaires given to dietitians who worked in a clinical capacity. One hundred and fifty-eight were returned from the postal survey and 29 returned by the extra group. The questionnaire explored topics such as causes for overweight and obesity, attitudes, responsibility of those who are obese or overweight, the reported practices of dietitians, and the association between views and practice.

The results showed that attitudes towards overweight and obese people were neutral to positive, although overweight people were rated more positively than those who were obese. Dietitians also perceived obese people as being more responsible for their excess weight than those who were overweight. Interestingly, greater merit was perceived in treating obese rather than overweight people.285

15.3.8.4. Patients’/clients’ views on the management of obesity

Tod and Lacey286 conducted a qualitative research project that outlined factors that encouraged or hindered overweight people from low-income groups accessing weight loss services. They recruited 16 people from the South Yorkshire Coalfields Health Action Zone, who attended Slimming World (a slimming club). Semi-structured interviews were conducted with all participants. Despite the sample size and lack of generalisability of this study, some themes emerged.

The main triggers that encouraged people to take action were embarrassment and humiliation, health-related problems or warnings, fear, and critical events such as holidays, weddings or birthdays. Main barriers reported related to issues of denial of their weight and a previous bad experience (such as being the biggest in the group, previous failure, public weighing). The authors contended that, as the experience of being overweight and obese adds frailty and vulnerability, and the triggers to take action are somehow distressing, services should be encouraged to have the adequate level of motivational support.286

Barker and Cooke287 also looked at the perspectives of those who were overweight, obese or in the process of losing weight, by conducting one-to-one interviews and qualitative discussion groups. Those who were defined as overweight had a BMI ranging from 25 kg/m2 to 29 kg/m2; and those defined as obese had a BMI of 30 kg/m2 or more. Slimmers were defined as someone who had been overweight in the last 2 years and had lost at least 6 kg and maintained that loss for at least 6 months. The study addressed issues ranging from perceptions of being overweight to reasons to be overweight, and the strategies that could be undertaken to lose weight.

Most relevant barriers were related to people being unclear about the benefits of weight reduction, and the extreme sacrifices that represented being on a diet which meant having to give up certain pleasures. This then had an effect of inducing long delays in initiating diets.287

15.3.8.5. Family/parents’ views on the management of obesity

Edmunds288 explored, through indepth interviews, parental perceptions of help-seeking experiences with healthcare professionals in central and south-west England. The children were aged from 4 to 15 years, and volunteers were recruited through healthcare professionals, posters in primary care settings and advertising in local papers. Children who attended weight loss groups were also recruited. Included in the weight history components of the interview were standardised shapes of children.

Parents were generally quite positive in their responses regarding healthcare professionals, although a diverse range of responses were reported when they sought child weight management help. The authors suggested that such a finding could be due to the healthcare professional’s concern in not adding further distress to either parent or child, or simply not knowing in which way to help. Again, there appeared to be a tendency for healthcare professionals to leave weight as an individual responsibility and for blaming the parent for the child being overweight.288

15.3.9. Role of professionally organised therapies in the management of overweight and obesity

The following review was presented to the GDG to inform any recommendations about the use of ‘alternative’ or ‘complementary’ therapies. The GDG considered that no recommendations could be made based on the available evidence. Therefore, the review is presented here for completeness only.

15.3.9.1. Evidence statements

(Table 15.39)

Table 15.39. Evidence statement and grading.

Table 15.39

Evidence statement and grading.

15.3.9.2. Evidence review on professionally organised therapies

Professionally organised therapies as defined by the House of Commons are: acupuncture, chiropractic, herbal medicine, homeopathy and osteopathy. We found one high-quality systematic review that was both comprehensive and recent.289 This paper, published by Pittler and Ernst in 2005, aimed to review complementary therapies for reducing body weight, and included six systematic reviews and meta-analyses based on RCTs and 25 additional RCTs. Only one of the included RCTs was eligible for inclusion against our established criteria for treatment and follow-up duration of a minimum of 12 months for adults, reporting an 18-month trial. No studies on the effectiveness of these therapies on children or adolescents were retrieved. Furthermore, no studies on the effectiveness of chiropractic or osteopathy in the treatment of obesity were found.

Acupuncture/acupressure

Pittler and Ernst referred to one systematic review that included four sham controlled RCTs.290 Two of the trials reported a decrease in hunger, and the other two did not report any difference in body weight compared with the sham acupuncture. The review concluded that the effect of acupuncture or acupressure for weight loss was not based in the results of rigorous clinical studies. Similar conclusions were drawn in another non-systematic review.291

Dietary supplements

Although Pittler and Ernst reviewed the use of over-the-counter dietary, supplements, this was outside the scope of the guidance.

Homeopathy

The authors reported that two RCTs were identified. In one study, Helianthus tuberosus D1 was given to patients with a mean BMI of 29 kg/m2, and after 3 months the treatment groups had lost a mean 7.1 kg which was significantly different from the placebo group. The other trial, which aimed to study Thyroidinum 30 cH, did not reach any difference between treatment group and placebo.

Hypnotherapy

The authors reported one meta-analysis (that included six RCTs) that aimed to compare hypnotherapy plus cognitive behaviour therapy with cognitive behaviour therapy alone. The results showed that adding hypnotherapy to cognitive behaviour therapy only slightly decreased body weight. The authors also referred to a further RCT with hypnotherapy directed at either stress reduction or energy intake, compared to dietary advice. Results showed a significantly greater weight reduction compared with control groups.

15.3.10. Effectiveness of brief interventions in primary care and other general clinical settings in improving outcomes for people who are overweight and obese

15.3.10.1. Evidence statements

(Table 15.40)

Table 15.40. Evidence statements and grading.

Table 15.40

Evidence statements and grading.

15.3.10.2. Evidence review on brief interventions

This section reviews evidence on the effectiveness of brief interventions in primary care and other general clinical settings in improving outcomes for adults who are overweight and obese.

Types of study
  • RCTs
  • clinical controlled trials based in the UK (corroborative evidence).
Types of participant
  • Children and adults who were overweight or obese.
Types of outcome
  • primary outcomes to be measured (not self-reported) estimates of overweight in per cent and BMI
  • secondary outcomes to be behaviour change, participants’ views, measures of self-esteem, health status, well-being and quality of life.

Only RCTs with a minimum duration of 12 months (including follow-up, 6 months for children) that aimed to assess the effectiveness of brief interventions in primary care and other general clinical settings in improving outcomes for adults who are overweight and obese, were included. Moreover, studies were required to specifically include overweight and/or obese participants.

The majority of studies on such interventions were chiefly in alcohol use disorders and smoking cessation. Beyond the lack of studies that met the above parameters, another challenge was how to define a ‘brief intervention’. Across the literature, different studies have given diverse definitions of ‘brief interventions’.292–293 As an example, Babor294 defines one contact as ‘minimal’, one to three sessions as ‘brief’, five to seven sessions as ‘moderate’ and eight or more sessions as ‘intensive’ treatment (Moyer et al. 2002293). What can also happen is that what is mean by ‘brief’ can be reported as ‘extended’ in another study (Jonson et al.292 quoted by Moyer et al.293). Despite these contradictions in the available literature, we chose to use Babor’s definition,294 by including studies with no more than four sessions.

From our searches we retrieved three studies that matched our inclusion criteria. However, two of these were excluded after discussion, as one was based in a blood pressure clinic and was published in 1978, and the other study included non-overweight participants and did not perform subgroup analysis.

Pritchard and coworkers18 studied the clinical and cost outcomes of providing nutritional counselling to patients aged between 25 and 65 years with either hypertension, type 2 diabetes or were overweight (BMI > 25 kg/m2). Results of the study were stratified per group, thus its inclusion in the review. Two groups (dietitian and doctor/dietitian) were given counselling focused on principles of good nutrition and exercise and addressed problem areas in lifestyle and dietary patterns. In the latter group, the participants were also seen by the GP at baseline and by the same GP on two other occasions during the 12 months for 5 minutes each time to encourage and monitor the participant. The control group (which matches Babor’s definition of brief intervention) received results of initial screening and were advised to discuss their queries with the GP at appointment. They received usual care (including monitoring, advice and prescriptions) from GP but no dietitian counselling. At 12 months, mean ± SD weight change in kilograms was −5.10 ± 7.36 kg in the dietitian group, 0.60 ± 6.08 kg in the control and −6.20 ± 7.67 kg in the dietitian and doctor group. Weight change in both intervention groups was significant (−5.70 kg [95% CI −8.05 to −3.35]; −6.80 kg [95% CI −9.17 to −4.43], respectively) compared to the control group. Moreover, significant decreases in mean blood pressure in participants with hypertension were reported in both intervention groups compared with the control group.18

No studies including children were identified.

Conceptual input from Professor Jane Wardle.

Lifestyle exercise relates to integrating exercise into the person’s/children’s lifestyle without the focus on exercise intensity. It can be walking or cycling to school, walking up and down stairs or walking at lunch.

For further details, see evidence tables for each study.

The majority of the studies from Epstein and coworkers (which are extensively used as evidence throughout interventions to treat childhood obesity) are based on the traffic light diet. This is a calorie-based food-exchange system. Foods are divided into five groups (fruits and vegetables, grains, proteins, dairy and other foods) and the foods in each group are colour coded according to nutrient density: green for ‘go’, yellow for ‘eat with care’; and red for ‘stop’. Green foods are foods containing fewer than 20 calories per serving, yellow foods are the staple of the diet and provide most of the basic nutrition, and red foods are those foods high in fat and simple carbohydrates. All sweets and sugared beverages are classified as red foods. Families are then instructed to count calories and cannot have more than four red foods a week.

Piers EV. Piers–Harris children’s self-concept scale revised manual. Los Angeles, CA: Western Psychological Services, 1984.

Both Braet and coworkers57,58 other outcomes are in section 15.2.8.

Both Braet and coworkers57,58 other outcomes are in section 15.2.8.

We have excluded studies based solely on VGB, although included those which combine VGB with other surgery techniques, or have one group that was given VGB.

We have excluded studies based solely on VGB, although included those which combine VGB with other surgery techniques, or have one group that was given VGB.

These studies do not include evidence tables, as they are older procedures, and are not used in the NHS.

Towbin A, Inge TH, Garcia VF et al. Beriberi after gastric bypass surgery in adolescence. Journal of Pediatrics 2004;145:263–7.

Due to lack of reporting, healthy eating advice and diets where the fat or calorie restriction was not estimable are included in this category.

Recalculated with sample size halved for control group in Pritchard 199918

Defined as a diet with 40 g or less of carbohydrate, irrespective of calorie content.

Defined as a diet with 40g or less of carbohydrate, irrespective of calorie content

Problems with HTA analysis, but recalculated figures were still non-significant.

Defined as a diet with 40 g or less of carbohydrate, irrespective of calorie content.

Defined as a diet with 40 g or less of carbohydrate, irrespective of calorie content.

Detailed information on previous technology appraisals are described in section 15.3.11.

Footnotes

*

Recommendations on classification deleted as not relevant to this review.

Watching television, playing computer games, imaginative play, talking on the phone, and playing board games.

Lifestyle exercise relates to integrating exercise into the person’s lifestyle without the focus on exercise intensity. It can be walking or cycling to school, walking up and down stairs or walking at lunch.

§

Conceptual input from Professor Jane Wardle.

**

Lifestyle exercise relates to integrating exercise into the person’s/children’s lifestyle without the focus on exercise intensity. It can be walking or cycling to school, walking up and down stairs or walking at lunch.

††

For further details, see evidence tables for each study.

‡‡

The majority of the studies from Epstein and coworkers (which are extensively used as evidence throughout interventions to treat childhood obesity) are based on the traffic light diet. This is a calorie-based food-exchange system. Foods are divided into five groups (fruits and vegetables, grains, proteins, dairy and other foods) and the foods in each group are colour coded according to nutrient density: green for ‘go’, yellow for ‘eat with care’; and red for ‘stop’. Green foods are foods containing fewer than 20 calories per serving, yellow foods are the staple of the diet and provide most of the basic nutrition, and red foods are those foods high in fat and simple carbohydrates. All sweets and sugared beverages are classified as red foods. Families are then instructed to count calories and cannot have more than four red foods a week.

§§

Piers EV. Piers–Harris children’s self-concept scale revised manual. Los Angeles, CA: Western Psychological Services, 1984.

***

Both Braet and coworkers57,58 other outcomes are in section 15.2.8.

†††

Both Braet and coworkers57,58 other outcomes are in section 15.2.8.

‡‡‡

We have excluded studies based solely on VGB, although included those which combine VGB with other surgery techniques, or have one group that was given VGB.

§§§

We have excluded studies based solely on VGB, although included those which combine VGB with other surgery techniques, or have one group that was given VGB.

****

These studies do not include evidence tables, as they are older procedures, and are not used in the NHS.

††††

Towbin A, Inge TH, Garcia VF et al. Beriberi after gastric bypass surgery in adolescence. Journal of Pediatrics 2004;145:263–7.

*****

Due to lack of reporting, healthy eating advice and diets where the fat or calorie restriction was not estimable are included in this category.

†††††

Recalculated with sample size halved for control group in Pritchard 199918

‡‡‡‡

Defined as a diet with 40 g or less of carbohydrate, irrespective of calorie content.

§§§§

Defined as a diet with 40g or less of carbohydrate, irrespective of calorie content

******

Problems with HTA analysis, but recalculated figures were still non-significant.

††††††

Defined as a diet with 40 g or less of carbohydrate, irrespective of calorie content.

‡‡‡‡‡

Defined as a diet with 40 g or less of carbohydrate, irrespective of calorie content.

Detailed information on previous technology appraisals are described in section 15.3.11.

Copyright © 2006, National Institute for Health and Clinical Excellence.
Bookshelf ID: NBK63705
PubReader format: click here to try

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (7.3M)
  • Disable Glossary Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...