aRelative risk of death from all causes adjusted for age, smoking, menopausal status, use of oral contraceptives and post-menopausal hormones, and parental history of myocardial infarction before age 60 years.
^b95% confidence interval for relative risk.
^cExcluding first four years of follow-up and women with > 4 kg weight change during those four years.

On inspection of the relative risk of death during the 14-year follow-up of these women it appears that those in the whole group with a BMI between 19 and 27 kg/m² have the least mortality risk (0.8 relative to 1.0 for those with a BMI < 19). Only in the range 27.0-28.9 does the risk rise to the level of those < 19 kg/m², and above 29 the risk is definitely increased. However, when those women who had never smoked are analysed separately, a different story emerges. Among non-smokers the mortality risk starts to increase at 22 kg/m² and is significantly increased at 27 kg/m², so the increased mortality among the women < 19 kg/m² disappears when the smokers are removed from the analysis. An even greater change occurs if the analysis is restricted to those women who had never smoked, and who had not either died or gained > 4 kg in the first four years of follow-up. Now the J-shaped curve has disappeared, the minimum mortality is with the thinnest women, and a significant increase in mortality risk occurs above 27 kg/m². Some of the women who died within four years of enrolment in the survey probably had a disease at that time, which may explain why they were thin and also why they died. This example has been considered in some detail because crude data on mortality do not provide reliable information on the health risks of obesity if the confounding effects of cigarette smoking, adult weight gain and previous disease are not allowed for. If these factors are removed, a woman aged 30-55 years is more than twice as likely to die in the next 14 years if her BMI is > 29 kg/m² than if it was < 19 kg/m².

Top of Page

The effect of age on excess mortality from all causes associated with obesity is controversial. In a large cohort of obese persons (n = 6193), obesity-related excess mortality declined with age at all levels of obesity.(14) This is not because obesity causes less ill health in older people, but because in older people death from causes unrelated to obesity become more common.

It is important to note that the obesity-related health risks cited above do not necessarily apply to populations of different ethnic origins. However, a recent systematic review (15) concluded that differences in absolute risk by ethnicity are not relevant to individuals in a clinical setting.

It is also important to note that there is evidence which suggests that the health benefits of leanness are limited to fit men, and being fit may reduce the hazards of obesity.(16)

Top of Page

Obesity: socioeconomic consequences

From the viewpoints of human suffering and healthcare expenditure, morbidity is as important as mortality. Obesity causes disability as well as death. For example, Rissanen et al.(17) analysed data from the Finnish Social Security system and found that the risk of drawing a disability pension increased significantly with BMI, even within the 'desirable' range of 20-25 kg/m²; one quarter of all disability pensions in women and half as many in men. These disabilities arose mainly from cardiovascular and musculoskeletal disease.

The impact of obesity on a number of specific diseases are outlined below.

Obesity, insulin sensitivity and diabetes

Diabetes is not as impressive a cause of mortality among obese people as heart disease, but it is itself a risk factor for heart disease, and also a very important cause of morbidity from neuropathy, nephropathy and eye disease.

A man more than 140% of average weight is 5.2 times more likely to die of diabetes than a normal-weight man, and for women the mortality ratio is 7.9 times for a similar degree of overweight.(18) A classic study of experimental obesity in Vermont has shown that the association between obesity and reduced insulin sensitivity (which is the primary problem in non-insulin-depndent diabetes mellitus, NIDDM) is a causal one. Young male volunteers, with no family history of diabetes or obesity, overate for six months so they increased their weight by 21%, of which 73% was fat, and they then showed significant changes in biochemistry in the direction of diabetes. After weight loss to normal values these changes reverted to normal.(19) Even among pre-pubertal children obesity is associated with peripheral and hepatic insulin resistance.(20)

Although diabetes is not directly the cause of most of the excess mortality among obese people, the metabolic defect underlying NIDDM is clearly the result of obesity,(21) which itself predisposes to hypertension and heart disease. These defects are reversible with weight loss, with corresponding improvement in mortality. A deliberate weight loss of 0.5-9.0 kg is associated with a 30-40% reduction in diabetes-related mortality.(22)

Coronary heart disease

The main cause of the excess mortality among obese people is coronary heart disease (CHD). Obesity is itself strongly related to hypertension and stroke, particularly in young people.(23) However, these risk factors improve when obese people lose weight.(24) After adjustment for age and smoking, the risk of a fatal or non-fatal myocardial infarction (MI) among women > 29 kg/m² is three times that among lean women.(13),(25) High blood pressure, high triacylglycerol and high low-density lipoproteins favour the formation of atheromatous lesions, but obese people have the added hazard of abnormalities of blood clotting factors, which further increase the risk of thrombosis and MI.(26) These abnormalities improve with therapeutic weight reduction.(27)

Top of Page

Cancer

A very large survey by the American Cancer Society found that the mortality ratio for cancer among men who were 40% overweight was 1.33, and for women 1.55. The most important increase is for breast cancer in post-menopausal women, but there is also an increased risk of cancer of the endometrium, uterus, cervix ovary and gall bladder in women, and of colon, rectum and prostate in men.(28) Intentional weight loss of 0.5-9.0 kg is associated with a decrease of 40-50% in mortality from obesity-related cancers.(22)

Osteoarthritis

Degenerative disease of weight-bearing joints is a very common complication of obesity, particularly in the knees of middle-aged women and causes significant disability.(17) Unlike the risk of heart disease or diabetes, the risk of osteoarthritis is related to the total amount of fat, and not in particular to the amount of intra-abdominal fat.(29)

Gallstones

Obese people have a higher output of cholesterol in bile, with a lower concentration of bile salts, so their bile is constantly in danger of forming gallstones. Rapid weight loss increases the release of cholesterol from adipose tissue, and hence increases the load to be excreted in bile.(30) Bile stasis contributes to biliary infections, and also to the risk of gall-bladder cancer, mentioned above.

Reproductive disorders

Obesity is associated with disorders of menstrual function, fertility and childbirth. Examination of obese pregnant women is particularly difficult, whether by abdominal palpation, ultrasound or laparoscopy. The difficulty in monitoring fetal wellbeing in severely obese mothers partly explains their increased rate of caesarian section. However, even moderate degrees of obesity are associated with an increased incidence of hypertension, toxaemia, gestational diabetes, urinary tract infections and fetal macrosomia.(31) There is also an increased risk of neural tube defects in the children of obese mothers: among women over 70 kg in weight dietary intake of folic acid does not have the same protective effect as it has in leaner women.(32),(33)

Top of Page

Sleep apnoea

Obesity causes inefficiency of respiratory function by several mechanisms. The mechanical load of fat on the chest wall increases the mechanical work of inspiration, especially when the subject is recumbent, and a large mass of intra-abdominal fat tends to push the liver upwards, thus decreasing the intrathoracic space. There is also a mismatch of pulmonary ventilation and perfusion, so much of the blood flowing through the lung capillaries is at the base of the lung, where ventilation is poor. These problems may cause the Pickwickian syndrome of chronic hypoxia and carbon dioxide retention, which may manifest itself as inappropriate somnolence, vividly described by Dickens in the fat boy in Pickwick papers, and obstructive sleep apnoea (OSA). This is a serious condition, which is associated with pulmonary hypertension and right-sided heart failure. Data from the Swedish Obese Subjects (SOS) study show that OSA was an important contributor to morbidity in severe obesity, and contributed to cardiovascular mortality.(34) Respiratory function improves when obese people lose weight.(35)

Psychological and social disorders

The health hazards of the obese person which have been listed above, become increasingly evident as the person becomes older: heart disease, hypertension, stroke, osteoarthritis, cancer and gallstones are all conditions which occur mainly in older people, so the obese young person does not experience these as a threat. However, the psychological and social penalties of obesity fall mainly on the child and young adult. Indeed, there is a view, often promoted in the media, that the penalties of obesity are mainly due to social discrimination, so if society treated obese people with respect and tolerance it would cease to be a problem.(36) This is wrong: the health hazards of obesity listed above would remain, however respectfully obese people were treated, but there is compelling evidence that our society discriminates against fat people.(37)

Social discrimination continues through adult life. Sonne-Holm and Sorensen (38) showed that, for a given parental social class, intelligence and education, severely obese people achieved less favourable social status than non-obese people. In the US, Gortmaker et al.(39) studied a nationally representative sample of 10, 039 men and women who were 16 to 24 years old in 1981, and obtained follow-up data on 65-79% of the cohort seven years later. Women who were initially above the 95th centile for BMI had completed fewer years in school, were less likely to be married and had higher rates of household poverty than the women who had not been overweight, independent of their baseline socioeconomic status and aptitude-test scores. However, people with chronic conditions such as asthma and musculo-skeletal abnormalities did not differ from non-overweight people in these ways.

In the general population, those who are overweight or obese are not significantly more depressed than lean people.(37) Among those who are depressed, it is difficult to establish if this is caused by obesity or caused by dieting in an unsuccessful attempt to reverse the obesity. Among the severely obese, there have been numerous reports of psychopathology: for example volunteers for the SOS study, both men and women, showed very poor ratings for mental wellbeing, and more symptoms of anxiety and depression than the reference population. The score on psychometric scales were as bad as, or worse than, those of patients with chronic pain, generalised malignant melanoma or tetraplegia after neck injury.(40) It is particularly important for doctors or dietitians who are treating obese patients to remember that their duty is to help to restore the self-esteem of obese patients, as well as to help them to lose their excess weight. This point is considered again when discussing treatment of obesity.

It is useful to note that there is some evidence that weight concerns vary with ethnic group.(41),(42)

All of these penalties of obesity decrease with weight loss, with the exception of the risk of gallstone formation. During weight loss in an obese person, the cholesterol in adipose tissue is mobilised and the bile may become even more liable to form cholesterol stones.

Top of Page

Fat distribution and health risk

A study in Gothenburg, Sweden, showed that people with a high WHR (indicating that fat was largely in the abdominal cavity, rather than subcutaneously on the limbs) had a greater risk of heart disease and diabetes than people with a similar amount of fat distributed peripherally.(43) This probably relates to the insulin insensitivity which is caused by a high flux of free fatty acids in the portal circulation, because intra-abdominal fat cells can release fatty acids very rapidly. However, further studies have shown that the increased mortality among men was not significantly related to WHR when the follow-up period was extended to 20 years,(44) and the central distribution of fat is associated with both cigarette smoking and a high alcohol intake, which may have contributed to some of the observed excess mortality risk.

Children

Effect on adult morbidity and mortality

There is limited evidence for an association between adolescent obesity and increased risk of adult morbidity (5) and mortality.(10) Must et al.(10) showed that obesity in adolescence predicts a broad range of adverse health effects that are independent of adult weight after 55 years of follow-up: the risks of morbidity from CHD and atherosclerosis are increased among men and women who were overweight as adolescents.

There is better evidence that the childhood period is important for adult obesity (and all the associated health risks, as described above) because tracking of overweight, albeit moderate, is observed between childhood and adulthood. This topic has recently been reviewed.(5),(45) The figures vary according to the definition of obesity and length of follow-up, but fat children have a high risk of going on to become fat adults. For example, in the 1958 British birth cohort, 38% of boys and 44% of girls above the 95th BMI centile at age 7, were obese at age 33.(46) Even so, only a small proportion of fat adults were fat in childhood. It is likely that there are factors operating in early adulthood that promote obesity, but there may also be factors operating in childhood that promote adult obesity. It is still a matter of debate whether there are particular stages in childhood, during which physiological alterations increase the risk of later obesity. These stages are termed critical periods, and may include the prenatal period, the adiposity rebound (second rise in adiposity occurring at about 6 years), and puberty.(47)

Top of Page

Parental fatness has also been identified by a recent systematic review as the most important predictor in childhood of adult obesity,(9) although the contribution of genes and inherited lifestyle factors to the parent-child fatness association remains largely unknown. Other important risk factors in childhood of adult obesity included social factors, birth weight, timing or rate of maturation, physical activity, dietary factors and other behavioural or psychological factors. The relationship between low socioeconomic status (SES) in childhood and increased fatness in adulthood is remarkably consistent, but when fatness is measured in childhood, the association with SES is less consistent. Studies investigating SES were generally large, but very few considered confounding by parental fatness. Women who change social class (social mobility) show the prevalence of obesity of the class they join, an association which is not present in men, and the influence of other social factors such as family size, number of parents at home and childcare have been little researched. Parsons et al.(9) found good evidence from large and reasonably long-term studies, for an apparently clear relationship for increased fatness with higher birth weight, but in studies which attempted to address potential confounding by gestational age, parental fatness or social group, the relationship was less consistent. The relationship between earlier maturation and greater subsequent fatness was investigated in predominantly smaller, but also a few large studies. Again, this relationship appeared to be consistent, but in general, the studies had not investigated whether there was confounding by other factors, including parental fatness, SES, earlier fatness in childhood, or dietary or activity behaviours. Studies investigating the role of diet or activity were generally small, and included diverse methods of risk factor measurement. There was almost no evidence for an influence of activity in infancy on later fatness, and inconsistent but suggestive evidence for a protective effect of activity in childhood on later fatness. No clear evidence for an effect of infant feeding on later fatness emerged, but follow-up to adulthood was rare, with only one study measuring fatness after the age of 7 years. Again, confounding variables were seldom accounted for. A few, diverse studies investigated associations between behaviour or psychological factors and fatness, but mechanisms through which energy balance might be influenced were rarely addressed.

Effects during childhood

Obese children are more prone to physical ailments, and are also liable to underperform at school relative to their potential. The health consequences of obesity in youth have recently been reviewed (48); increased blood lipids, glucose intolerance, hypertension and increases in liver enzymes associated with fatty liver, have all been observed to be more common in obese children or adolescents. The diseases to which obese children are more liable are tabulated in Table 4.(3)

Top of Page

The main penalties of obesity experienced by children are social isolation and peer problems.(49) Although there is little evidence to suggest that self-esteem is significantly affected in obese young children, on reaching the teenage years the effect is striking. Obese children are believed by their peers at school to be lazy, dirty, stupid, ugly, cheats and liars, and these perceptions have been reported by children as young as 9 years old.(50)

Overweight adolescent women have a lower educational attainment, lower incomes and are less likely to marry than those not overweight.(39) If these relationships are indeed causal, then they imply far-reaching consequences for costs to health services, and the total healthcare.

Why does the NHS think that obesity is not important?

The dramatic effect which obesity has on quality of life, morbidity and mortality (and equally how these can be reversed with weight loss) has resulted in repeated and strongly worded reports by national and international expert committees that 'obesity is one of the most important public health hazards of our time'.(3),(15),(51-53) There is a stark contrast between these authoritative warning, and the absence in any country of any effective preventative or treatment programme to reverse the increasing prevalence of obesity. In the UK, a target for obesity was omitted from Our Healthier Nation: saving lives.(54) The Health Education Authority (HEA) has not developed a campaign for the treatment of obesity. They did run a campaign to increase levels of physical activity in 1997,(55) which may increase physical fitness and help maintenance of weight loss, but cannot be expected to reduce body weight significantly in obese people. No national clinical guidelines for the treatment of obesity have been developed the UK (although they have for Scotland; SIGN obesity guidelines (52)). Indeed the most important campaign to date is probably the Fighting fit, fighting fat campaign run by the BBC in spring 1999, which tells us something of the paradox between health service providers and the public demand for help in losing weight.

So why does the health service refuse to treat obesity seriously? We do not have hard evidence on which to base an answer to this crucial question, but believe there are several factors that, in varying proportion, inhibit health services from controlling obesity. Some of these obstacles have a factual basis, others are based in misapprehension of the facts. If progress is to be made these obstacles need to be removed. Their nature and basis is briefly reviewed below.

Top of Page

'Obesity per se is not a health hazard'

In the 1950s, international epidemiology was in its infancy. Keys and colleagues set up a prospective study of the factors which predisposed to CHD. They recruited healthy men aged 40-59 years in seven countries (USA, Japan and five European countries), and found after 15 years of ollow-up that adiposity did not significantly predict mortality if age, cigarette smoking, blood pressure and cholesterol were already entered into a multiple regression equation.(56) This finding led public health policy makers to the view that obesity per se was benign, and that the real villains were smoking, blood pressure and cholesterol. The flaw in the argument, as described above, is that obesity itself predisposes to hypertension and hypercholesterolaemia, but even when these factors are allowed for, obesity remains an independent risk factor for heart disease and total mortality both in men (57) and women.(25)

'Obesity is genetically determined, and hence untreatable'

The genetics of obesity in some laboratory rodents has been extensively explored, and it has been shown that the genetically determined absence of 'leptin' entirely accounts for the obesity of some strains of mouse. The role of leptin in the aetiology of obesity in human subjects is not clear: obese people have high, not low, leptin concentrations. There is also good evidence from twin studies that the susceptibility to weight gain in a given environment is affected by hereditary factors, probably by the interaction of many genes. However, it is obvious that the recent doubling of the prevalence of obesity in the UK cannot be ascribed to a change in the genetic make-up of the population because genetic make-up can change over generations, not during the lifetime of an individual. Obesity is determined by the interaction between genes and environment, and is greatly influenced by psychosocial factors and cognitive actions. Environmental and lifestyle factors can certainly be changed, and it is by this route that obesity can be prevented or treated in the individual and in the community.

'Fat is a feminist issue'

Orbach (58) made the valid point that many young women struggled vainly to achieve unphysiological thinness in order to meet a feminine stereotype which was thought to be attractive, and they would be better if they gave up this futile endeavour. This is true, and it should remind healthcarers that it is important not to damage the self-esteem of obese people when trying to help them to lose weight. However, it is not a valid argument against providing help for those who are obese, so they can avoid the health hazards listed above.
Orbach's original thesis has been taken further by pressure groups who claim that health education campaigns designed to control obesity are a form of unfair discrimination, analogous to racism, and should be made illegal. Health educators who say (truthfully) that obesity is unhealthy, and can be avoided or reversed by lifestyle changes, may be accused of 'victim blaming'. No doubt this in part explains the very muted response of organisations such as the HEA to the obesity problem.

Top of Page

Media misinformation

There is huge public demand (described below) for an easy solution to the problem of weight loss, so it is not surprising that an unlimited amount of bizarre advice, magic potions and pseudo-science is on offer to the bemused consumer. This is part of the price we pay for freedom of expression, but it is an obstacle to those who try to offer reliable guidance about the health hazards of obesity, and what can and should be done to avoid them.

Public demand and putting the NHS services for obesity into context

Demand may be defined as what the public would be willing to pay for, or might wish to use, in a system of free healthcare. The public demand for help to acquire the perfect body is immense; it is a national obsession, particularly among women. This is one of the key differences in the planning of obesity health services as compared with those for other diseases. The public demand is catered for by the many sources of help from outside the health service. Help in the form of specialist magazines and books, slimming clubs and slimming foods have soared in the past 20 years. The fact that so many people within England seek advice on obesity from outside the health service highlights two important issues for health services commissioners.

• Most individuals who would benefit from losing weight (currently at 17-20% of population; see section 4) probably use non-NHS services, but only a minority are also offered/demand help by the NHS. Unfortunately the constant background information on dieting from other sources is usually more compelling. The patient may request advice from the health professional regarding a specific diet or food which they have heard will enhance their weight loss (or even cause weight loss without the tedium of dieting). Regardless of the credibility of reported claims for such diets or foods, it is important that the health professional addresses them seriously and does not dismiss them as simply ridiculous; if the patient thought that these claims were ridiculous he or she would not have asked for advice on them in the first place. We suggest that effective care in this field requires provision of services by both the NHS and others working in close collaboration.
• Demand for obesity services is greater than the need for them since many individuals who seek advice on weight management are not overweight or obese. However, it is also important to recognise that these individuals might benefit from sound advice on lifestyle issues through the health promotion services.

Children are quite different from adults in terms of non-NHS sources of help on weight management. The sources mentioned above are targeted only at adults, and this provides the NHS with a niche on this childhood obesity 'market'. Earlier in this section we highlighted the risks of childhood obesity, and in section 7 we suggest ways in which the NHS may deliver services to this vulnerable group.

Top of Page

Prospect for new effective treatments

Readers of this chapter will be aware that the media are constantly heralding a 'breakthrough' in the treatment or prevention of obesity, since such items are effective in selling newspapers. The two developments that have received most attention recently are the discovery of the hormone leptin and the licensing of a new type of anti-obesity drug orlistat. The evidence relating to these newly discovered compounds is reviewed later in the chapter. The purpose of this short note is to consider the possibility that they will revolutionise the management of obesity and thus render the contents of this chapter obsolete.

Leptin is a hormone, released from adipose tissue, which has the effect of reducing food intake in genetically obese mice, and curing their obesity and infertility. Initially, therefore, there was great optimism that it would have similar effects in obese human subjects. It was found, however, that obese human subjects did not (like obese mice) have abnormally low levels of leptin, but abnormally high levels. So far (except in a single case of a rare genetic disorder) leptin has not been shown to be therapeutically effective in human obesity.

Orlistat is a drug that inactivates the enzymes that digest fat in the human small intestine, and thus reduces fat absorption by about 30%. The results of RCTs are reviewed in section 6. In very large, multicentre, international trials orlistat has been shown to cause greater weight loss in obese subjects on a low-fat diet than that observed in control subjects; this difference is statistically significant (since the trials involved about 1000 subjects) but clinically not very impressive.

Sibutramine is a drug not yet licenced for the treatment of obesity in the UK, but trials have been reported from France (59) and from the USA.(60) In the French trial, patients with BMI > 30 were screened using a four-week treatment on a very-low-calorie diet; only those who lost > 6 kg in this phase entered the trial. An intention to treat analysis showed the mean weight change after one year among 81 patients on sibutramine (10 mg) was -5.2 kg, and among 78 patients on placebo was +0.5 kg. In the American trial, 1463 patients were screened, 1047 were randomised and 683 completed the 24-week study. The weight loss at completion ranged from 1.2% in the placebo group to 9.4% among those on a dose of 30 mg/day. These results show that sibutramine causes weight loss which is statistically greater than placebo, but not impressive compared with the initial overweight of the volunteers, and with a large variation in response between individuals.

We are confident that, at least for the next decade, the health problems associated with obesity, and the methods available for effective treatment, will be little changed from the present situation. We do not expect that a new therapeutic 'breakthrough' will greatly affect the assumptions on which this chapter is based.

Top of Page

3 Subcategories and risk

Subcategories of obesity related to risk of associated diseases

Health risk, in terms of mortality, increases progressively from normality to very severe obesity (section 2, Figure 2). At each weight, this health risk is greatest in young people, people with diabetes or hypertension, people with a relatively high WHR. The costs of obesity, both to the individual and to the community, arise mainly from the co-morbidities rather than from obesity itself (see section 5).

From the viewpoint of healthcare planners there is a dilemma. One strategy (which has been widely adopted in the past) is to plan to treat the co-morbidities - heart disease, stroke, hypertension, diabetes, osteoarthritis, gallstones, certain cancers, reproductive disorders, sleep apnoea, psychological and social disorders - and ignore the underlying obesity. This strategy is superficially plausible, since patients with these co-morbidites clearly need treatment. However, experience shows that the strategy is expensive and ineffective if the underlying obesity is allowed to increase.

An alternative strategy is to seek to prevent the development of obesity by health education campaigns that promote physical activity and healthy diets of low energy density. This also fails for several reasons. First, people will not adopt healthier lifestyles unless the facilities are available to modify their diet and exercise more in conditions that are affordable and safe, and unless they have a clear understanding of the relationship between overweight and health risk. These requirements are not met at present. Second, a campaign to prevent overweight and obesity is inadequate to meet the needs of a population in which half are already overweight and one fifth are already obese. Third, campaigns that exhort adults to 'fight the flab' do not address the problem of increasing obesity among children. Government and health authorities potentially have some control over the diet and physical activity of schoolchildren, but this opportunity is not being effectively used to prevent obesity at its earliest stage.

The solution to the dilemma lies in the ability of healthcare planners to take a broader view of the problem of obesity, and integrate its three subcategories. These are:

• Obese patients with existing co-morbidities. These are the most visible component of the problem, to which the most attention is at present paid. With better management of the other two components there will be fewer patients in this category, and so costs in this sector will decrease, and the level of public health will improve.
• Obese individuals who do not yet have co-morbidities. These are the people who are at present least well served by the NHS.
• Adults and children who are not obese, but particularly at risk to become obese (see below).

The evidence for categorising the health risk in children by degree of obesity is not available. Section 7 describes the evidence which suggests that primary school children, compared with younger or older children, are more effectively managed.

Top of Page

Subcategories of the non-obese population who are most likely to become obese (predictors of obesity)

It is important to state at the outset that anybody is at risk of becoming obese, so long as they consume more energy than they use. Adults who are most likely to become obese are:

• those who were previously obese and who have lost weight
• smokers who have stopped smoking
• those who change from an active to an inactive lifestyle
• those with poor educational achievement.

Children at high risk of becoming obese adults (as detailed in section 2) are:

• children who have high weight-for-height
• children who have obese parents.

  Top of Page

4 Prevalence and incidence

Adults

The prevalence of obesity, particularly in the developed world, has increased particularly rapidly over the past two decades. The WHO calls it a 'global epidemic'.(3)

There are good data from National Health Examinations showing prevalence in the UK and elsewhere (Table 5). The prevalence of overweight and obesity in the UK population was first determined in 1980, in a survey of a representative sample of 5000 man and 5000 women aged 16-64 years. At that time, the proportion of men in the 'pre-obese' stage was 34%, and 6% were obese; for women there were 24% pre-obese and 8% obese.

The prevalence was higher among the older subjects. Another survey carried out in 1987 using the same methodology showed an alarming increase in the prevalence of obesity (BMI > 30), which overall had increased from 6 to 8% in men and from 8 to 12% in women. The increase occurred in all age groups, but particularly among women aged 25-34 years, in whom the prevalence appears to have doubled over the seven-year interval. There is now a regular cycle of surveys which continue to show an increasing prevalence of obesity: the trend up to 1997 (which is the latest year for which data are available at the time of writing (61) is shown in Table 5. The data on prevalence become available about two years after the fieldwork is done, and the prevalence among adult men and women in the UK has increased by approximately one percentage point annually over the past decade. It is therefore reasonable to estimate that the prevalence will be more than 20% among both men and women by the year 2000. Serial surveys in other countries show a similar trend.

Top of Page

It is evident that since the first survey in 1980, the prevalence of obesity in the UK has more than doubled among both men and women. The general level of obesity in affluent countries ranges from the strikingly low values in Japan, through moderate levels in Netherlands and Sweden, to the high levels in USA and Australia. Among less affluent countries there is also a huge range from low levels in Brazil to astonishingly high levels in Samoa. Despite these differences the trend towards increasing obesity is evident everywhere, whatever the baseline level.

Relation to age, social class and region. Data for adults and children in England in 1996 taken from the Health Survey of England 1996 (62)

Adults

Sex

Overall, mean BMI is 26.3 kg/m² in men and 26.0 kg/m² in women. Among those aged 16-64 years, mean BMI is 26.2 kg/m² in men and 25.8 kg/m² in women. Overall, 61% of men and 52% of women are overweight or obese. A greater proportion of men (45%) than women (34%) are overweight, but slightly lower proportions of men (16%) than women (18%) are obese. Among those aged 16-64 years, 16% of men and 17% of women are obese.

Age

Men tend to reach maximum prevalence of obesity about age 45 years, but in women the prevalence increases to age 65 years and then starts to decline.

Social class

The prevalence of obesity is inversely related to social class, but there is no satisfactory explanation for this observed association.

Region

There are no marked regional differences in the prevalence of obesity in the UK when the effects of age, social class and smoking habit are allowed for. In men, the prevalence of being overweight is lowest in West Midlands (39%) and highest in North Thames and Northern and Yorkshire (45%), while in women it is lowest in North Thames (31%) and highest in West Midlands (36%). The prevalence of obesity is lowest in North Thames in men (13%) and both North and South Thames in women (17%), and highest in Trent and West Midlands in men (18%) and Trent in women (20%).

Top of Page

Smoking and ethnicity

Data on obesity by smoking or ethnicity is not published in the Health Survey of England 1996 report.(62)

Incidence

Incidenceof obesity is impossible to measure, since there is so much variability in weight with time in any individual. At any moment there will be quite a lot of people oscillating above and below any given threshold of severity.

Children

Sex and age

Using BMI reference curves recently derived for UK children, 4% of boys and 4% of girls aged 2-15 are above the 98th percentile. Mean BMI falls between age 2 and age 5 or 6. It then increases fairly rapidly to adulthood. Compared with males, females have a slightly higher BMI during adolescence.

Social class, region, smoking and ethnicity

Data on obesity by social class, region, smoking and ethnicity is not published in the Health Survey of England 1996 report.(62)

Incidence

Incidence of obesity is even more difficult to measure in children since variability in weight and height is greater during growth.

Top of Page

5 Services available and their cost

What happens at present and why?

Obesity has features that do not apply to the other major causes of ill health and disability (e.g. hypertension, cancer or diabetes) with which the health service should cope. A patient with such an illness has probably been diagnosed by a doctor, and looks to medical aid (NHS or private or alternative) for treatment. The problem is one that is known only to those whom the patient chooses to inform, and it carries no social stigma.

In contrast, obesity is usually a condition the patient has self-diagnosed, it is obvious to the public and it does carry a social stigma in many cases. It is unusual for the obese person to seek medical aid in the first instance. The media is full of advice on what to do, so typically the obese person who eventually has a medical consultation gives a history of a series of self-help measures, which did not have a satisfactory outcome. This gives doctors the impression that the self-help measures are invariably ineffective, but this is not so: medical referrals are always self-help failures, but self-help successes do not register in the medical statistics. Field studies of unselected populations show that those who are now overweight were often previously normal weight, and those now of normal weight were often previously overweight.(63)

So, from the viewpoint of the primary care physician seeing an obese patient in surgery, the patient has typically 'tried everything' without satisfaction, and hopes that the GP will have a pill which will solve the problem. The patient already carries a large amount of misinformation about obesity: especially about the rate of weight loss which is to be expected, since this is routinely exaggerated by some commercial weight-loss organisations and magazines. The patient is a self-help failure who may well have been told by the previous therapist that his or her failure indicates a rare metabolic abnormality with which he or she expects the GP to deal. This is not a basis on which a successful consultation can be expected.

This unsatisfactory situation is made much worse by the fact that the GP has had little or no training on how to manage obesity unless, in exceptional circumstances, they attended a postgraduate course,(64) and has no clinical guidelines on which to anchor the consultation and help with decision making.

The next part of this section outlines, for doctors (GPs and consultants), practice nurses (PNs) and dietitians, information about treatments offered and delivery of this service. As there are no significant services that specifically tackle the prevention of obesity or childhood obesity in the UK, the information below is restricted to the treatment of adult obesity. The final part of this section deals with the cost of obesity.

Top of Page

What do doctors do?

The first NHS port of call for the obese patient is the GP who can manage the patient him or herself, or refer them to a PN or a dietitian. No good information is currently available on how GPs manage obesity. However, the National Audit Office, London, carried out a survey during 1999 which will provide this information (personal communication, National Audit Office).

• Only one GP (5%) had treatment guidelines for obesity
• There was great variation in the number of patients treated directly by the GP, and in the number referred for treatment: in both cases from no patients to over 200 per year
• One GP never treated or advised patients in person, while one other GP never referred patients for treatment
• 22% of GPs had referred patients to a surgical consultant, but only 11% to a specialist physician.

Referral to a hospital consultant who runs a specialist obesity clinic is unusual, simply because there as so few. There are nine specialist clinics in the UK; two in London, one in Luton, one in Nottingham, one in Leeds, one in Liverpool, one in Manchester and two in Scotland (Aberdeen and Dundee) (personal communication, Dr Julian Barth, Leeds). Kopelman has argued that there should be more obesity clinics in the NHS.(65)

Specialist obesity clinics are usually multidisciplinary, and often have the facility to refer patients on to a clinical psychologist or surgeon. An audit of one such specialist clinic (at the Royal London Hospital, London) showed that 69% of patients were treated by diet alone, 23% by diet and anti-obesity drugs, and 7% by psychological interventions.(66)

What do practice nurses do?

As for GPs, little good information is currently available on how PNs manage obesity; the National Audit Office survey will provide many of the answers.

Ogden and Hoppe (67) provide some information on what PNs do. Thirty percent of the 586 PNs who responded to a questionnaire survey reported running weight loss clinics. The majority of the PNs who responded gave advice to more than one patient a week, but most spent 10 minutes or less discussing weight loss. Further, all PNs reported offering general nutritional advice and exercise most frequently, whereas advising patients to eat less in general and referrals to self-help groups were offered only sometimes, and calorie-controlled diets were offered the least frequently. From the work by Ogden and Hoppe (67) it appears that PNs offer general nutritional advice and exercise.

Top of Page

What do dietitians do?

From a survey conducted in 1998, only one out of 99 NHS dietetic services questioned did not accept overweight or obese client referrals for management.(68) The majority (87%) of respondents allocated up to 30 minutes for a new client appointment. Although 27% had no specific policy about the number of follow-up appointments during a 'course' of treatment, 20% offered one or two follow-up appointments; 35% offered three or four; 16% offered five or more, and only 2% offered no follow-up. Most follow-up appointments were every four weeks (8% were three weeks or more frequently; 50% were every four weeks; 35% were five weeks or less frequently, and data for 7% were missing). The duration of follow-up appointment was as follows: 24% were allocated less than 11 minutes; 73% 11-20 minutes; 2% 21-30 minutes).

The majority of dietetic services did not prioritise (or match) patients for different dietary treatments. Of those who did, criteria such as existing disease, other risk factors, diagnosed eating disorder, degree of fatness, weight history and mental health or learning difficulties were used to identify different approaches for management. However, no common strategies for matching patients to treatments could be identified. Most dietetic services (78%) saw overweight or obese clients only on an individual basis; 20% offered treatment on a combination of individual and group styles. Very few (only 23%) had a protocol, although some others (20%) stated that they currently had protocols under development.

The economic cost of obesity

Only two papers containing data on the healthcare cost of obesity in the UK have been published to date. The first appears as a brief section in a report conducted by the Office for Health Economics (69); the estimates were calculated for the early 1990s. West (69) estimated the direct healthcare cost of overweight and obesity to the NHS as £29.35 million per year; £13.8 million from general practice, £8.3 million from inpatient and £0.85 million from outpatient hospital services, £2.9 million from pharmaceutical services and £3.5 million from dietetic services. West (69) also estimated the total cost of overweight and obesity as a risk factor for some other diseases as £165.25 million per year; 5% of the total cost for stroke, 80% for diabetes, 10% for arthritis and 20% for hypertension.

The second paper estimated the additional direct costs of major co-morbidities associated with obesity to the health service, and also the potential for resource savings from effective interventions, also for the early 1990s.(70) Since data on the consumption of health service resources and personal details for the same individuals in England are not available, the authors obtained prevalence and resources data from separate sources. Prevalence data were obtained from a large population survey conducted in northwest England in 1992, providing information on height, weight, gender and recent health problems for which respondents had consulted a doctor in the previous 12 months. The principal source of resource use data was the 1995-1996 Hospital Episode Statistics data set for the northwestern region of the NHS, and the national Healthcare Resource Group cost figures were used to estimate costs.

Top of Page

The authors concluded that the number of grossly obese patients in the UK remains small and the burden they place on the health service is not very serious. Most of the additional cost is for those moderately overweight. Initial estimates suggest annual reductions in healthcare spending in England of up to £131 million per annum (1996 prices) with a BMI threshold of 25, or £15 million with a threshold of 30, may be possible with effective interventions for being overweight and obese. This is equivalent to 1% or 0.1% of total healthcare spending respectively. These data are important since they show that the healthcare needs should prioritise providing effective interventions for the overweight rather than the grossly obese.

Estimates for UK dietetics costs in 1998 are presented at the end of this section. Data on the cost of drugs used to treat obesity in the UK are available but because this is such a rapidly evolving area of medicine (section 6), the data currently available are inappropriate.

A survey is currently underway which will provide more information on the total cost of obesity in the UK; this survey forms part of the agenda of the WHO International Taskforce on Obesity.

The cost of managing obesity in other countries has been estimated, and some of these data are mentioned in brief below to provide some estimate of what the cost might be for the UK. The healthcare costs of excess weight have been estimated for a number of countries other than the UK.(71-73) These costs are derived by estimating the economic costs of different diseases, ascribing a proportion of this cost to obesity and summing the obesity-related portions. The calculations used by Levy et al.(72) are shown in Table 6. The estimated costs therefore depend on the total cost of the health services and assumptions about the proportions of disease costs attributable to obesity. Quesenberry et al.(74) recently found that there was an association between BMI and annual rates of inpatient days, number and costs of outpatients visits, costs of outpatient pharmacy and laboratory services, and total costs. Relative to BMI of 20-24.9, mean annual total costs were 25% greater among those with BMI of 30-34.9 (rate ratio, 1.25; 95% confidence interval, 1.10-1.41), and 44% greater among those with BMI of 35 or greater (rate ratio, 1.44; 95% confidence interval, 1.22-1.71). The association between BMI and CHD, hypertension and diabetes largely explained these elevated costs. In summary, estimated costs of obesity vary in different countries, at different times, and depend on the cut-off used to define obesity, but all available estimates conclude that somewhere between 1 and 5% of total healthcare costs are attributable to overweight and obesity.

Top of Page

Authors of the various economic analyses comment that their estimates are conservative, as there are several items for which no costing is available. Furthermore, most of these analyses only estimated direct costs. Where estimates of indirect cost (e.g. related to the loss of productivity by absenteeism, disability pensions and premature death) and personal costs (since obese people generally earn less than their lean counterparts) are available, they add significantly to the direct costs. It is also obvious that, with increasing prevalence of obesity in all affluent countries, the total burden on the healthcare budget is always increasing. We can, therefore, safely conclude that there are massive economic savings to be made in healthcare costs, apart from any other benefits to wellbeing, if the prevalence of obesity is reduced.

Top of Page

UK estimates for dietetic costs

Personnel

In terms of the costs specifically incurred by dietetic time, data on contact time from two large districts show that about one third of total contact time cost may be on weight management. In Leeds, 26% of referrals in 1996/7 and 1997/8 were for simple obesity; 12% were for diabetes and 3.5% were for lipids, and about half of the advice given in these consultations was on weight management (personal communication, Head of Dietetic Services for Leeds). In Durham, 15% of all contacts in 1997/8 were for simple obesity; 30% were for diabetes and 10% were for dyslipidaemia, of which 50% were for weight management (personal communication, Head of Dietetic Services for Durham). These data show that about 35% of all dietetic contact time is spent on weight management.

To estimate how much this activity costs the NHS per year, we used data supplied by the British Dietetic Association for the number of dietitians employed in each grade in the UK at 17 December 1998. Salary was calculated from the mid-point for each grade (not including discretionary points) at 1 December 1998 (Table 7). The total number of dietitians (full-time equivalents) managing obesity in the UK was estimated as 998, costing the NHS £16 776 834 per year. This equates to one dietitian per 59, 127 people in the UK general population (based on the 1997 estimate of 59, 009, 000; personal communication, Office of National Statistics, London) or one dietitian per 10, 052 obese people in the general population (based on 17% of the UK population being obese in 1996).

NHS dietitians are all linked to a dietetic department, and dietetic departments can be found attached to all district general hospitals (DGHs) or teaching hospitals. Dietetic departments attached to DGHs are usually smaller and deal with fewer tertially referrals, compared with dietetic departments attached to teaching hospitals. Excluding tertiary referrals, there is no regional difference in the number of dietitians per population served.

Top of Page

Treatments

The costs involved in the lifestyle treatment options (diet, exercise and behavioural therapy) are essentially the costs of the personnel involved. There are no good data on which to base reliable estimates of the cost of drug treatments or surgical treatments.

6 Effectiveness of services and interventions

Assessment of effectiveness

To assess the effectiveness of interventions for the treatment of obesity we have relied primarily on systematic reviews of RCTs. Randomised controlled trials are (almost) the best source of evidence for assessing the efficacy healthcare interventions, including lifestyle interventions.(75) Results from systematic reviews (the meta-analysis) of several RCTs which address the same question are even better.(76) A conclusion based on several RCTs, selected by previously defined criteria, is likely to be nearer the truth than one based on a group of RCTs which happen to support the reviewer's own prejudices.

There have been three such reviews on the treatment of obesity published in the past few years, and those which have assessed the efficacy (in terms of long-term weight loss) at one year or more are reviewed here. The three reviews were carried out by the National Institutes of Health in the USA,(15) the Canadian Task Force on Preventative Health Care (77) and the Centre for Reviews and Dissemination in the UK.(78) For simplicity, these three reviews are referred to below as the American, Canadian and UK reviews, respectively. The findings from these three reviews vary a little because each asked a slightly different question and thus included different studies; these differences are highlighted in Table 8. The following section is based on the largest review, the American review, which included studies from 12 weeks duration upwards, but which assessed studies by their duration. The studies identified by the American review which are referred to in this section can be found in their full report which is available to view at their website.

Top of Page

Additional studies identified by the Canadian and UK reviews, which are:

• a direct comparison of at least two interventions
• a comparison of an intervention with a placebo or usual care
• an addition of an intervention to another intervention versus the other intervention (e.g. the addition of physical activity to diet versus diet)

are cited in this section.

Where evidence on efficacy of some interventions from long-term RCTs has not been identified by these three systematic reviews, the next best evidence known to the authors is presented.

The outcome presented is weight loss. This is not, of course, the only measure of efficacy in a treatment for obesity, but it is one measure of outcome that is always reported, and this enables us to make comparisons between treatments.

More attention in this section is given to dietary, physical activity and drug interventions, compared with surgical interventions, since these are the most commonly employed in practice. Behavioural strategies are not primary treatments of obesity; they do not directly affect energy balance, but they help the patient comply with those aspects of treatment that do affect energy balance. The evidence of efficacy of many different types of behavioural strategies is reviewed by the American, Canadian and UK systematic reviews.

Top of Page

Estimates of the effectiveness, where known, are graded by the size of the effect where

and quality of the supporting evidence where:

Effectiveness of patient-centered interventions

Dietary treatments for obesity

Low-calorie diets (LCDs)

The American review examined 34 RCT articles for the impact of an LCD consisting of approximately 1000-1200 kcals/day on weight loss. Many of these LCDs also promoted low fat intake as a practical way to reduce energy intake. Twenty five of these RCTs covered interventions lasting six months or more. All of the studies, regardless of the length of the intervention, showed that LCDs result in weight loss. From the 25 RCTs with a duration of six months or more, compared with controls, LCDs brought about a mean weight loss of approximately 8% of body weight over a period of six months and up to one year. Four studies that included a long-term follow-up lasting 3 to 4.5 years reported an average weight loss of 4% over the long term.

In addition, the Canadian review identified a trial by Skender et al.,(7) who found no difference in weight loss at two years between a group who exercised and another group who exercised and took a LCD.

Top of Page

Quality of supporting evidence = I-1 Size of effect = (six months to one year) A

Low-fat diets

The American review identified nine RCTs testing the efficacy of diets that varied in fat and energy content on weight loss. (Note: two of these RCTs were also included in their assessment of LCDs,(80),(81) and one of these RCTs was also included in their list of excluded studies.(82)). Lower-fat diets varied from 20 to 30% of energy from fat, and energy ranged from 1200 to 2300 kcals/day. Three RCTs, all six months or greater in duration, promoted lower-fat diets with ad libitum energy intake. Two of the three RCTs reported that lower-fat diets with ad libitum energy resulted in a reduction in energy intake of 85 kcal and 300 kcals, and all three RCT's produced a greater weight loss by a mean of 1 to 3.9 kg compared with a higher-fat diet. Three RCTs compared lower-fat diets with targeted energy reduction to lower-fat diets alone, and all three found that weight loss was greater in the lower-fat diet with energy reduction than in the lower-fat diet alone. When similar energy levels occur between lower- and higher-fat diets, similar amounts of weight loss were reported in two studies, whereas one study showed 1.8 kg greater weight loss on the lower-fat compared with the higher-fat diet, despite similar reported energy levels.

The UK review identified an additional study by Pascale et al.,(83) who found that people with NIDDM receiving calorie- and fat-restricted diets lost more weight than those only restricting calories.

Taken together, these studies show that lower-fat diets ranging from 20 to 30% of energy can contribute to lower energy intake even when energy reduction is not the focus of the intervention, but when LCDs are targeted with lower-fat diets, better weight loss is achieved. However, there is little evidence that lower-fat diets per se cause weight loss independent of energy reduction.

Top of Page

Low-fat, low-energy diets:
Quality of supporting evidence = I-1 Size of effect = A

Low-fat, ad libitum diets:
Quality of supporting evidence = I-2 Size of effect = C

High-fibre diets

The UK review identified only two studies which directly compared diet alone, both examining the effects of dietary fibre.(84),(85) Fibre supplements were more effective than placebo at increasing weight loss when given in conjunction with a 1200-1600 kcal/day diet.(85) However, this does not appear to lead to greater mean weight loss than a low-fibre/low-energy diet.(84)

Fibre supplements:
Quality of supporting evidence = I-2 Size of effect = (at six months) C

Very-low-calorie diets (VLCDs)

The American review identified four RCTs comparing VLCDs (providing about 400-500 kcal/day) with LCDs (providing about 1000-1200 kcal/day). During the active phase, VLCDs were given exclusively for 12-16 weeks and then were followed by LCDs for a total duration ranging from 24 weeks to five years. The participants in these trials were primarily women who were extremely obese. VLCDs either alone or, usually, combined with behavioural therapy, promoted weight loss of approximately 13-23 kg during the active phase of the VLCD intervention, compared with 9-13 kg with LCDs. In three of the four studies, VLCDs resulted in 4-12 kg greater weight loss than the LCDs at the end of the active phase. Over the medium term of 6-12 months, weight loss on the VLCDs ranged from 1.1 to as much as 10.4 kg greater than weight loss on LCDs. (One of the studies which was excluded by the American review (86) but included in the UK review provided similar findings at one-year follow-up.) One study did not show a particular advantage of VLCD over LCD either during the active phase or at 24 weeks. After one year, there was no long-term advantage of VLCDs over LCDs.

The Canadian review included three additional trials in their assessment of VLCDs. At two-year follow-up they found that a VLCD plus behavioural therapy produced greater weight loss compared with behavioural therapy alone; Miura et al.(87) found a mean difference of 1.3 kg, Torgerson et al.(88) found a mean difference of 2.9 kg. (Note: the American review identified the trial by Miura et al. but excluded it from their assessment.) Unlike the other studies, Miura et al.(87) found that a VLCD alone produced a greater weight loss than either behavioural therapy alone or behavioural therapy plus a VLCD. In a comparison of continuous VLCD versus VLCD followed by LCD, Ryttig et al.(89) found no difference in weight loss at two years.

Top of Page

VLCDs versus LCDs:
Quality of supporting evidence = I-1 Size of effect = A

Meal replacements

An interesting paper on meal replacements was identified by the UK review; Anderson et al.(90) compared a liquid supplement diet providing 800 kcal/day with a diet consisting of a combination of liquid supplements, providing 320 kcal/day, and conventional foods to the value of approximately 500 kcal/day. Both groups also received individual meal plans that gradually introduced conventional foods over a period of 4-6 weeks. Both groups had lost significant amounts of weight (about 15% of their initial weight) by the end of the 12-week weight-loss phase. At one-year follow-up, the two groups combined sustained weight losses over time (p < 0.001), but between-group differences were not analysed. Another RCT (91) found that weight loss in patients on a LCD was less at three months (mean 1.3 kg) compared with that in patients using meal replacements (mean 7.1 kg); this difference in weight loss was maintained at one year (11.3 and 5.9 kg, respectively) while all patients used a combination of LCD and meal replacements.

Meal replacements versus LCD:
Quality of supporting evidence = I-2 Size of effect = A

Milk diet

The efficacy of the milk diet has recently been subject to a RCT which assessed the short-term (over 16 weeks) efficacy of the milk diet versus a conventional weight-reducing diet.(92) Patients on the milk diet lost significantly more weight than those on the conventional weight-reducing diet, although part of the reason for this success may be that the milk diet was novel to the patients, and that this novelty may have worn off with time.

Milk diet versus LCD
Quality of supporting evidence = I-2 Size of effect = (at 16 weeks) A

Food provision

The UK review also looked for studies which assessed the usefulness of food provision. Standard behavioural therapy combined with the provision of meal plans and grocery lists to obese women produced significantly greater weight loss than standard behavioural therapy alone.(93) A smaller but significant result was found when all required food was provided to participants.(94),(95)

Top of Page

Quality of supporting evidence = I-2 Size of effect = C

Note: Three studies by Wing et al.,(96-98) which are included in the UK review, are not included in this section on dietary treatments since one assessed the efficacy of a nutritional educational package.(96) and the others were confounded by financial incentives.(97),(98) Note also that the 1991 study by Wing et al.(97) was identified but excluded by the American review because of a lack of control group, and the 1994 study by Wing et al.(98) was included by the American review in their comparison of behavioural interventions.

 Exercise

Exercise versus controls

Twelve RCTs were included in this section of the American review. Ten of these articles reported a mean weight loss of 2.4 kg in the exercise group compared with the control group; two reported no difference.

Quality of supporting evidence = I-1 Size of effect = B

A meta-analysis of 28 publications on the effect on weight loss of exercise compared with diet or control groups showed that aerobic exercise alone produces a modest weight loss of 3 kg in men and 1.4 kg in women compared with controls.(99) There was no significance in weight loss related to the type of physical activity. Andersen et al.(100) compared, in obese women on a low-fat diet, the effect of adding lifestyle activity or structured aerobic exercise: at one year there was little difference between the groups who lost 6.7 kg and 7.1 kg respectively.

Exercise versus diet

Ten RCTs (nine of which were the same studies as those above) reported on RCTs that had a diet-only group and an exercise-only group. In every case except one, the exercise-only group did not experience as much weight loss as the diet-only group, with a difference of approximately 3 kg.

The UK review also identified and included a study by King et al.,(101) which assessed the effect of diet or exercise over one year in men (the study continued with the aim of assessing different types of maintenance strategies); the diet group lost significantly more weight than the exercise group at one year.

Quality of supporting evidence = I-1 Size of effect = E (exercise produces less weight loss compared with diet)

Top of Page

Exercise plus diet versus diet or exercise

The American review included 15 studies. Each of them had a diet-only arm and six also had an exercise-only arm. Twelve of these 15 studies observed that the exercise-plus-diet group had a mean greater weight loss of 1.9 kg compared with the diet-only group. Three RCTs compared the longer-term versus the shorter-term effects of exercise plus diet versus diet alone. All three studies found that exercise plus diet resulted in approximately 1.5-3 kg greater weight loss than diet alone over nine months to two years.

In addition, the UK review identified a trial by Skender et al.,(79) who found no difference in weight loss at two years between a group who took a LCD and another group who exercised and took a LCD.

Exercise plus diet versus diet:
Quality of supporting evidence = I-1 Size of effect = C/B

The UK review included two additional studies which assessed the efficacy of behavioural interventions with an exercise component. One study found that supervised aerobic exercise sessions resulted in more sustained weight loss than the provision of exercise education.(102) The other study (103) compared the use of stimulus control versus stimulus control and exercise. At one year, the stimulus control group had lost 2.4 kg compared with 7.4 kg lost by the stimulus control and exercise group.

Pharmacotherapy

Data reporting the results of 52 weeks of drug treatment for obesity are available only for dexfenfluramine, orlistat, phentermine, d-1-fenfluramine and sibutramine. The American review included all of this data. However, this section considers only the evidence relating to sibutramine and orlistat, since they are the only two drugs licensed in the UK for the treatment of obesity.

Top of Page

Sibutramine

Sibutramine promotes a sense of satiety through its action as a serotonin and noradrenaline re-uptake inhibitor. In addition, it may have an enhancing effect on thermogenesis through stimulation of peripheral noradrenergic receptors.

The American review included two studies.(104),(105) Lean (106) reviewed the clinical efficacy of sibutramine, and included the study by Jones et al.(104) cited above, and in addition studies by Jones et al.,(107) Drouin et al.(108) and Bray et al.(109) Lean found that sibutramine produces dose-related weight loss when given in the range 5-30 mg/day, with an optimal dose of 10-15 mg/day. Weight loss with sibutramine is 3-5 kg better than placebo at 24 weeks, and weight loss is maintained to 52 weeks at doses of 10 and 15 mg. By six months, 69% of patients treated with sibutramine (15 mg) achieve a 5% or greater reduction in their baseline weight.

Quality of supporting evidence = I-1 Size of effect = B

Orlistat

Orlistat, which was licensed for prescription in the UK and the rest of the European Community in August 1998, is a powerful inhibitor of pancreatic lipase, so some 30% of dietary fat is not digested but is excreted in faeces. The media have suggested that this new drug will enable fat people to eat what they like and still lose weight. As Garrow points out in a BMJ editorial, this is highly misleading.(110) Anyone taking orlistat who eats a high-fat diet will receive a powerful incentive to reduce fat intake.

The American review identified two studies.(111),(112) The first study found an additional mean weight loss of 2.2 kg with 150 mg/day in the sibutramine group compared with the placebo group after 12 weeks of treatment. The second study found an additional mean weight loss of 0.63 kg with 30 mg/day, 0.71 kg with 180 mg/day and 1.75 kg with 360 mg/day in the sibutramine group compared with the placebo group after 12 weeks of treatment.

Since the American review was completed there have been three more publications reporting RCTs on orlistat. The first (113) was a trial lasting one year on 46 obese subjects (BMI 38), half of whom received 120 mg orlistat three times daily after a four-week run-in period, while the control group received placebo capsules. All subjects were prescribed a low-fat diet, which was designed to cause an energy deficit of 600 kcal/day initially, and after week 24 it was reduced by a further 300 kcal/day. Of the patients on orlistat, nine out of 23 did not complete the trial, and in the control group, 11 out of 23 dropped out. In the intention-to-treat analysis the last recorded weight was carried forward, so drop-outs were assumed to have maintained their last weight, rather than regaining weight. The weight loss (mean (s.d.)) at six months was 8.6 (5.4) kg in the treated group and 5.5 (4.5) kg in the placebo group.

The second report (114) was about 733 patients who were recruited for a similar protocol to the trial described above, but who were followed during a second year. After screening, 340 were assigned to placebo and 343 to 120 mg orlistat three times daily. Mean weight loss after one year was 10.3 kg for the orlistat group and 6.1 kg for the placebo group.

Top of Page

The third report, by Davidson et al.,(115) was of a multicentre trial in the USA in which 1187 subjects entered, of whom 223 were given placebo and 657 orlistat. Among those on orlistat weight loss (mean (s.e.m.)) was 8.76 (0.37) kg, compared with the placebo group, whose loss was 5.81 (0.67) kg, which was significantly different (p < 0.001). It should be noted that if the s.e.m. of the weight loss among 657 patients on orlistat was 0.37 kg, the s.d. was 9.48 kg, which is larger than the mean weight loss of 8.76 kg. These three trials show that patients on orlistat lose more weight than those on the same low-fat diet with placebo. However, the mean difference between orlistat and placebo is about 3 kg after one year of treatment, and about one in six of the patients on orlistat for one year were heavier than they had been at baseline.

Quality of supporting evidence = I-1 Size of effect = B

The Royal College of Physicians of London published a report in December 1998 on the clinical management of overweight and obese patients with particular reference to the use of drugs.(51) This report provides no information on the efficacy of the various drug treatments for obesity (it gives reference to the UK review), but it does provide a useful clinical guideline. Other useful information on pharmacotherapy for obesity is provided by Kopelman (53) and Williamson.(116) Both of these summaries stress that this is a rapidly evolving area of healthcare.

Surgery

Eight RCTs were included. All of these studies included individuals who had a BMI of > 40, or 35-40 with co-morbidity. Weight loss due to surgical intervention, such as gastric bypass, ranged from 50 kg to 100 kg over a period of six months to one year. Further information on the efficacy of different types of surgery can be found in the American review (pp54-55).

The Canadian review identified an additional paper which presented longer-term outcomes for a study identified by the American review.(117) Both the UK and Canadian reviews identified an additional paper which presented longer-term outcomes for a study published in 1984 and identified by the American review.(118) In addition, the UK review identified eight RCTs, two of which were excluded by the American review. Some of these studies used jejunoileostomy, or the use of the gastric balloon compared with sham procedures, which are no longer recommended. The remaining three papers examined gastric bypass versus gastroplasty. Each showed that gastric bypass produced greater weight losses.(119-121)

Top of Page

Quality of supporting evidence = I-1 Size of effect = A*

The National Institutes of Health Consensus Development Conference consensus statement, 'Gastrointestinal Surgery for Severe Obesity',(122) concluded that the benefits outweigh the risks and that this more aggressive approach is reasonable in individuals who strongly desire substantial weight loss and have life-threatening co-morbid conditions.

Jaw wiring and waist cords

Wiring the jaws together is a standard procedure in the treatment of fractures of the lower jaw. Garrow (123) suggested this procedure as an alternative to the then-prevalent jejunoileal bypass for severely obese patients who were unable voluntarily to restrict their food intake. Rodgers et al.(124) reported a series of 17 patients, initially twice their ideal weight, who lost a median of 25.3 kg in six months without serious complications. The lost weight was, however, rapidly regained at an unusual rate when the wires were removed. Subsequently, Garrow and Gardiner (125) suggested fitting a waist cord when the wires were removed. Garrow and Webster (126) reported a series of 35 patients who were offered jaw wiring combined with a waist cord. Of these, nine dropped out during the jaw-wiring phase and another 12 during the waist-cord follow-up (mainly because they moved away or became pregnant). The remaining 14 (who were initially 54.1 kg overweight) on average lost 42.4 kg during 11 months of jaw wiring. There was a regain of 9.6 kg during the waist-cord phase, which resulted in an average weight loss of 32.8 kg over three years. This procedure has not been tested in a RCT, but these results compare well with other surgical procedures, with a much lower cost or morbidity.

Quality of supporting evidence = II-2 Size of effect = A*

Other types of interventions

Examples of such treatments include acupuncture, herbal remedies, supplements and hypnotherapy. No studies of acceptable quality were identified by any of the three reviews.

Maintenance of weight loss in adults

The evidence relating to maintenance and combined treatment and maintenance programmes is reviewed separately by the UK review. There were many different types and combinations of interventions used in the 21 studies that were identified. Overall, greater frequency of (or any) contact over the long term appears to result in smaller weight gains/greater weight losses. A commentary on the usefulness (or rather uselessness) of maintenance programmes after weight loss was published recently.(127) The value of waist cords for the maintenance of weight loss is discussed above.

Top of Page

Quality of supporting evidence = I-1 Size of effect = too variable to summarise (due to the many different types and combinations)

Prevention of obesity in adults

The evidence relating to the prevention of obesity in adults is reviewed by the UK and Canadian reviews, and also by Hardeman et al.,(128) who included studies regardless of study design or length of follow-up.

The UK review included three RCTs, all community-based studies with no-intervention control groups. In the Stanford Five Cities Project, a multimedia education programme resulted in reduced average weight gain. (129) The provision of a monthly newsletter including information relevant to weight control combined with a financial incentive to maintain weight, resulted in a large reduction in the proportion of people gaining weight at 12 months.(130) In the third study, a community-based programme using social learning and communication skills failed to demonstrate any benefit.(131) The Canadian review identified no RCTs. Hardeman et al.(128) identified three studies, but could not come to any overall conclusions about their efficacy since the papers provided little detail on the interventions, underlying theoretical model or study methodology. Furthermore, outcome measures were usually self-reported, drop-out rates were high and follow-up was generally short.

Quality of supporting evidence = I-1 Size of effect = too variable to summarise (due to the different types and combinations of interventions used)

Obesity in children

Treatment of obesity in children

The evidence relating to the treatment of obesity in children is reviewed by the UK review. Eleven small trials were found, six from a single research group. Two trials, both of good quality, suggest that an effective strategy is to use interventions designed to reduce sedentary behaviour. No long-term benefit for weight loss in children was demonstrated by the addition of controlled exercise to diet (one trial). Other trials (two) were too small to be interpretable.

There is conflicting evidence over the effectiveness of treating obese parents and children together, and the involvement of parents in children's treatment programmes. The treatment of obese parents and children together may be of more value for the treatment of younger children, i.e. 5-8 year olds. In studies where one group had parental involvement and the other did not, one study found no difference and another found that the control group gained weight at a slower rate. In another study, parental involvement seemed most effective when the child and parent attended support sessions separately.

Top of Page

An excellent paper, prepared by an expert committee in the USA, provides practical guidance to paediatric clinicians who treat obese children.(132)

Quality of supporting evidence = I-1 Size of effect = too variable to summarise (due to the many different types and combinations of interventions used)

Prevention of obesity in children

The evidence relating to the prevention of obesity in children is reviewed by the UK review and also by Hardeman et al.(128) The UK review included only one RCT; family therapy was shown to be more effective in preventing the progression of obesity in 10-11-year-olds than conventional dietary and exercise treatments or no intervention.(133) Hardeman et al. identified five studies, but could not come to any overall conclusions about their efficacy for the same reasons as those cited above.

Quality of supporting evidence = I-2 Size of effect = D

Effectiveness of healthcare professional-centered interventions

Harvey et al.(134) have carried out a systematic review of intervention studies to determine the existence and effectiveness of interventions to improve health professionals' management of obesity or the organisation of care for overweight and obese people. Objective measures of health professionals' practice and behaviours, and patient outcomes, including satisfaction, behaviour, psychological factors, disease status, risk factors and measures of body weight, fat or BMI, were collected. Twelve studies that met all the review inclusion criteria were identified. Three were RCTs of health professional-oriented interventions (such as the use of reminders and training) and one was a controlled before and after study to improve collaboration between a hospital clinic and GPs. A further eight RCTs were identified of interventions comparing either the deliverer of weight loss interventions or the setting of the delivery of the intervention.

The authors found that the heterogeneity and generally poor quality of identified studies made it difficult to provide recommendations for improving health professionals' obesity management. They concluded that, at present, decisions about improving the provision of services for overweight and obese people must be based on the evidence from patient interventions and good clinical judgement. However, Harvey et al.(134) identified a number of options which may warrant further exploration: reminders to health professionals, the use of intensive inpatient services and interventions to improve shared care across existing services.

Top of Page

A potentially important barrier to effective obesity management is a lack of motivation to work with this patient group due to negative perceptions of overweight and obese people or the efficacy of treatments. Although Harvey et al.(134) searched for studies that had assessed whether negative attitudes among providers were impinging on good practice and whether interventions to change attitudes might result in improved clinical decisions, none was included. One potentially relevant study was identified, but this fell outside the inclusion criteria because the intervention was delivered to medical students and focused on attitude change without an objective measure of behaviour change. Given that much commentary has been passed on the possible implications of negative views towards this group of patients, it is surprising there have been no rigorous evaluations of strategies to improve negative attitudes and related practices.

Effectiveness and cost

Clearly most expenditure on obesity in England is spent on advice on diet. Occasionally advice on exercise is also given, although this is usually no more than asking the patient to increase their levels of physical activity. Most of this advice is given as part of one-to-one consultations by GPs, PNs or dietitians. The findings of the systematic reviews would suggest that this is essentially a waste of NHS resources, and that the NHS would be better to spend the same amount of money on treating fewer patients with more effective treatments, such as some drugs or surgery. However, we suggest (in section 8) how the NHS could most effectively spend the existing resources allocated for obesity, while explaining why resources should be diverted towards this activity

7 Quantified models of care/recommendations

Objectives of treatment

In the treatment of patients with established obesity there are three objectives:

• to assist the patient to achieve a weight at which the health risks of obesity are reduced to the lowest possible level for this patient
• to help the patient to maintain this weight loss indefinitely
• to maintain, or restore if necessary, the patient's self esteem.

  Top of Page

Target weight and rate of weight loss

Some guidelines emphasise that 'priority in obesity management should be on weight maintenance and modest weight loss, rather than a return to ideal or normal weight'.(52) The reasoning behind this advice is that there can be a substantial reduction in mortality and risk factors in overweight people who lose 5-10 kg in one year, and that repeated failures to achieve normal weight amplify a patient's depression and lack of self-esteem. These are valid arguments, but there is no evidence that a patient who is, say, 40 kg overweight does better if advised to lose 5-10 kg than if advised to lose 30-40 kg. If the patient who is 40 kg overweight is aged 65 years, short in stature and crippled with osteoarthritis of hips or knees, it is quite likely that a loss of 5-10 kg is as much as can reasonably be achieved. In this situation, the modest weight loss may reduce the health risks to the lowest possible level for this patient, and thus achieve the first objective above.

On the other hand, if the 40 kg-overweight patient is aged 25 years and otherwise fit, but has a family history of diabetes or heart disease, a loss of only 5-10 kg is unlikely to be the optimum for reduction of health risk. Thus the appropriate target weight loss should be assessed in the light of the individual circumstances, rather than by the application of any fixed rule.

Concerning the optimum rate of weight loss there is consensus that a rate greater than 1 kg/week is undesirable, since it is likely to cause excessive loss of lean tissue.

Priority for treatment: obesity or co-morbidity?

In section 3, three subcategories of obesity were identified: obese patients with existing co-morbidities, obese patients without co-morbidities and non-obese individuals with special risk factors for the development of obesity. At present the usual policy is to treat the co-morbidities of the first group, and to ignore the remainder.

Top of Page

This policy may not be appropriate. French et al.(135) studied the effect of weight change between the ages of 18 and 50 years on disease prevalence later in life. They obtained by postal questionnaire reported weight change between the ages of 18 and 50 years in 41, 837 women, and were able to classify the weight change pattern of 17, 252 of these women into one of five classes. The pattern of the remaining 23, 710 women (mostly weight gainers) would not fit into these patterns, and were excluded from the series. Between age 18 and 30 years those in whom weight changed by less than 5% were classified 'weight stable'. Those in whom weight increased, or decreased, by 10% or more were classified as 'weight gainers' or 'weight losers'. A similar designation was used for weight change between the age of 30 and 50 years. The number of women who showed each of the five patterns of weight gain, and their subsequent odds ratio of developing diabetes, high blood pressure, heart attack, other heart disease, cancer or perceived poor health, is shown in Table 9. The data are also analysed separately for those who were, or were not, overweight at age 18 years.

*Stable = < 5 kg change, loss = 5 kg+ loss, gain = 5 kg+ gain.
(a) not overweight at age 18 years, (b) overweight at age 18 years.

Three important messages emerge from this table.

Top of Page

1. Weight loss of 5 kg or more between the age of 18 and 30 years is uncommon: it was observed in only 8% of the population, half of whom regained this lost weight during the period between 30 and 50 years of age. The largest group (47%) are those who gain at least 5 kg between 18 and 30, and again between 30 and 50 years.
2. The odds ratio for diabetes, hypertension, heart attack, other heart disease and perceived poor health (but not cancer) is significantly increased in those women who gained weight compared with those whose weight was stable. The diseases most strongly associated with weight increase are diabetes and hypertension.
3. Eight percent (712 of 9153) of the women were overweight at age 18. In virtually every instance, the odds ratio of the diseases analysed (except cancer) is markedly higher among those who were overweight at 18 compared with those who were not.

This indicates that a policy of laissez faire is not justified. Individuals who are overweight at age 18 years, or who gain more than 5 kg during adult life, have a significantly increased risk of developing the co-morbidities of obesity. Both medically and economically it may be better to prevent this weight gain than to treat the co-morbidity when it arises.

Health gain from weight loss

The data presented above show that weight gain in adult life is a common event, and is associated with greater risks of ill health. In adults, there is conflicting evidence about the health effects of weight loss, or fat loss. This evidence is reviewed at length in a systematic review.(15) Most of the evidence comes from prospective cohort studies. In particular, there are strong epidemiological data, which show that weight loss is associated with increased total mortality. This has led some commentators to conclude that, although weight gain is bad for health, losing the excess weight does not confer any benefit.

This paradox has been resolved by a secondary analysis of data from the Tecumseh and the Framingham studies.(136) In these two studies, change in body weight, and also in fat mass (by skinfolds) was recorded. They found that each standard deviation in weight loss increased the mortality hazard by 29% and 39% respectively in the two studies. Contrarily, each standard deviation of fat loss reduced the hazard rate by 15% and 17% respectively. This supports the interpretation that unintentional weight loss in the general population is a sign of ill health, but that intentional fat loss in obese people is beneficial.

Top of Page

The reason why evidence relating intentional weight loss to health outcomes comes from cohort studies is obvious; although one can randomise individuals to different types of treatments for obesity, one cannot randomise them to weight-loss or no weight-loss groups. Indeed, this problem was the topic of a workshop convened in 1997 by the National Institutes of Health and Centres for Disease Control and Prevention. (137),(138) The workshop participants agreed that a well-designed RCT should be undertaken to estimate the magnitude and direction of the long-term health benefits of intentional weight loss. The results of this study, and those from the SOS study will provide much better evidence. Recently, data on sick leave and disability pension from the SOS study at four years were published.(139) Severely obese patients who underwent surgical treatment for obesity lost more weight and experienced a reduction in sick leave and disability pension compared with controls, particularly in patients aged 47-60 years.

Matching patients to resources

It is obvious that the number of people involved, and type of care needed, varies greatly between the subgroups identified in section 3. The relatively small number of obese people with established co-morbidities urgently need medical care, whereas the large number of non-obese people at high risk to develop obesity do not. This suggests that the appropriate model of care should have different levels ranging from low-cost advice centres, which are easily accessible to concerned members of the public and capable of coping with large numbers, to tertiary referral centres capable of providing medical care for the most complex problems.

• Level 1 is essentially a self-help group, which may be guided by a health professional - either a state registered dietitian (140) or a specially trained public health nurse.(141) It is aimed at management of overweight people for whom the SIGN (52) target of weight maintenance or modest weight loss is entirely suitable. It is also suitable for the long-term follow-up of patients who have been assessed and/or treated at Level 2 or 3. The GP of the individual attending a Level 1 centre should be so informed, to ensure that there is no medical contraindication to the proposed weight management plan.
• Level 2. For those individuals who do not achieve satisfactory results at Level 1, or whose obesity or co-morbidity is sufficiently severe to justify it, the next step is to Level 2. This is based on a primary care doctor who may, having assessed the patient, refer back to Level 1. The assessment of current disease and risk factors at Level 2, which is suggested by the SIGN (52) guidelines includes the following:
  - measurement of weight, height and waist: BMI calculated
  - risk factors assessed (e.g. smoking)
  - blood pressure measured
  - urine tested for glucose
  - plasma g-glutamyl transferase
  - total plasma cholesterol
  - thyroid-stimulating hormone.
  Advice on risk factor management and dietary advice may be offered at level 2, or the patient may be sent with a referring letter to level for this advice and further follow-up.
• Level 3. The last step, which is appropriate for a small minority of difficult cases, is Level 3, based on a specialist in a tertiary referral centre. Any attempt to quantify the workload at each of these levels involves assumptions about the efficacy of each level (especially Level 1), and the criteria which are used to make the judgement that a change to another level is indicated. Furthermore, if the system works well the prevalence of obesity will decrease, and hence the load at every level will decrease also (see below).

  Top of Page

Integration within the NHS and relationship with private facilities

The components of the above model, which involve NHS resources (Levels 2 and 3), should be integrated with NHS activities elsewhere. Clearly the existing facilities for managing diabetes, heart disease, osteoarthritis, etc., should be available to obese patients entering through the Level 1 route, and conversely patients who enter treatment for a co-morbidity, and who are also overweight, should have access to the Level 1 facilities.

Any model of obesity management based on the NHS will run in parallel with other non-NHS facilities, such as commercial clinics, slimming clubs, health farms and leisure centres. At presentm there is justifiable suspicion and even hostility between NHS and non-NHS healthcare providers in the field of obesity, since there are many charlatans who are untrained and unregulated in the private sector. For example, under rule 51.3 of the Advertising Standards Agency, slimming clubs are not allowed to advertise themselves as treating obesity (BMI > 30 kg/m2) because this is a medical condition which should be treated 'under the supervision of a suitably qualified health professional'. This is a commendable attempt to protect the public from exploitation by charlatans, but it also limits the ability of well-run private facilities to provide much-needed Level 1 facilities.

Logistics of the proposed model

Since a very large proportion of the population need help from this model it is logistically necessary to prioritise services to those who would derive most benefit from weight control. The only health benefit associated with obesity is a protection against osteoporosis: probably this is partly because adipose tissue contains aromatase, which converts androgens to oestrogens, and hence tends to preserve skeletal mass in post-menopausal women. The health hazards of obesity are greater in young people than in older ones, and weight loss is easier to achieve in younger people, so there are grounds for offering help primarily to younger adults.(142) Management of obesity in children requires a different approach. The implications of this policy related to demographic patterns are discussed below.

Age structure of a model population of 100 000, and of those requiring obesity management services

The age structure of the population of England and Wales is shown in Table 10, divided into five categories. For three of these (pre-school children, secondary school children and retired adults) the case for providing obesity management services is much weaker than for the remaining two - primary school children and pre-retirement adults. The reasons for these priorities are indicated above.

Top of Page

NK = not known

For purposes of the following calculation, the population is assumed to be of a constant size, and numbers have been rounded to give no more than three significant figures. In the model population of 100 000 with a typical age structure there will be 8900 primary school children aged 5-11 years, with an annual intake of 1270. Among each year's entry 9% will be above the 91st centile, giving an annual intake of 120 children, who are in this category for seven years, giving a total of 840 primary school children requiring management of their overweight. Facilities for children are discussed later in this section.

The other sector requiring help are the pre-retirement adults of which there are 63 600. With the current prevalence of overweight and obesity in the UK population there will (at any given time) be about 30 000 who are overweight (BMI > 25), of whom about 12 000 are obese (BMI > 30). To achieve the second objective of treatment given at the start of this section, help should be available to 30 000 young adult (aged 16-64 years).

Top of Page

Weight management clinics required to help 30 000 overweight adults

Although typically a population of 100 000 people will at any given moment have about 30 000 who need help with weight management, it will not be the same 30 000 from one year to the next. In Table 10, the annual intake of overweight adults is shown as 'not known'. Within a population individuals are constantly passing above and below whatever threshold is used to define an 'overweight' category (143) so it is not possible to identify a subsection of the population who are particularly at risk for obesity. The facility must therefore be some form of clinic capable of dealing with those individuals who are overweight, and who perceive themselves to be in need of help at this time.

For reasons given above, it is not possible to estimate how many people in a population of 100 000 would wish to use a Level 1 facility. Some indication can be gained from the experience of Bush et al.(140) in Harrow, and of Karvetti and Hakala (141) in Finland. The Harrow Slimming Club operated in a northern suburb of London, and registered 1090 members in 50 courses over a period of ten years. On average, therefore, each year there were five 10-week courses, which were attended by 109 people. Karvetti and Hakala recruited 243 overweight people in six weeks in the city of Turku, which has a population of about 160 000 and 'several' health centres. If a district with a population of 100 000 had five clinics, and each clinic ran ten courses per year for 20 people per course, this would provide facilities for 1000 members per year. This would probably be an adequate facility, because a family member (usually a mother of children) who attended the course would often be able to transmit what she had learned to other members of the family.

Level 2 help should be available by referral of Level 1 members who have additional problems. If this is estimated to be 10% of those attending at Level 1, then there would be 100 referrals per year to GPs of overweight people with additional problems. This would not be a great burden, since it would be fewer than ten patients per year to individual GPs.

If there were ten specialist centres in the UK, then each Level 3 facility would serve about 600 districts. If practitioners at Level 2 referred 98% of the patients back to Level 1 for management, and 2% on to Level 3, then each specialist unit would receive about 1200 referrals per year. Assessment at Level 3 would initially involve an outpatient consultation with a specialist (and preferably also a dietitian), for which about 30 minutes would be required, so a clinic session of 2.5 hours would provide slots for five new patients. At this input, the Level 3 facility would have to run about five clinics per week. At present there is no experience of a national Level 3 facility such as that suggested, so the estimate of workload is a guess which would have to be modified in the light of experience, and in the light of the confounding factors mentioned at the start of this section.

Facilities required to manage 30 000 overweight adults

The facilities needed at Level 1 are easily provided by local authority school health clinics, which are not used in the evenings. The most important requirement is a good leader - preferably a state registered dietitian or specially trained health nurse. A room in which it is possible to conduct seminar-type teaching with a group of up to 25 members, and facilities for weighing the members of the group and for showing audiovisual teaching aids is also required.

The facilities required at Level 2 are normally found in the surgery premises of GPs. These are for a clinical consultation, physical examination and assessment according to the SIGN (52) guidelines. These are weight, height and waist measurement, blood pressure, urine glucose, plasma g glutamyl transferase, plasma cholesterol and thyroid-stimulating hormone.

Top of Page

The facilities for Level 3 care will normally be in a teaching hospital. Apart from the advantage of the facilities mentioned below, it is important that the management of obesity at this level should be part of the education of healthcare students. In addition to the normal clinical and diagnostic facilities, there should be a physician with a special interest in obesity, full dietetic support, and preferably facilities for measuring resting metabolic rate by indirect calorimetry and for measuring body composition by either densitometry, total body water or total body potassium. The consultant physician needs to have the option of referral for advice to a surgeon or clinical psychologist, and to be familiar with the indications for drug treatment. It is very helpful to have the option to use up to three hospital beds, so that patients who have been referred from a distance can be admitted for investigation if necessary.

Hospital admission 'to get the patient started on a diet' is a waste of resources. Of course it is possible to initiate weight loss in a metabolic ward, but that has no long-term benefit because if the patient is totally unable to diet outside hospital the prognosis is hopeless. However a brief admission (up to three weeks) may be justified in some cases to establish the type of diet and exercise programme which a given patient can reasonably be expected to maintain in the long term. Even the most highly motivated patients do not adhere to dietary restrictions as rigidly at home as they do in a well-run metabolic ward.(144)

Management of obesity in children

The evidence for childhood obesity as a risk for ill health in adulthood is good (see section 2), but data on the effects of intentional weight control in childhood on health outcomes are lacking. Table 9 shows that the subsequent health problems of women who are already overweight at the age of 18 years are significantly greater than of those who are not overweight at age 18 years, so it is evidently necessary to prevent excessive weight gain in childhood.

The change in BMI from birth to age 20 years is shown in Figure 3. In the first year of life, the normal child trebles its birth weight (from about 3.4 kg to 10 kg). Over the next four years height increases more rapidly than weight, so BMI decreases, and the nadir of the 50th centile line is below 16 kg/m2 at age 5 years. During this period of rapid height growth there are potential dangers of causing permanent stunting if energy intake is restricted. In fact, fat babies often do not stay fat,(145) so it seems prudent to delay weight control measures until after the age of 5 years. By age 12, the 50th centile line has risen to 18 kg/m². The broken arrow in Figure 3 shows that if a child who was on the 91st centile at age 5 years did not increase BMI over the next seven years, he or she would then be on the 50th centile of weight-for-height. A strategy by which this might be achieved is discussed below. The rationale for ceasing weight control measures at age 12 years is that children during and after puberty become much more economically and emotionally independent, and some rebellion against adult authority is a normal phenomenon. If the proposed programme has achieved its objective by age 12 years there will be little need for weight control in the secondary school years.
Indeed, a primary school-based intervention study in two towns in northern France is currently assessing the impact of an educational programme in all 6-12-year-olds. The study began in 1992, and body weight evolution will be reported after the ten-year follow-up.(146)

Top of Page

Figure 3: The relationship of BMI to age in a child on the 9th, 50th or 91st centile between birth and age 20 years. (7) The broken horizontal arrow shows that if a child aged 5 years, who is then on the 91st centile, maintains a constant BMI over the next seven years, then by age 12 years he or she would be on the 50th centile of weight-for-height. The years of primary schooling offer an opportunity to prevent obesity in adolescents.

Facilities required to manage overweight among 840 primary school children

A child on the 50th centile for BMI at age 5 years has a weight of 19 kg, height 1.15 m, BMI = 14.4 kg/m². By age 12 years the corresponding values are weight 40 kg, height 1.5 m, BMI = 17.8 kg/m². A child with the same height but who is travelling along the 91st centile for BMI will weigh 23 kg, BMI = 17.4 kg/m² at 5 years, and by 12 years will weigh 50 kg, BMI = 22.2 kg/m². Therefore the 50th-centile child will gain 21 kg over the seven years in primary school, while the 91st centile child will gain 27 kg. However, if the 91st-centile child gained only 17 kg (instead of 27 kg) and continued to achieve the same height growth, the result would be that the two children would converge to the same value of BMI, as shown by the broken arrow in Figure 3. In terms of energy balance, an increase of weight by 21 kg indicates a storage of (147) Mcal (600 MJ), and an increase of 17 kg indicates a storage of 119 Mcal (500 MJ). The difference in rate of energy storage between the two over a period of seven years is approximately 11 kcal (46 kJ)/day.

The purpose of this calculation is to show that the degree of average daily energy restriction that is required to convert an overweight 5-year-old into a normal-weight 12-year-old is very small. It can be achieved by substituting fruit and low-energy drinks for sweets and high-energy drinks, and by encouraging extra exercise.

It should be noted that weight loss is not part of the objective: what is required is a slowing of weight gain.
The problem with this lies not in the thermodynamics, but in the social implications of weight control of young children. Great care is required that well-intentioned schemes do not make matters worse by causing social stigmatisation of the overweight child. It is for this reason that Figure 3 has the 9th centile line also, in order that underweight children are also selected as having special healthcare needs, as would children with problems of hearing or vision. It seems likely that the ideal model would be for children at entry to primary school entry to have their BMI determined. Those who fell below the 9th, or above the 91st, centile line in Figure 3 would be referred to a special needs nurse, who would undertake monitoring of weight and height growth of these children during the next seven years. This should be done with the minimum of fuss, and with the help of the child's parents, teachers and school physical education and catering staff , with referral to a paediatrician if necessary.

Top of Page

8 Outcome measures and audit methods

The simplest criterion of the prevalence of obesity among adults is measured height and weight so that BMI can be calculated. BMI can be audited at any level, and a cross-sectional sample is measured annually by the Health Surveys for England. The success of a programme to manage obesity will also be indicated by the decrease in co-morbidities such as heart disease, diabetes, gallstones, osteoarthritis, sleep apnoea, some sex-hormone-sensitive cancers and the psychopathologies associated with severe obesity. These criteria, though important, are difficult to interpret, since changes in the prevalence of these diseases may be confounded by changes not associated with obesity status, such as cigarette smoking.

The success of a programme for preventing obesity in children can be assessed by measuring the prevalence of overweight in children at age 12 years.

9 Information and research requirements

There are many questions to which we need answers. For example:

• The influence of dietary fat intake on obesity prevalence and treatment.
• The efficacy of different preventative programmes, especially those involving exercise.
• The relative influence of genetics and environment to the familial aggregation of obesity.
• The efficacy of strategies for maintenance of weight loss.
• The cost of preventing and treating obesity

A more extensive list of topics that require further research is set out in a recent report.(53)

Top of Page

References

1. Waterlow JC (chairman). Research on obesity: report of a DHSS/MRC group. London: HMSO, 1976

2. Bray GA (ed). Obesity in America. NIH Publication No 79-358. Washington: US Dept of Health Education and Welfare, 1979.

3. World Health Organisation (WHO), International Obesity Task Force (IOTF). Obesity: preventing and managing the global epidemic. Report of a WHO consultation on obesity, Geneva, June 3-5 1997. Geneva: WHO, 1998.

4. Seidell JC, Flegal KM. Assessing obesity: classification and epidemiology. Br Med Bull 1997; 53: 238-52.

5. Power C, Lake JK, Cole TJ. Measurement and long-term health risks of child and adolescent fatness. Int J Ob 1997; 21: 507-26.

6. Cole TJ. Changing prevalence of childhood obesity in the Western world. Int J Ob 1998; 22 (Suppl 4): S1 (abstract No 2).
   

7. Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch Dis Child 1995; 73: 25-9.

8. Rona RJ, Chinn S. The National Study of Health and Growth. Oxford: Oxford University Press.

9. Parsons T, Power C, Logan S, Summerbell CD. Childhood characteristics of adult obesity: a systematic review. Int J Ob 1999; 23(Suppl 8).

10. Must A, Jacques PF, Dallai GE, Bajema CJ, Dietz WH. Long term morbidity and mortality of overweight adolescents. N Engl J Med 1992; 327: 1350-5.

11. Crawley HF, Portides G. Self-reported versus measured height, weight and body mass index amongst 16-17 year old British teenagers. Int J Ob 1995; 19: 579-84.

12. Eckel RH, Krauss RM, for the AHA Nutrition Committee. American Heart Association call to action: obesity as a major risk factor for coronary heart disease. Circulation 1998; 97: 2099-100.

13. Manson JE, Willett WC, Stamfer MJ, Colditz GA, Hunter DJ, Hankinson SE, Hennekins CH, Speizer FE. Body weight and mortality among women. N Engl J Med 1995; 333: 677-85.

14. Bender R, Jockel K-H, Trautner C, Spraul M, Berger M. Effect of age on excess mortality in obesity JAMA 1999; 281: 1498-504.

15. The National Heart, Lung, and Blood Institute (NIH). Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Bethesda, Maryland: National Institutes of Health, 1998. Available at website: http://www.nhlbi.nih.gov/guidelines/obesity/ob_home.htm

16. Lee CD, Blair BN, Jackson AS. Cardiorespiratory fitness, body composition, and all-cause and cardiovascular disease mortality in men. Am J Clin Nutr 1998; 69: 373-80

17. Rissanen A, Heliovaara M, Knekt P, Reunanen A, Aromaa A, Maatela J. Risk of disability and mortality due to overweight in a Finnish population. BMJ 1990; 301: 835-6.

18. Lew EA, Garfinkel L. Variations in mortality by weight among 750,000 men and women. J Chronic Dis 1979; 32: 563-76.

19. Sims EAH, Danforth EJr, Horton ES, Bray GA, Glennon JA, Salans LB. Endocrine and metabolic effects of experimental obesity in man. Recent Progress Hormone Res 1973; 29: 457-96.

20. Hoffman RP, Armstrong PT. Glucose effectiveness, peripheral and hepatic insulin sensitivity in obese and lean prepubertal children. Int J Ob 1996; 20: 521-5. 

    
    Top of Page

21. Scheen AJ, Paquot N, Letiexhe MR, Paolisso G, Castillo MJ, Lefebvre PJ. Glucose metabolism in obese subjects: lessons from OGTT, IVGTT and clamp studies. Int J Obes 1995; 19 (suppl 3): S14-S20.

22. Williamson DF, Pamuk E, Thun M, Flanders D, Byers T, Heath C. Prospective study of intentional weight loss and mortality in never-smoking overweight US white women aged 40-64 years. Am J Epid 1995; 141: 1128-41.

23. Chen Y, Rennie DC, Reeder BA. Age-related association between body mass index and blood pressure: the Humbolt study. Int J Obes 1995; 19: 825-31.

24. Dattilo AM, Kris-Etherton PM. Effects of weight reduction on blood lipids and lipoproteins: a meta-analysis. Am J Clin Nutr 1992; 56: 320-8.

25. Manson JE, Colditz GA, Stamfer MJ, Wuillett WC, Rosner B, Monson RR, Speizer FE, Hennekens CH. A prospective study of obesity and risk of coronary heart disease in women. N Engl J Med 1990; 322: 822-9.Meade TW, Ruddock V, Stirling Y, Chakrabati R, Miller GJ. Fibrinolytic activity, clotting factors and long-term incidence of ischaemic heart disease in the Northwick Park study. Lancet 1993; 342: 1076-9.

26. Ernst E, Matrai A. Normalisation of hemorheological abnormalities during weight reduction in obese patients. Nutrition 1987; 3: 337-9.

27. Garfinkel L. Overweight and cancer. Ann Intern Med 1985; 103: 1034-6.

28. Davis MA, Neuhaus JM, Ettingrer WH, Mueller WH. Body fat distribution and osteoarthritis. Am J Epid 1990; 132: 701-7.

29. Spirt BA, Graves LW, Weinstock R, Bartlett SJ, Wadden TA. Gallstone formation in obese women treated by a low-calorie diet. Int J Ob 1995; 19: 593-5.

30. Galtier-Dereure F, Montpeyroux F, Boulot P, Bringer J, Jaffiol. Weight excess before pregnancy: complications and cost. Int J Ob 1995; 19: 443-8.

31. Shaw GM, Velie EM, Schaffer D. Risk of neural tube defect-affected pregnancies among obese women. JAMA 1996; 275: 1093-6.

32. Werler MM, Louik C, Shapiro S, Mitchell AA. Pre-pregnant weight in relation to risk of neural tube defects. JAMA 1996; 275: 1089-92.

33. Grunstein RR, Stenlof K, Hedner J, Sjostrom L. Impact of obstructive apnoea and sleepiness on metabolic and cardiovascular risk factors in the Swedish Obese Subjects (SOS) study. Int J Ob 1995; 19: 410-18.

    Top of Page

34. Harman EM, Block AJ. Why does weight loss improve the respiratory insufficiency of obesity? Chest 1986; 90: 153-4.

35. Chernin K. The Tyranny of Slenderness. London: The Women's Press, 1981.

36. Wadden TA, Stunkard AJ. Social and psychological consequences of obesity. Ann Intern Med 1985; 103: 1062-7.

37. Sonne-Holm S, Sorensen TI. Prospective study of attainment of social class of severely obese subjects in relation to parents social class, intelligence and education. BMJ 1986; 292: 586-9.

38. Gortmaker SL, Must A, Perrin JM, Sobol AM, Dietz WH. Social and economic consequences of overweight in adolescence and young adulthood. N Engl J Med 1993; 329: 1008-12.

39. Sullivan M, Karlsson J, Sjostrom L, Backman L,Bengtson C, Bouchard C, Dahlgren S, Jonsson E, Larsson B, Lindstedt S, Naslund I, Olbe L, Wedel H. Swedish obese subjects (SOS) - an intervention study of obesity. Baseline evaluation of health and psychosocial functioning in the first 1743 subjects examined. Int J Ob 1993; 17: 503-12.

40. Ogden J, Chanana A. Explaining the effect of ethnic group on weight concern: finding a role for family values. Int J Ob 1998; 22: 641-7.

41. Striegel-Moore RH, Wilfley DE, Caldwell MB, Needham ML, Brownell KD. Weight-related attitudes and behaviors of women who diet to lose weight: a comparison of black dieters and white dieters. Ob Res 1996; 4: 109-16.

42. Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E. Sjostrom L. Distribution of adipose tissue and risk of cardiovascular disease and death: a 12 year follow-up of participants in the population study of women in Gothenburg, Sweden. BMJ 1984; 289: 1257-61.

43. Larsson B. Regional obesity as a health hazard in men - prospective studies. Acta Med Scand Suppl 1988; 723: 45-51.

44. Serdula MK, Ivery D, Coates RJ, Freedman DS, Williamson DF, Byers T. Do obese children become obese adults? A review of the literature. Prev Med 1993; 22: 167-77.

45. Power C, Lake JK, Cole TJ. Body mass index and height from childhood to adulthood in the 1958 British born cohort. Am J Clin Nutr 1997b; 66: 1094-101.

46. Dietz WH. Critical periods in childhood for the development of obesity. Am J Clin Nutr 1994; 59: 955-9.

47. Dietz WH. Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics 1998; 101(Suppl): 518-25.

48. Stunkard A, Burt V. Obesity and the body image. II. Age at onset of disturbances in the body image. Am J Psychiatry 1967; 123: 1443-7.

49. Hill AJ, Silver EK. Fat, friendless and unhealthy: 9-year old children's perception of body shape stereotypes. Int J Ob 1995; 19: 423-30.

50. Royal College of Physicians (RCP). Clinical Management of Overweight and Obese Patients with Particular Reference to the Use of Drugs. London: RCP, 1998.

51. Scottish Intercollegiate Guidelines Network (SIGN). Obesity in Scotland: integrating prevention with weight management. Edingburgh: SIGN, 1996. Also available (updated version) at website: http://show.cee.hw.ac.uk/sign/home.htm

    Top of Page
    

52. British Nutrition Foundation (BNF). Obesity. Report of the British Nutrition Foundation Task Force (J Garrow, chairman). Oxford: Blackwell Science, 1999.

53. Department of Health. Our Healthier Nation: saving lives. London: HMSO, 1999.

54. Health Education Authority (HEA). Young and Active: draft policy framework for young people and health-enhancing physical activity. London: HEA, 1997.

55. Keys A, Menotti A, Aravanis C, Blackburn H, Djordevic BS, Buzina R, Dontas AS, Fidanza F, Karvonen MJ, Kimura N, Mohacek I, Nedeljkovic S,Puddu V, Punsar S, Taylor HL, Conti S, Kromhout D, Toshima H. The seven countries study: 2289 deaths in 15 years. Prev Med 1984; 13: 141-54.

56. Hubert HB. The nature of the relationship between obesity and cardiovascular disease. Int J Cardiol 1984; 6: 268-74.

57. Orbach S. Fat is a Feminist Issue. London: Hamlyn, 1978.

58. Apfelbaum M, Vague P, Ziegler O, Hanotin C, Thomas F, Leutenegger E. Long-term maintenance of weight loss after a very-low-calorie diet: a randomized blinded trial of the efficacy and tolerability of sibutramine. Am J Med 1999; 106: 179-84.

59. Bray GA, Blackburn GL, Ferguson JM, Greenway FL, Jain AK, Mendel CM, Mendels J, Ryan DH, Schwartz SL, Scheinbaum ML, Seaton TB. Sibutramine produces dose-related weight loss. Ob Res 1999; 7(2): 189-98.

60. Department of Health. Health Survey for England: Adult references tables 1997 (unpublished). Available at: www.doh.gov.uk/stats/hstab97/intro.htm

61. Prescott-Clarke P, Primatesta P. Health Survey of England 1996. London: HMSO, 1998.

62. Effery RW, Folsom AR, Leupker RV, Jacobs DR, Gillum RF, Taylor HL, Blackburn H. Prevalence of overweight and weight loss behaviour in a metropolitan adult population: the Minnesota heart survey experience. Am J Public Health 1984; 74: 349-52.

63. Summerbell CD. Teaching nutrition to medical doctors: the potential role of the State Registered Dietitian. J Hum Nutr Diet 1996; 9: 349-56.

64. Kopelman, P. Place of obesity clinics in the NHS. Brit J Hosp Med 1993; 49 (8); 533-5.

65. Grace C, Summerbell C, Kopelman P. An audit of dietary treatment modalities and weight loss outcomes in a specialist obesity clinic. J Hum Nutr Diet 1998; 11(3): 197-202.

66. Ogden J, Hoppe R. The relative effectiveness of two styles of educational package to change practice nurses' management of obesity. Int J Obes 1997; 21: 963-71.

67. Cowburn G, Summerbell C. A survey of dietetic practice in obesity management. J Hum Nutr Diet 1998; 11(3): 191-6.

68. West R. Obesity. London: Office of Health Economics, 1994.

    Top of Page
    

69. Bagust A, Roberts BL, Haycox AR, Barrow MA. The additional cost of obesity to the health service and the potential for resource savings from effective interventions. Euro J Pub Health 1999; 9: 000-000.

70. Seidell JS. What is the cost of obesity? Int J Obes 1997; 21 (suppl 2): S1 (abstract No 1).

71. Levy E, Levy P, Le Pen C, Basdevant A The economic cost of obesity: the French situation. Int J Obes 1995; 19: 788-92.

72. Colditz GA. Economic costs of obesity. Am J Clin Nutr 1992; 55: 503S-7S.

73. Quesenberry CP, Caan B, Jacobson A. Obesity, health services use, and health care costs among members of a health maintenance organisation. Arch Intern Med 1998; 158: 466-72.

74. Stephenson J, Imrie J. Why do we need randomised controlled trials to assess behavioural interventions? BMJ 1999; 316: 611-13.

75. Summerbell CD, Higgins JPT. Systematic reviews and meta-analysis. British Nutrition Foundation (BNF) Nutrition Bulletin 1997; 22: 111-18.

76.  Douketis JD, Feightner JW, Attia J, Feldman WF, with the Canadian Task Force on Preventative Health Care. Periodic health examination, 1999 update: 1. Detection, prevention and treatment of obesity. J Can Med Ass 1999; 160: 513-25.

77. Glenny A-M, O'Meara S, Melville A, Sheldon TA, Wilson C. The treatment and prevention of obesity: a systematic review of the literature. Int J Ob 1997; 21: 715-37. Also available at website: www.york.ac.uk/inst/crd/obesity.htm

78. Skender ML, Goodrick GK, Del Junco DJ, Reeves RS, Darnell L, Gotto AM, Foreyt JP. Comparison of 2-year weight loss trends in behavioural treatments of obesity: diet, exercise, and combination intervention. J Am Diet Ass 1996; 96: 342-6.

79. Hakala P, Karvetti RL. Weight reduction on lactovegetarian and mixed diets. Changes in weight, nutrient intake, skinfold thickness and blood pressure. Eur J Clin Nutr 1989; 43: 421-30.

80. Katzel LI, Bleecker ER, Colman EG, Rogus EM, Sorkin JD, Goldberg AP. Effects of weight loss vs aerobic exercise training on risk factors for coronary disease in health, obese, middle-aged and older men. A randomised controlled trial. JAMA 1995; 274: 1915-21.

81. Schlundt DG, Hill JO, Pope-Cordle J Arnold D, Virts KL, Katahn M. Randomised evaluation of a low fat ad libitum carbohydrate diet for weight reduction. Int J Ob 1993; 17: 623-9.

82. Pascale RW, Wing RR, Butler BA, Mullen M, Bononi P. Effects of a behavioural weight loss program stressing calorie restriction versus calorie plus fat restriction in obese individuals with NIDDM or a family history of diabetes. Diabetes Care 1995; 18: 1241-8.

83. Baron JA, Schori A, Crow B, Carter R, Mann JI. A randomised controlled trial of low carbohydrate and low fat/high fibre diet. Am J Public Health 1986; 76: 1293-6.

84. Ryttig KR, Tellnes G, Haegh L, Boe E, Fagerthun H. A dietary fibre supplement and weight maintenance after weight reduction: a randomised, double blind placebo controlled long term trial. Int J Ob 1989; 13: 165-71.

85. Wadden TA, Stunkard AJ. Controlled trial of very-low calorie diet, behavioural therapy, and their combination in the treatment of obesity. J Consult Clin Psychol 1986; 54: 482-8.

    Top of Page
    

86. Muira J, Arai K, Tsukahara S, Ohno M, Ikeda Y. The long term effectiveness of combined therapy by behaviour modification and very low calorie diets: 2 years follow-up. Int J Obes 1989; 13 (Suppl 2): S39-S46.

87. Torgerson JS, Lissner L, Lindroos AK, Kruijer H, Sjostrom L. VLCD plus dietary and behavioural support versus support alone in the treatment of severe obesity. A randomised two-year clinical trials. Int J Ob 1997; 21: 987-94.

88. Ryttig KR, Flaten H, Rossner S. Long term effects of a very low calorie diet (Nutrilett) in obesity treatment. A prospective, randomised, comparison between VLCD and a hypocaloric diet + behaviour modification and their combination. Int J Ob 1997; 21: 574-9.

89. Anderson JW, Brinkman-Kaplan V, Hamilton CC, Logan JE, Collins RW, Gustafson NJ. Food containing hypocaloric diets are as effective as liquid supplement diets for obese individuals with NIDDM. Diabetes Care 1994; 17: 602-4.

90. Ditschuneit HH, Flechtner-Mors M, Johnson TD, Alder G. Metabolic and weight loss effects of a long-term dietary intervention in obese patients. Am J Clin Nutr 1999; 69: 198-204.

91. Summerbell CD, Watts C, Higgins JPT, Garrow JS. Outpatient weight-reduction diets should be novel, simple and well-supervised. BMJ 1998; 317: 1487-9.

92. Wing RR, Jeffery RW, Burton LR, Thorson C, Nissinoff KS, Baxter JE. Food provision versus structured meal plans in the behavioural treatment of obesity. Int J Ob 1996; 20: 56-62.

93. Jeffery RW, Wing RR, Thorson C, Burton LR, Raether C, Harvey J, Mullen M. Strengthening behavioural interventions for weight loss: a randomised trial of food provision and monetary incentives. J Consult Clin Psychol 1993; 61: 1038-45.

94. Jeffery RW, Wing RR. Long-term effects of interventions for weight loss using food provision and monetary incentives. J Consult Clin Psychol 1995; 63: 793-6.

95. Wing RR, Epstein LH, Nowalk MP, Koeske R, Hagg S. Behaviour change, weight loss improvements in type 2 diabetics. J Consul Clin Psychol 1985; 53: 111-22.

96. Wing RR, Marcus MD, Epstein LH, Jawad A. A family based approach to the treatment of obese type 2 diabetic patients. J Consul Clin Psychol 1991; 59: 156-62.

97. Wing RR, Blair E, Marcus M, Epstein LH, Harvey J. Year long weight loss treatment for obese patients with type 2 diabetes: does including an intermittent very-low calorie diet improve outcome? Am J Med 1994; 97: 354-62.

98. Garrow JS, Summerbell CD. Meta-analysis: effect of exercise, with or without dieting, on body composition of overweight subjects. Eur J Clin Nutr 1994; 49: 1-10.

99. Andersen RE, Wadden TA, Bartlett SJ, Zemel B, Verde TJ, Franckowiak SC. Effects of lifestyle activity vs structured exercise in obese women. JAMA 1999; 281: 335-40.

    Top of Page
    

100. King AC, Frey Hewitt B, Dreon DM, Wood PD. Diet vs exercise in weight maintenance: the effects of minimal intervention strategies on long-term outcomes in men. Arch Intern Med 1989; 149: 2741-6.

101. Pavlou KN, Krey S, Steffee WP. Exercise as an adjunct to weight loss and maintenance in moderately obese subjects. Am J Clin Nutr 1989; 49: 1115-23.

102. Johnson WG, Stalonas PM, Christ MA, Pock SR. The development and evaluation of a behavioural weight reduction program. Int J Ob 1979; 3: 229-38.

103. Jones SP, Smith IG, Kelly F, Gray JA. Long term weight loss with sibutramine. Int J Obes 1995; 19 (Suppl 2): 41 (abstract 071).

104. Bray GA, Ryan DH, Gordon D, Heidingsfelder S, Cerise F, Wilson K. A double-blind randomised placebo controlled trial of sibutramine. Obes Res 1996; 4: 263-70.

105. Lean MEJ. Sibutramine - a review of clinical efficacy. Int J Ob 1997; 21: S30-S36.

106. Jones SP, Newman BM, Romance FM. Sibutramine hydrochloride: weight loss in overweight subjects. Int J Ob 1994; 18(suppl 2): 61 (abstract 0234).

107. Drouin P, Hanotin C, Coureier S, Leutenegger E. A dose ranging study: efficacy and tolerability of sibutramine in weight loss. Int J Ob 1994; 18(Suppl 2): 60 (abstract 0232).

108. Bray GA, Blackburn GL, Ferguson JM, Greenway FL, Jain A, Kaiser Pe, Mendels J, Ryan D, Schwartz SL. Sibutramine - dose response and long-term efficacy in weight loss, a double blind study. Int J Ob  
     1994; 18(Suppl 2): 60 (abstract 0231).

109. Garrow JS. Flushing away the fat. BMJ 1998; 317: 830-1.

110. Drent ML, van der Veen EA. Lipase inhibition: a novel concept in the treatment of obesity. Int J Ob 1993; 17: 241-4.

111. Drent ML, Larsson I, William-Olsson T, Quaade F, Czubayko F, von Bergmann K, Strobel W, Sjostrom L, van der Veen EA. Orlistat (Ro 18-0647), a lipase inhibitor, in the treatment of human obesity: a multiple dose study. Int J Ob 1995; 19: 221-6.

112. James WPT, Avenil A, Broom J, Whitehead J. A one-year trial to assess the value of orlistat in the management of obesity. Int J Obes 1997; 21(Suppl 3): S24-S30.

113. Sjostrom L, Rissanen A, Andersen T, Boldrin M, Golay A, Koppeschaar HPF, Krempf M. Randomised placebo-controlled trial of orlistat for weight loss and prevention of weight regain in obese patients. Lancet 1998; 352: 167-73.

114. Davidson MH, Hauptman J, DiGirolamo M, Foreynt JP, Halstead CH, Heber D, Heimburger DC, Lucas CP, Robbins DC, Chung J,Heymsfield SB. Weight control and risk factor reduction in obese subjects treated for 2 years with orlistat: a randomized controlled trial. JAMA 1999; 281: 235-42.

115. Williamson DF. Pharmacotherapy for obesity JAMA 1999; 281: 278-80.

     Top of Page
     

116. Lechner GW, Elliot DW. Comparison of weight loss after gastric exclusion and partitioning. Arch Surgery 1983; 118: 685-92.

117. Andersen T, Stockholm KH, Backer OG, Quaade F. Long-term (5-year) results after either horizontal gastroplasty or very-low-calorie diet for morbid obesity. Int J Ob 1988; 12: 277-84.

118. Sugerman HJ, Starkey JV, Birkenhauer R. A randomised prospective trial of gastric bypass versus vertical banded gastroplasty for morbid obesity and their effects on sweets versus non sweet eaters. Ann Surgery 1978; 205: 613-24.

119. MacLean LD, Rhode BM, Sampalis J, Forse RA. Results of the surgical treatment of obesity. Am J Surgery 1993; 165: 155-60.

120. Howard L, Malone M, Michalek A, Carter J, Alger S, Van Woert J. Gastric bypass and vertical banded gastroplasty - a prospective randomised comparison and 5-year follow up. Ob Surgery 1995; 5: 55-60.

121. The National Heart, Lung, and Blood Institute (NIH). Consensus Statement on Gastrointestinal Surgery for Severe Obesity. NIH Consensus Development Conference, 25-27 March 1991.

122. Garrow JS. Dental splinting in the treatment of hyperphagic obesity. Proc Nutr Soc 1974; 33: 29A.

123. Rodgers S, Burnet R, Goss A, Phillips P, Goldney R, Kimber C, Thomas D, Harding P, Wise P. Jaw wiring in the treatment of obesity. Lancet 1977; 1: 1221-3.

124. Garrow JS, Gardiner GT. Maintenance of weight loss in obese patients after jaw wiring. BMJ 1981; 282: 858-60.

125. Garrow JS, Webster JD. Long-term results of treatment of severe obesity with jaw wiring and waist cord. Proc Nutr Soc 1986; 45: 119A.

126. Mustajoki P, Pekkarinen T. Maintenance programmes after weight reduction - how useful are they? Int J Ob 1999; 23: 553-5.

127. Hardeman W, Griffin S, Johnston M, Kinmouth A l Wareham NJ. Interventions to prevent weight gain: a systematic review. Int J Ob 2000; 24: 000-000.

128. Taylor CB, Fortmann SP, Flora J, Kayman S, Barrett DC, Jatulis D, Farqhar JW. Effect of long term community health education on body mass index: the Stanford Five City Project. Am J Epid 1991; 134: 235-49.

129. Forster JL, Jeffery RW, Schmid TL, Kramer FM. Preventing weight gain in adults: a pound of prevention. Health Psychol 1988; 7: 515-25.

130. Jeffery RW, Gray CW, French SA, Hellerstedt WL, Murray D, Luepker RV, Blackburn H. Evaluation of weight reduction in a community intervention for cardiovascular disease risk: changes in body mass index in the Minnesota Heart Health program Int J Ob 1995; 19: 30-9.

131. Barlow SE, Dietz WH. Obesity evaluation and treatment: expert committee recommendations. Pediatrics 1998; 102: E29. Also available at http://www.pediatrics.org

132. Floodmark CE, Ohlsson T, Ryden O, Sveger T. Prevention of progression to severe obesity in a group of obese school children treated with family therapy. Pediatrics 1993; 91: 880-4.

     Top of Page
     

133. Harvey EL, Glenny A-M, Kirk SFL, Summerbell CD. Effective Professional Practice: a systematic review of health professionals' management of obesity. The Cochrane Library. Issue 1. Oxford: Update Software, 1999. Int J Ob 1999; 23 (12): 1213-22.

134. French SA, Jeffery RW, Folsom AR, McGovern P, Williamson DF. Weight loss maintenance in young adulthood: prevalence and correlations with health behaviour and disease in population-based sample of women aged 55-69 years. Int J Ob 1996; 20: 303-10.

135. Allison DB, Zannolli R, Faith MS, Heo M, Pietrobelli A, Van Itallie TB, Pi-Sunyer FX, Heymsfield SB. Weight loss increases and fat loss decreases all-cause mortality rate: results from two independent cohort studies. Int J Ob 1999; 23: 603-11.

136. Yanovski SZ, Bain RP, Williamson DF. Report of a National Institute of Health-Centres for Disease Control and Prevention Workshop on the feasibility of conducting a randomised clinical trial to estimate the long-term health effects of intentional weight loss in obese persons. Am J Clin Nutr 1999; 69: 366-72.

137. Blackburn GL. Benefits of weight loss in the treatment of obesity (editorial) Am J Clin Nutr 1999; 69: 347-9.

138. Narbro K, Agren G, Jonsson E, Larsson B, Naslund I, Wedel H, Sjostrom L. Sick leave and disability pension before and after treatment for obesity: a report from the Swedish Obese Subjects (SOS) study. Int J Ob 1999; 23: 619-24.

139. Bush A, Webster J, Chalmers G, Pearson M, Penfold P, Brereton P, Garrow JS. The Harrow slimming club: Report on 1090 enrollments in 50 courses, 1977-1986. J Hum Nutr Diet 1988; 1: 429-36.

140. Karvetti RL, Hakala P. A 7-year follow-up of a weight reduction programme in Finnish primary health care. Eur J Clin Nutr 1992; 46: 743-52.

141. Andres R, Elahi D, Tobin JD, Muller DC, Brant L. Impact of age on weight goals. Ann Int Med 1985; 103: 1030-3.

142. French SA, Jeffery RW, Folsom AR, Williamson DF, Byers T. Weight variability in a population-based sample of older women: reliability and intercorrelation of measures. Int J Ob 1995; 19: 22-9.

143. Garrow JS, Webster JD, Pearson M, Pacy PJ, Harpin G. Inpatient-outpatient randomised comparison of Cambridge diet versus milk diet in 17 obese women over 24 weeks. Int J Obes 1989; 13: 521-9.

144. Poskitt EME, Cole TJ. Do fat babies stay fat? BMJ 1977; 1: 7-9.

145. Basedevant A, Boute D, Borys JM. Who should be educated? Education strategies: could children educate their parents? Int J Ob 1999; 23(Suppl 4): S10-S13.

     Top of Page