You are on page 1of 24

OBESITY IS THE EXCESS accumulation of body fat.

It can be assessed by the body


mass index (weight in kilograms divided by the square of height in metres
[BMI]). The relationship of BMI to total body and visceral fat, and consequent
complications, varies between ethnic groups1 (Box 1). Asian populations,
particularly those from South East Asia and the Indian subcontinent, have more
fat and more comorbidities for any given BMI, resulting in different suggested
BMI cut-off points.3 In contrast, Polynesians have higher BMI cut-off points.4

The risk of comorbidities rises with increasing BMI, with a mild rise in the
overweight range, moderate in class I, severe in class II and very severe in
class III obesity. As well as the total fat, distribution is important to the
incidence of comorbidities. Central abdominal fat, particularly visceral fat, is a
risk factor for the metabolic syndrome. Waist circumference cut-off points in
white European populations are shown in Box 2.
Adult BMI cut-off points cannot be used for assessing children and adolescents,
as BMI varies throughout childhood, being high in the 2nd year of life, dropping
to a nadir at age 47 years and then rising again to adult values (Box 3). BMI-
for-age charts, such as those developed by the Centers for Disease Control and
Prevention in the United States,6 can be used in clinical practice to assess and
monitor BMI over time in children (Box 3). Overweight is defined as BMI
between the 85th and 95th percentiles, and obesity as BMI > 95th
percentile,7 but these definitions are arbitrary, as, unlike adult BMI cut-offs,
they are not linked to morbidity data. Although a table of age- and sex-specific
cut-offs developed for epidemiological research8 allows international comparison
of trends in overweight and obesity in children, it is not intended for routine
clinical use.

In children and adolescents, as in adults, waist circumference is strongly


correlated with abdominal fat and markers for comorbidities, such as adverse
lipid and glucose profiles and hypertension.9 However, there are no
internationally accepted criteria for waist circumference in this age group and,
as for adults, racial and ethnic variations exist. For example, African American,
Mexican American and Mohawk Indian children carry more abdominal fat than
white children.10

The epidemic
The prevalence of obesity appears to have more than doubled in Australian
adults in the decades from 1980 to 2000. A 1980 survey of people aged 25
64 years living in capital cities found that 7.1% were obese.11 In contrast, the
19992000 Australian Diabetes, Obesity and Lifestyle (AusDiab) study found in
the same urban age group that 18.4% were obese, a 2.5-fold
increase.12 Overall, the AusDiab study found, among 11 000 adults from around
Australia, that 48.2% of men and 29.9% of women were overweight, while
19.3% of men and 22.2% of women were obese.12

The increase in prevalence of overweight and obesity is even more striking


among Australian children.13 Over the decade 19851995, the combined
prevalence of the two conditions in children almost doubled (from 10.7% to
19.5% in boys and from 11.8% to 21.1% in girls), while that of obesity on its
own more than tripled (from 1.4% to 4.5% in boys and from 1.2% to 5.3% in
girls).13

Aetiology and effects of obesity


The aetiology of obesity is complex, with both environmental and genetic
influences. The recent increase in prevalence is clearly due to the continuous
availability of high-energy foods, together with a major reduction in the
obligatory need for physical activity that has characterised human existence
until very recently.

However, not everyone in an obesogenic environment becomes obese,


indicating that a genetic predisposition is required. Studies of twin pairs reared
together or apart suggest that about 70% of the influence on body weight is
genetic, while about 30% is environmental.14 The dominance of genetic
influences has been confirmed by adoption studies, which found that adoptees
resemble their biological parents in body size, with very little resemblance to
their adoptive parents.15 Some genes that could predispose to obesity have
been identified (including genes encoding leptin, the leptin receptor, pro-
opiomelanocortin and the melanocortin-4 receptor), but many more are
probably as yet undiscovered.16

In children, environmental contributors to obesity include prolonged television


viewing, playing of computer and video games,17 decreased physical activity
(especially incidental activity) and increased consumption of energy-dense foods
and sugar-containing drinks.18

Obesity has been linked to complications in many body systems (Box 4).

Preventing obesity
The magnitude of the problem of obesity in most westernised countries
mandates population-level strategies. Interventions focusing on simply
educating individuals and communities about behaviour change have had
limited or no success.19 Instead, there is a need to produce an environment
that supports healthy eating and physical activity throughout the community.
A recent systematic review of interventions for preventing obesity in children
concluded that there are limited high quality data on the effectiveness of
prevention programs.20 This may well reflect the methodological and ethical
challenges of conducting such studies, often in a sociopolitical environment not
conducive to change. However, the review did highlight the potential effects of
reducing sedentary behaviours and increasing physical activity.

Possible strategies include:

regulating the nature and amount of food advertising directed at


children;

providing high-quality recreation areas, safe cycle paths and safe


street lighting in local neighbourhoods;

improving public transport;

providing economic incentives for production and distribution of


vegetables and fruit; and

developing town planning policies that promote active or public


transport over private motor cars.

Such multifaceted large-scale interventions require cooperation and support


from many sectors of society and government, with adequate resourcing and
significant community ownership.

Lessons may be learnt from previous successful campaigns for long-term social
change, such as tobacco control and the promotion of breastfeeding. 21 However,
as few previously studied behaviours are driven by biological drives as powerful
as the drive to eat, it is imperative that the efficacy of any public health
measure be tested before it is widely applied.

Individual therapy for obesity


A weight loss and maintenance strategy for individual patients is outlined inBox
5. The weight-loss phase involves establishing a negative energy balance by
reducing energy intake and increasing physical activity. Most motivated patients
can succeed in this phase, but different strategies suit different individuals.

The next phase, weight maintenance, is more difficult and requires more
emphasis. Lifestyle modification plus pharmacotherapy, if required, offer the
best chance of success. For those who have failed in weight maintenance,
bariatric surgery currently produces the best results and should be considered.

Clearly, it is important to set realistic goals from the start. Weight loss that is
useful from a medical point of view (5%10%) often does not satisfy patients.
It may help to point out that more modest losses may be easier to maintain in
the long term.

Lifestyle change in adults


It is important to ensure that patients are ready and enthusiastic about
attempting weight loss, as their cooperation is essential. A negative energy
balance must be established, by reducing energy intake and increasing energy
expenditure. A practical way to reduce energy intake is to limit fat intake.
Patients may need information about foods that contain fat. It may help to also
limit carbohydrates, especially those with a high glycaemic index (rapidly
digested and absorbed carbohydrates), as high insulin levels can encourage
weight gain22,23 (see also National Health and Medical Research guidelines on
managing obesity,7 section 5, available
atwww.obesityguidelines.gov.au/pdf/adults_part5.pdf). A reduction of 2500 kJ
(600 calories) from the stable prior intake is generally advised, which should
lead to weight loss of 0.6 kg per week. This can be calculated from a diet diary
kept over a week before starting treatment. The advice of a trained dietitian will
be of great help.

Increased physical activity is an important component of lifestyle modification.


The increase must be substantial (80 minutes of moderate-intensity activity per
day), but cannot usually be achieved immediately. A more modest initial target
can be set (eg, 30 minutes of walking 35 days per week) and increased
gradually. While reduced energy intake is the major method of losing weight, it
has been shown that a high level of physical activity is essential to assist with
maintaining weight loss. Irrespective of its impact on weight, physical activity
has wider benefits on well-being, including improved cardiovascular fitness.

A Cochrane review on the effectiveness of intervention by health professionals


in weight management concluded that there are few solid leads about
improving obesity management, although reminder systems, brief training
interventions, shared care, in-patient care and dietitian-led treatments may all
be worth further investigation.24

Lifestyle change in children and adolescents


For children and adolescents, the broad principles of management are well
recognised and comprise behaviour modification (eg, patterns of television
watching, computer use, socialising and leisure activity), family support, dietary
change, increased physical activity, decreased sedentary behaviour, and a
developmentally appropriate approach.

To be effective, therapy must take account of family influences on food and


activity habits.24 Parental involvement is necessary with both young children
and adolescents. Several studies have now shown that long-term (210-year)
maintenance of weight loss can be achieved with family-based
intervention.25 There is no direct evidence on which dietary modification is most
effective in this age group. Interventions should follow national nutrition
guidelines and emphasise lower-fat options, increasing vegetable and fruit
intake, healthier snack-food choices and probably decreased portion sizes.
Reducing soft drink and fruit juice intake is also important.18

Participation in an exercise program is a long-term predictor of successful


weight control in children.26 The type of exercise appears important for
sustained weight loss. While both lifestyle (eg, walking and cycling) and
programmed aerobic exercise promote weight loss in the initial phase, lifestyle
exercise is more likely to be continued long term.

In pre-adolescents, an approach using parents as the exclusive agents of


lifestyle change appears superior to a child-centred approach, where the child is
expected to take major responsibility for making changes.27,28 Thus, sessions
involving one or both parents without the child are likely to be the most
effective.

In adolescents, features of successful interventions include separate sessions for


the adolescent and parent, and a structured but flexible program that
encourages sustainable modifications in lifestyle, relationships and
attitudes29 (see case report, Box 6). There is also a report of success, at least in
the short term (3 months), with a 4-month behavioural weight control program
for overweight adolescents initiated in a primary care setting and extended
through telephone and mail contact.30

The National Health and Medical Research Council recently published clinical
practice guidelines for management of overweight and obesity in children and
adolescents, which review in detail the evidence on interventions. 31

Drug therapy
Role of leptin
The use of drugs to assist in management of obesity has always been
controversial. This is partly due to the widespread belief that losing weight and
maintaining weight loss is simply a matter of exercising free will, which has
even led to a suggestion that it is unethical to be obese.32 However, this view
cannot explain why the overwhelming majority of obese individuals regain
weight after successful weight loss,33 despite an often desperate desire to
remain lean. The discovery of the hormone leptin in 1994 shed light on this
paradox.

Leptin is a cytokine hormone produced in fat cells in proportion to their size. It


is secreted into the bloodstream and crosses the bloodbrain barrier via an
active, saturable process. It acts in the hypothalamus to alter the expression of
neurotransmitters (eg, it inhibits production of neuropeptide Yand stimulates
production of melanocyte-stimulating hormone), resulting in suppression of
hunger and an increase in spontaneous activity. Hence, as weight increases, fat
cells increase in number and size, increasing the production of leptin, which
feeds back to inhibit food intake and increase energy expenditure, limiting
weight gain. It appears that a genetically lean individual will gain an extra 7
8 kg before leptin increases sufficiently to stop the weight gain. Individuals who
gain more than this amount must be unresponsive to the hormone, either
because it cannot enter the brain efficiently or because there is a mutation in
one of the many steps required for leptin action, such as at the melanocortin
receptor. When obese individuals with a high leptin level lose weight, the leptin
level falls dramatically,34 resulting in relative leptin deficiency.

It is known from a study of two children with a homozygous mutation in the


leptin gene that leptin deficiency leads to insatiable hunger.35 Thus, it is likely
that weight-reduced obese individuals regain weight despite a great desire to
remain lean because they cannot tolerate hunger long term in the face of
abundant food. This has led to reappraisal of the role of drugs and surgery in
management of obesity (see case report, Box 7).

Available drugs
Drugs available for treating obesity are shown in Box 8. Both the intestinal
lipase inhibitor orlistat and the serotonin- and noradrenaline-reuptake inhibitor
sibutramine have been shown to limit weight regain in large randomised
placebo-controlled trials.37,38 The lack of long-term studies of the noradrenergic
agonists phentermine and diethylpropion limits their usefulness in long-term
management.

Some antidepressants affect body weight, including the selective serotonin-


reuptake inhibitors (fluoxetine, paroxetine, fluvoxamine, citalopram and
sertraline) and the serotonin- and noradrenaline-reuptake inhibitor venlafaxine.
The effects of fluoxetine are best studied, with several studies showing a
modest but reproducible effect on weight loss. These drugs are not approved for
treating obesity, but should be the drugs of choice when treating depression in
overweight patients.

Many compounds are currently undergoing clinical trials including:

leptin and leptin analogues;

topiramate, an antiepileptic drug with appetite-suppressant action;

rimonabant, an inhibitor of the cannabinoid-1 receptor;


amylin, a protein secreted by pancreatic beta cells; and

AOD 9604, a fragment of the growth hormone molecule.

Many other compounds are in earlier stages of development.

Surgery
Surgery should be considered for patients with significant comorbidities
associated with obesity, especially if medical therapy has failed repeatedly.
Studies in Australia39 and overseas40 have shown major benefits of gastric
surgery on weight loss, diabetes, hypertension, dyslipidaemia, insulin
sensitivity, sleep apnoea, asthma, infertility and quality of life. For example, a
study of 50 patients with type 2 diabetes and an initial mean BMI of
48 kg/m2 found that, a year after laparoscopic gastric banding, fasting serum
glucose level had decreased from 9.4 to 6.2 mmol/L, glycosylated haemoglobin
level from 7.8% to 6.2%, serum triglycerides from 2.4 to 1.4 mmol/L, and
blood pressure from 154/96 to 130/79 mmHg, while high-density lipoprotein
cholesterol level had increased from 1.03 to 1.22 mmol/L.41 Of the 29 patients
who had been taking oral hypoglycaemic agents, only eight still required these,
after an average 27 kg weight loss. While laparoscopic gastric banding (Box 9)
is becoming the favoured approach because of its reversibility and low
morbidity, some surgeons still perform gastric stapling or biliopancreatic
diversion with good outcomes.42Bilio-pancreatic diversion is indicated when
previous restrictive surgery has failed and could be considered in superobesity.

Box 1: Definitions of obesity in adults2

Box 2: Waist circumference associated with increased risk of


metabolic complications in white European adults*

Box 3: Example of chart of body mass index (BMI) for age6


Box 4: Complications of obesity
Cardiovascular: Hypertension, dyslipidaemia, increased risk of coronary heart
disease and stroke

Respiratory: Obstructive sleep apnoea, asthma

Endocrine: Glucose intolerance, insulin resistance, type 2 diabetes, polycystic


ovary syndrome

Orthopaedic: Back pain, osteoarthritis, flat feet

Dermatological: Acanthosis nigricans, skin tags, intertrigo

Gastrointestinal: Non-alcoholic steatohepatitis, reflux oesophagitis, gallstones

Psychosocial: Social isolation and discrimination, decreased self-esteem,


binge-eating disorder, bulimia, and depression

Other: Increased risk of breast and other cancers, increased intracranial


pressure, proteinuria

Box 5: Management of obesity in motivated adult patients


Step 1. History, examination and investigation (plus management of
comorbidities if not resolved by weight loss)

Step 2. Weight loss (312 months)

Lifestyle modification (decrease energy intake and increase physical


activity)

Diet

Low-fat, reduced-carbohydrate diet

Very low energy diet (VLED) (commercial, over-the-counter


mixture of essential nutrients with defined energy content, usually
supplied as a powder to be mixed with water and drunk three times daily)
Meal replacement program (eg, commercial program that
provides preprepared meals)

Commercial weight loss centre

Step 3: Weight maintenance (lifelong)

Lifestyle modification

Pharmacotherapy (if required)

Step 4: Bariatric surgery (for obese patients who have failed medical
therapy)

Box 6: Case report an overweight adolescent


Presentation: A 13-year-old girl presented to her general practitioner with a
respiratory tract infection. During the consultation, her mother commented that
she was concerned about her weight and was being teased about this at school.
She had left her previous school because of bullying. Aware of the time needed
to discuss this issue, the GP made a separate appointment for the girl and her
mother to return.

History: The girl was an only child and appeared to have good relationships
with her parents and peers. Her general health was good. Several family
members were obese (mother, maternal aunts, three grandparents), and the
mother described the family as heavy-boned. However, there was no
significant family history of other disorders associated with insulin resistance
(eg, type 2 diabetes, premature heart disease). The girls lifestyle was
sedentary, with her main interests being playing music, sewing, reading and
talking on the telephone. She was driven to and from school and watched about
3 hours of television per day. Her diet included at-risk features, such as
occasionally skipping breakfast, consuming full cream milk, something nice for
morning and afternoon tea and about 500 mL of soft drink per day, regularly
buying food from a milk-bar in the afternoon, and grazing at home (eg, on
biscuits from the cupboard).

Examination: Weight was 72.6 kg (> 97th percentile) and height 161.5 cm
(< 75th percentile), giving a body mass index (BMI) of 28.0 kg/m2 (> 95th
percentile for age; adult overweight range). Her waist circumference was 85 cm
(adult at risk of metabolic complications range). She was in mid-puberty, with
no striae on the abdomen or upper thighs. Blood pressure was 120/80 mmHg.
Investigations: A fasting blood test showed normal lipid profile, liver function,
glucose and insulin levels. Urinalysis gave normal results.

Management: The GP arranged to see the girl and her mother together and
separately, initially every 3 weeks and then less often. Two visits to a local
dietitian were arranged; a long waiting list made further follow-up impossible.
The girl was encouraged to set her own goals for food and activity changes,
which were discussed with the GP at each consultation. The family was
supported in changing their eating patterns and television use. The dietitian
looked with the girl at ways to recognise and deal with eating cues.

Course: Over time, the mother herself started to lose weight because of altered
cooking practices and being more active. Water rather than soft drink was
offered at the evening meal, biscuits and less-healthy snacks were no longer
stored in the cupboards, and the whole family moved to eating more vegetables
and having smaller meat portions at the evening meal. The girl ate something
for breakfast each morning and started walking to and from school each day.
She began tennis lessons and found an interest in tap dancing.

Ten months later, the girls weight was 69.3 kg, height 163.0 cm, BMI
26.1 kg/m2, and waist circumference 80 cm. She reported being fitter and said
that she was greatly enjoying school and was no longer being bullied.

Box 7: Case report relapsing obesity in an adult


Presentation: A 42-year-old woman was referred for management of severe
obesity and type 2 diabetes.

History: She had been thin as a child but started to gain weight after the birth
of her first child, 20 years before. She had attempted weight loss several times
previously, at one time losing 30 kg using a commercial weight-loss program,
but always regained the weight.

Two years prior she was diagnosed with diabetes after presenting with polyuria
and thrush. This was treated with metformin (1000 mg twice daily) and
glibenclamide (5 mg once daily). She had a past history of
hypercholesterolaemia and hypertension, treated with atorvastatin (40 mg in
the morning) and ramipril (2.5 mg in the morning), and pre-eclampsia during
one pregnancy. She had a maternal history of obesity and diabetes. She was a
non-smoker. She was married with two daughters and was a full-time mature-
age university student.

Examination: She appeared well. Weight was 114.5 kg, height 1.65 m (body
mass index, 42.1 kg/m2, corresponding to obesity class III), waist
circumference 125 cm, and hip circumference 131 cm. Pulse was regular at
98 beats/min, blood pressure was 130/90 mmHg. She had mild acanthosis
nigricans around the neck and a few skin tags. Cardiorespiratory examination
revealed no abnormalities. Pulses were difficult to feel. There was no evidence
of microangiopathy.

Investigations: Serum levels were: glycosylated haemoglobin, 10.2%


(reference range [RR], < 6.1%); total cholesterol, 4.8 mmol/L (RR,
< 5.6 mmol/L); triglyceride, 3.3 mmol/L (RR, < 2.6 mmol/L); high-density
lipoprotein (HDL) cholesterol, 0.86 mmol/L (RR, > 1.0 mmol/L); low-density
lipoprotein (LDL) cholesterol, 3.34 mmol/L (RR, < 3.51 mmol/L); LDL/HDL
cholesterol ratio, 3.9 (RR, < 3.6). Serum levels of electrolytes, urea and
creatinine, and liver function results, were in the reference range.

Management: The patient said that she was ready to attempt weight loss as
she understood that obesity was contributing to her diabetes, hypertension and
dyslipidaemia. A diet and exercise regimen was discussed, and she was referred
to a dietitian for help in calculating a 2 500 kJ (600 calorie) deficit diet. She
undertook to try to increase her physical activity.

1-month review: The patient said she was struggling, as combining full-time
study, part-time work and maintaining a household left little time for exercise
and, although she tried to change her diet, she found it difficult not to snack.
She said she had found it easier to diet when she was younger. Her diabetes
had not improved.

Management: Given the patients medical comorbidities, it was worth


considering more aggressive management of obesity.

Gastric banding was discussed. She had considered this after seeing it on
television. However, she had no private insurance, and waiting lists at the few
public hospitals that perform the operation are up to 5 years.

A medical strategy was adopted. She began a 3-month course of:

A very low energy diet (VLED), comprising a sachet of essential


nutrients mixed with water three times daily (providing vitamins, minerals,
amino acids, essential fatty acids and 1914 kJ [456 calories] per day). This
was purchased over-the-counter from a pharmacist.

A bowl of vegetables or salad (to supply roughage) sprinkled with a


teaspoon of olive oil (to contract the gall bladder) in the evenings.
She stopped taking glibenclamide, both to remove the risk of hypoglycaemia
and to allow insulin levels to fall, producing mild ketosis to help control hunger.
She was asked to have serum electrolytes, urea, creatinine and uric acid levels
measured 6 weeks into the diet. She saw the dietitian fortnightly.

3-month review: After 3 months of the diet she had lost 17 kg, and blood
glucose levels had improved dramatically, with glycosylated haemoglobin
dropping to 7.9%.

Management: She was advised to phase out the VLED over the next 2 months
and to adopt a low-fat, low-carbohydrate diet. However, as previous lost weight
had always been regained, sibutramine (10 mg, mornings) was prescribed to
assist weight maintenance. She reported it had a noticeable effect on hunger
and satiety, but she was struggling financially because of its cost ($120 per
month).

Treatment and prevention of obesityare there


critical periods for intervention?

Next Section

Why obesity?

Both professionals and the public view obesity, increasingly apparent in childhood, and
already highly prevalent in adults in the Western world, as one of, if not, the most
important public health problem of our times. The considerable effort expended on
researching risk factors for obesity (a Medline search for studies examining risk factors
for obesity conducted at the time of writing this editorial (November 2005) resulted in
264 326 hits) contrasts starkly with the simplicity of the key underlying problem, that
obesity is largely a consequence of over-nutrition and under-activity. Despite the clarity
of this message, there is little evidence-based guidance on successful, viable long-term
strategies to prevent or treat obesity. We believe there is a need to develop findings
from epidemiological research into coherent decisions regarding prevention and
treatment interventions and ultimately appropriate polices for the improvement of
public health. Our intention was that a themed issue on obesity in the International
Journal of Epidemiology would contribute towards this aim.
In the first half of this editorial we review the current evidence for the treatment of adult
obesity and conclude that to date there is no strong evidence that such treatments have
long-term benefits in terms of health gain. Clearly, lack of evidence does not equate to
lack of effect and there is no doubt that most trials to date have not been large enough
or had sufficiently long-term follow-up to answer these questions. On the other hand
treating established obesity in adulthood may be shutting the gate after the horse has
bolted. Further, epidemiology tells us that obesity is socially patterned, varies between
countries, but in recent years has shown marked increases in all countries, and that
what we eat and the exercise we take is largely determined by the food industry,
transport policy, and the built environment (see for example the piece by Cummins and
Macintyre in this issue1). Thus, a population approach to the primary prevention of
obesity and to the prevention of its associated diseases is more likely to be beneficial
than an individual or small group level approach such as treating established obesity.

Animal studies suggest that brief interventions during critical or sensitive periods of
development can have lasting effects in terms of disease prevention. This seems such
an exciting prospect to us that we spend the second half of this editorial considering
whether there is sufficient evidence relating critical/sensitive periods of development to
the risk of later obesity and its associated diseases to warrant trials in humans of such
interventions.

Previous SectionNext Section

Treating obesity

Ideally, any treatment for obesity should assess long-term impacts on obesity
associated cardiovascular risk factors, such as hypertension, dyslipidaemia, and insulin
resistance, as well as disease outcomes such as cardiovascular disease, diabetes, and
osteoporosis. However, to date, few studies of any intervention have been sufficiently
powered and sufficiently long-term to go beyond the assessment of weight loss itself.

Previous SectionNext Section

Diets

Several systematic reviews and meta-analyses have examined the effect of a variety of
dietary interventions on weight reduction in adults with obesity. Very low energy density
diets (<800 kcal/day) resulted in the greatest weight loss, 1525% of initial weight, over a
short period, in those who completed the programme.2,3 However, the authors noted that these
programmes were associated with high financial cost, high attrition rates and high odds
of regaining 50% or more of the lost weight over 1224 months of follow-up. With the
exception of weight-watchers (a weekly support group activity in the UK), for which
three randomized controlled trials suggested moderate weight loss (up to 3% of original
weight) over 2 years of follow-up, trials of self-help programmes and programmes
available over the internet do not suggest benefits in terms of weight loss or other
outcomes.3
In the long-term only low fat diets have been found to be beneficial, with a pooled
weight loss of 3.55kg (95% CI 4.54 to 2.55kg) at 36 months compared with control
groups.4 There were also long-term beneficial effects on dyslipidaemia and blood
pressure for low fat diets in obese individuals. The long-term effects of other diets,
including very low calorie diets were unclear, and the authors concluded that large long-
term randomized trials of different dietary regimes with disease outcomes were required
to determine the true health benefit of any diet.4

Previous SectionNext Section

Over-the-counter dietary supplements

As well as the numerous fad-diets that are available to the public there is a burgeoning
market in over-the-counter remedies that make extravagant claims about their weight
loss potential. Max Pittler and Edzard Ernst recently undertook a review in order to
assess the evidence from rigorous clinical trials, systematic reviews and meta-analyses
on the effectiveness of dietary supplements.5 They identified five systematic
reviews/meta-analyses and an additional 25 trials, which were not included in any
previous review. None of these studies provided convincing evidence to support the use
of supplements, with one exception. Ephedra sinica (also known as ma-huang) was
associated with modest short-term weight loss (in the order of 0.9 kg /month) when
compared with placebo. However, this is an ephedrine-containing supplement and has
been found to be associated with a 2-fold to 4-fold increase in the odds of psychiatric,
autonomic or gastrointestinal symptoms, and heart palpitations.5 Unsurprisingly, Pittler
and Ernst concluded that none of the reviewed supplements could be recommended for
over-the-counter use.5

Previous SectionNext Section

Pharmocological treatment

Over recent decades there has been increased interest in the use of drug treatment for
obesity. Two anti-obesity drugsorlistat and sibutraminehave been widely assessed in
a number of randomized controlled trials.

Orlistat, a pancreatic lipase inhibitor that reduces the absorption of dietary fat, is
effective both in the short-term and long-term at reducing weight, particularly when
combined with dietary and exercise interventions.6,7 The combination of orlistat and
physical activity advice has been found to have long-term (1824 months) beneficial
effects on dyslipidaemia and hypertension, though it is unclear whether this effect is
primarily due to the drug or the increase in physical activity.6,7 To date there are no
results of the long-term effects of orlistat on disease endpoints or disease and disability
free survival.6,7 Orlistat is associated with a higher incidence of gastrointestinal adverse
events, which have in some trials resulted in poor compliance. Further, there is some
evidence that the effect of orlistat, in terms of weight loss and improvements in
cardiovascular disease risk factors, is weaker for obese individuals who are also diabetic
than for non-diabetic obese individuals.6

Sibutramine is a centrally acting serotoninnorepinephrine reuptake inhibitor that


enhances satiety and promotes energy expenditure. Short-term trials, and one trial over
2 years of follow-up, demonstrate sustained effects on weight loss.8 Beneficial effects of
sibutramine on triglyceride, high-density lipoprotein cholesterol, and glycaemic control
have also been reported, but there is no direct evidence that sibutramine reduces
obesity-associated morbidity and mortality. Because of its norepinephrine effect it has
been anticipated that sibutramine could increase blood pressure. This hypothesis is
supported by some, though not all trials.8,9Nevertheless, it is not recommended for use
in obese individuals with hypertension, which, given the concordance between obesity
and hypertension, somewhat limits its usefulness.

Many new pharmacological approaches are under investigation. These include gut
hormones, such as cholecystokinin that normally signal satiety, other centrally acting
serotonin agents, the anticonvulsant medications topiramate and zonisamide,
cannabinoid receptor antagonists, and drugs that act on other peptide
neurotransmitters. The first randomized controlled trial (n = 1507) in humans of a
selective cannabinoid-1 receptor antagonist (rimonabant) was recently published. 10 It
found a marked reduction in weight and waist circumference and improvements in high
density lipoprotein cholesterol, triglycerides, insulin resistance, and the prevalence of
the metabolic syndrome at 1 year of follow-up when rimonabant was given at a dose of
20 mg per day, but much weaker effects on weight reduction of a 5 mg dose and no
effects on metabolic syndrome components at this lower dose.10 Despite the marked
weight loss with the higher dose there was no effect on blood pressure, total cholesterol,
or low density lipoprotein cholesterol. Further, there is evidence from animal studies
that low dose cannabinoid therapy reduces progression of atherosclerosis,11 and it has,
therefore, been suggested that blocking cannabinoid receptors might actually increase
the risk of atherosclerosis.12 There is biological evidence that blocking cannabinoid
receptors might result in demyelination and one participant treated with rimonabant in
this trial developed multiple sclerosis.13 This may have been a chance occurrence, but
there is clearly a need to aggressively investigate potential side-effects of all new drugs.

Previous SectionNext Section

Surgery

In most populations surgery is reserved for morbid obesity [Body mass index (BMI) of 35
kg/m2 together with obesity related morbidity or BMI of 40 kg/m2 in the absence of
associated morbidity] and is considered when other non-surgical treatments have failed,
though there is evidence of extensive use of surgery to reduce weight outside of these
criteria in privately funded health care systems. A systematic review identified 26
studies, of which just 5 (2 randomized controlled trials and 3 cohort studies) compared
surgery with non-surgical management, with the remaining 21 (all randomized
controlled trials) comparing the effectiveness of different surgical procedures to each
other.14 The quality of the studies was noted to be generally poor, with just 3 of the 24
trials having adequate allocation concealment. The authors concluded that the limited
evidence suggests that surgery is more effective than conventional management for
weight loss in morbid obesity. The comparative safety and effectiveness of different
surgical procedures is unclear.14

Previous SectionNext Section

A population approach to the primary prevention of obesity and its related morbidity and mortality

Obesity is no longer a health problem confined to adults. The prevalence of childhood


obesity has increased 3-fold in most industrialized countries over the last 20 years.15 In the US,
often perceived as an extreme example, around a quarter of all children are overweight
or at risk of being overweight.16 The rest of the developed world is, however, not far
behind.17 Obese children often become obese adults. Childhood obesity increased the
risk of adult obesity 4-fold in men and 3.2-fold in women in the British 1958 birth cohort,
although child to adult BMI correlations across the range were modest.18 Among
contemporary children and adolescents obesity is associated with elevated blood
pressure, dyslipidaemia, glucose intolerance, hyperinsulinaemia, and greater left
ventricular mass,1923 though the evidence linking childhood overweight/obesity with
adult cardiovascular disease events is weak, perhaps in part because these studies are
based on individuals who were born several decades ago at times when childhood
obesity was less common.24,25 Increasingly, frank type 2 diabetes is being diagnosed in
obese adolescents.26,27 Further, obesity and its associated cardiovascular disease risk
factors are associated with atheroscelorosis in autopsy studies of adolescents and
young adults.28 Thus, there is evidence that obesity in contemporary children and
adolescents has already resulted in metabolic and vascular abnormalities that may be
long-lasting. As a consequence attempts to treat established obesity in adulthood may
be too late to have important impacts on disease prevention or health improvement.

Previous SectionNext Section

Standard approaches to obesity prevention in children and adolescents

One review of interventions to treat or prevent obesity in children found that, as with
adult obesity, there was no strong evidence that interventions to treat obesity in
children had long-term benefits in terms of weight loss or associated morbidity. 15 The
authors asked Why is substantial long-term weight loss so difficult to obtain? 15 They
concluded that increasing funds were required for research into new behavioural,
environmental, and pharmacological approaches for the prevention and treatment of
obesity in children, but emphasized that the epidemic of childhood obesity was unlikely
to be resolved without concerted political action to detoxify the obesogenic environment
in which we live.

A recent Cochrane systematic review identified just 22 controlled (with or without


randomization) trials of interventions in childhood and adolescence to prevent
obesity.29 Most were school based and most assessed outcomes over a short time
period only. Important methodological weaknesses were noted in many studies, and in
particular the authors noted that many of the studies included in this review have
unit of allocation errors, since allocation was often by institution (e.g. school) but
assessment was by individual child. The results of these studies are likely to be
misleadingly optimistic.29 Even with this caveat regarding their possible exaggeration
of true effects, most studies found that combined promotion of healthy eating and
physical activity were not effective at preventing childhood obesity. The impact of these
interventions on the adverse sequelae of obesity, such as glucose intolerance,
hypertension, and dyslipidaemia, were rarely assessed. While better designed studies of
these interventions may provide evidence of effectiveness in terms of both weight
control and metabolic and disease outcomes we believe there is also merit in exploring
whether brief interventions during key periods of development might have long-term
benefits in terms of obesity and obesity related disease prevention.

Previous SectionNext Section

Critical and sensitive periods for the primary prevention of obesity

Three periods in early life may be particularly important for the development of obesity
and its associated morbidity and mortality: the perinatal period; the period of adiposity
rebound; and puberty/adolescence.30,31

Previous SectionNext Section

The perinatal period

There is increasing evidence that intrauterine over-nutrition predicts life long


obesity.32,33 According to this hypothesis high maternal glucose, free fatty acid, and
amino acid plasma concentrations result in over-nutrition of the fetus which, through
permanent changes in appetite control, neuroendocrine functioning, or energy
metabolism in the developing fetus, leads to obesity in later life.32,33 Since maternal
obesity itself is associated with insulin resistance and glucose intolerance, and,
therefore, higher plasma concentrations of glucose and free fatty acids, maternal
obesity is seen as the prime factor in fetal over-nutrition. Recent evidence supports this
hypothesis, with two studies demonstrating a relationship between greater weight gain
during pregnancy and obesity in the offspring at 24 years. 34,35 The consequences of
these finding are potentially formidable: the obesity epidemic could accelerate through
successive generations independent of further genetic or environmental factors.36 The
mechanisms of such an association between maternal weight and weight gain during
pregnancy and obesity in her offspring are becoming clearer. Offspring of female rats
with diet-induced obesity have been found to be heavier than the offspring of rats with
the same genotype, but without the diet-induced maternal obesity. 37 In vitro, animal
and human studies have demonstrated that fetal pancreatic development and fat stores
are influenced by the availability of fetal fuelsin particular glucose, lipids, and amino-
acidswhich are in turn determined by maternal insulin secretion and responsiveness,
and maternal plasma levels of glucose and free fatty acids.33 These in vitrofindings are
confirmed by studies of women with gestational diabetes whose offspring have
considerably greater birth weights and a greater risk of obesity and diabetes in later
life.33,38 There is also evidence that these adverse sequelae are not confined to
maternal diabetes; rather there is a linear trend of increasing offspring birth weight with
increasing maternal gestational glucose concentration across the population
distribution.39

The long-term follow-up of the offspring of mothers who have been involved in
randomized trials of the effectiveness of strict glycaemic control during pregnancy will
provide particularly valuable insights into the potential of intervening during this period
to improve outcomes in the offspring. In the short term, improved perinatal outcomes
have been observed amongst those women with gestational diabetes randomized to
intensive glycaemic control vs those on standard care.40 There were fewer large for
gestational age infants amongst those in the intervention group (13% vs 22%, P <
0.001) and fewer infants with macrosomia (10% vs 21%, P< 0.001). However, these
differences may have been largely driven by the shorter period of gestation among the
intensively treated group, due mainly to the greater rate of inductions of labour in that
group. Nevertheless, long-term follow-up of these infants to determine whether a brief
intervention during the intrauterine period has long-term beneficial effects on the
offspring in terms of the development of obesity and its associated diseases is
important for testing the fetal overnutrition hypothesis and determining whether a brief
intervention during the intrauterine period among this high-risk group has a lasting
effect.

Previous SectionNext Section

Infancy

In normal physiological circumstances, during the first year of life BMI increases rapidly,
but then decreases, reaching a minimum usually 56 years of age. This point of minimum BMI
has been called the adiposity rebound, though would be more accurately termed BMI
rebound. Following this nadir, BMI then starts to increase again. Several studies have
found that an earlier adiposity rebound (based on the assessment of BMI) increases
the risk of later obesity.4143 However, the meaning and usefulness of these findings are
unclear. BMI is not a true measure of adiposity and other markers of adiposity do not
show the same patterns as BMI over early life. Thus, ponderal index (kg/m3) and
percentage body fat both decrease to about age 6 years and remain constant
thereafter, whereas triceps skinfold thickness shows two nadirs (at ages 68 and 1517
years).44 It has also been demonstrated that early age at adiposity rebound predicts
later fatness as it identifies children whose BMI centile is high and/or moving upwards
across centiles, suggesting that BMI centile crossing or actual BMI in childhood is a more
useful measure for predicting later fatness than is age at adiposity rebound. 45 This is
consistent with findings from the Bogalusa study, which, although finding an association
between early age at adiposity rebound and adult BMI, also noted that BMI at age 78
years was a stronger predictor of adult BMI than age at minimum BMI.46 As age at
adiposity rebound can only be determined in retrospect, prevention per se is difficult to
implement and assess.44 One could try to identify modifiable risk factors associated
with early adiposity rebound, but in one study that aimed to do just that the only
independent predictor of early adiposity rebound was parental obesity, which is a known
risk factor for offspring obesity.47

Several studies have found that breast-feeding is protective against later adiposity, but
a recent systematic review and meta-analysis concluded that while mean BMI in later
life was lower among breast-fed subjects, the difference was small and likely to have
been strongly influenced by publication bias and confounding factors. 48 These
conclusions are supported by findings from a large cluster randomized controlled trial of
the promotion of breast-feeding,49 which failed to show marked differences in obesity or
cardiovascular disease risk factors in later childhood. Thus, evidence to date does not
support infancy as a critical period during which interventions might have long-term
effects on the risk of obesity and its associated diseases.

Previous SectionNext Section

Puberty and adolescence

Puberty is a time of rapid change in size and shape for both females and males. In
females earlier age at menarche is associated with obesity, independently of childhood
BMI and other potential confounding factors.50 Puberty is associated with a
physiological increase in insulin resistance51 and is thought to contribute to a peak of
incidence in type 1 diabetes at that age. In relation to these changes in insulin
metabolism, post-pubertal fat deposition in both females and males tends to be more
central rather than general. In addition, behaviours such as dietary patterns and levels
of physical activity are largely formed in adolescence and persist into
adulthood.44 Adolescence may also be a critical period for the development of
atherosclerosis. Lipid rich deposits (fatty streaks) are found in the aortae of almost all
children >3 years of age, irrespective of ethnicity, sex, environment, diet, or later
CHD.28 Consequently, while these lesions may be the seed for atherosclerosis, their
relationship to extent of adult atherosclerosis is disputed. Autopsy studies in humans
show that late adolescence (i.e. from 1519 years) is the key time when fatty streaks convert to
raised atherosclerotic lesions.28 Intriguingly, this also corresponds to the age at which BMI
and skinfold thickness increase in young adults who go on to develop the metabolic
syndrome,52suggesting that the development of nascent metabolic syndrome may be
linked to the generation of raised atherosclerotic plaques. By the age of 30, raised
atherosclerotic lesions are present in arteries of 1 in 3 adults and are associated with
the same risk factors (central adiposity, dyslipidaemia, hypertension, glucose
intolerance/insulin resistance, chronic inflammation)28 that predict diabetes and
cardiovascular disease. These data suggest that adolescence offers a therapeutic
window with a unique opportunity to modify the risk of future obesity, diabetes, and
cardiovascular disease and achieve long-term prevention, perhaps via short-term
interventions.
Previous SectionNext Section

Pregnancy

The role of pregnancy in determining offspring obesity has already been discussed. But
in addition, there is increasing evidence that weight gain during pregnancy, and post-
partum weight retention, may be an important predictor of the mothers' risk of
subsequent obesity and diabetes.53 It has been proposed that the antenatal period
therefore offers a unique period in the life course during which women at risk of future
diabetes and cardiovascular disease might be identified, at a time when they might be
particularly receptive to health promotion or disease prevention interventions. 53

Previous SectionNext Section

Studies of interventions during key developmental periods: proof of concept

The idea that an intervention during a key developmental period can persistently modify
risk factors is supported by studies in animal models of human disease. Brief treatment
with angiotensin-converting enzyme (ACE) inhibitors54 or a selective angiotensin II
receptor antagonist55 in young (prior to their development of hypertension) genetically
spontaneous hypertensive rats causes a reduction in blood pressure that persists
throughout life and is associated with a reduction in target organ damage. In humans,
as a result of these findings in animal studies, there are now two on-going trials
investigating whether treatment of pre-hypertension with an angiotensin receptor
antagonist for a brief period only in young adults (average age 35 years) may delay or
prevent subsequent hypertension: the Trial of Prevention of Hypertension (TROPHY) and the
Danish Hypertension Project.56 Similarly, animal models of type 1 diabetes indicate that
intensive prophylactic treatment from weaning to 180 days of life in genetically
programmed diabetic mice is effective at reducing the risk of development of
diabetes.57 In a non-randomized controlled study of non-diabetic school children who
had islet cell antibodies (and thus increased risk of type 1 diabetes) brief treatment in
childhood with nicotinomide reduced the risk of future diabetes.58 Similarly, a small trial
of prophylaxis with insulin therapy among non-diabetic children with relatives who had
type 1 diabetes produced promising results.59 However, larger randomized trials of
these agents have been negative.60,61 Nevertheless, given the difficulty of establishing
the correct therapeutic window, duration of therapy, dose, and agent, these
disappointing findings should not curtail attempts to pursue this approach in this and
other disease areas.

Previous SectionNext Section

The future

Data on childhood obesity from the developing world are sparse, but indicate that not
only is obesity on the increase but also that obesity co-exists with the long-standing
problem of under-nutrition (see, for example, the paper by Andrew Prentice 62 in this
themed issue). The impact of interactions between these conditions is not known and
difficult to predict. In Asian Indians, long-term adaptation to scarce food supplies has, in
times of abundance, resulted in a classically insulin resistant population, with central
obesity, dyslipidaemia, glucose intolerance, and cardiovascular disease. Yet, perhaps as
a consequence of persistent maternal malnutrition (both under-provision of critical
nutrients, and overprovision of obesogenic foods), Indian Asian babies are both short
and thin, and are already more glucose intolerant, insulin resistant, dyslipidaemic, and,
importantly, have a greater percentage of body fat, than their European
counterparts.63,64 That this population has, in settings of food abundance, one of the
highest rates of diabetes and cardiovascular disease in the world, suggests that the
intergenerational effect of over-nutrition superimposed on under-nutrition may be
particularly toxic. This observation underlines the fact that we must be cautious when
extrapolating findings from studies performed largely in Western settings to the
developing world, where triggers for obesity and their outcomes may be very different.
In addition, in many developing countries, obesity in women is particularly prized as a
sign of affluence, and is often achieved at the cost of relative malnutrition for other
members of the family. Given the suggested vicious spiral between obesity during
pregnancy and childhood obesity, this has potentially dire implications for the likely
future patterns of obesity in these countries.

Standard approaches to obesity prevention in the long term have been disappointing.
Targeting the prevention of obesity during the key periods of development may be of
particular relevance in reducing subsequent risks of adult obesity and associated
chronic disease. To our knowledge there are no trials in humans that have examined the
long-term effects of maternal glycaemic control during pregnancy on the risks of obesity
and associated morbidity in their offspring (to provide causal evidence for the fetal
overnutrition hypothesis and to provide evidence on the possible prolonged and long-
term benefit of a brief intervention during a critical period of future health) and beyond
the trials of school-based health promotion interventions (described above) we are not
aware of other trials in adolescence (a possible critical period for the development of
obesity, metabolic disorders, and atherosclerosis) that have assessed brief interventions
aimed at permanent beneficial effects on obesity and cardiovascular risk factors. It is
possible that most investigators feel that the epidemiological evidence is still not
sufficiently robust to proceed with such trials. But we feel that the animal studies in
hypertension and diabetes discussed above and the progress from these to undertaking
trials of brief interventions in young adults offer exciting prospects for the future.
Perhaps if we undertake another themed issue of the journal in 10 years time we will be
able to report on the benefits of a brief intervention in a critical period of human
development on future obesity and health risk.