"Obesity As a Global Problem". In: International Textbook of Obesity

MMMM Part I

Obesity as a Global Problem

Vicki J. Antipatis and Tim P. Gill Rowett Research Institute, Aberdeen, UK

INTRODUCTION WHAT IS OBESITY AND HOW IS IT MEASURED? Obesity is a major public health and economic problem of global significance. Prevalence rates are At the physiological level, obesity can be defined as increasing in all parts of the world, both in affluent a condition of abnormal or excessive fat accumula- Western countries and in poorer nations. Men, tion in adipose tissue to the extent that health may women and children are affected. Indeed, over- be impaired. However, it is difficult to measure weight, obesity and health problems associated body fat directly and so surrogate measures such as with them are now so common that they are replac- the body mass index (BMI) are commonly used to ing the more traditional public health concerns indicate overweight and obesity in adults. Addi- such as undernutrition and infectious disease as the tional tools are available for identification of indi- most significant contributors to global ill health (1). viduals with increased health risks due to ‘central’ In 1995, the excess adult mortality attributable to fat distribution, and for the more detailed charac- overnutrition was estimated to be about 1 million terization of excess fat in special clinical situations deaths, double the 0.5 million attributable to under- and research. nution (2). This chapter looks at obesity as a global problem. It begins with a brief overview of methods of classi- Measuring General Obesity fication, a critical issue for estimating the extent of obesity in populations. The serious impact of excess The BMI provides the most useful and practical body weight on individuals and societies through- population-level indicator of overweight and obes- out the world in terms of associated health, social ity in adults. It is calculated by dividing body- and economic costs is considered next. The body of weight in kilograms by height in metres squared the chapter concentrates on current prevalence and (BMI : kg/m). Both height and weight are trends of adult obesity rates around the world, in- routinely collected in clinical and population health cluding projections for the year 2025. Comment is surveys. made on key features and patterns of the global In the new graded classification system develop- epidemic followed by discussion of the major fac- ed by the World Health Organization (WHO), a tors that are driving it. An overview of the emerging BMI of 30 kg/m or above denotes obesity (Table childhood obesity problem is given next. The chap- 1.1). There is a high likelihood that individuals with ter concludes with a call for global action to tackle a BMI at or above this level will have excessive the epidemic. body fat. However, the health risks associated with overweight and obesity appear to rise progressively

Table 1.1 Classification of overweight and obesity in adults Table 1.2 Sex-specific waist circumference measurements for according to BMI identification of individuals at increased health risk due to intra-abdominal fat accumulation Classification BMI (kg/m) Waist circumference Underweight :18.5 (cm) Normal range 18.5—24.9 Risk of metabolic Overweight P25 complications Men Women Pre-obese 25.0—29.9 Obese class I 30.0—34.9 Alerting zone Increased 94 80 Obese class II 35—39.9 Action zone Substantially increased 102 88 Obese class III P40 Adapted from WHO (1). Source: WHO (1). with increasing BMI from a value below 25 kg/m, believe that a health risk classification based on and it has been demonstrated that there are benefits waist circumference alone is more suitable as a to having a measurement nearer 20—22 kg/m,at health promotion tool than either BMI or waist-to- least within industrialized countries. To highlight hip ratio, alone or in combination (4). Recent work the health risks that can exist at BMI values below from the Netherlands has indicated that a waist the level of obesity, and to raise awareness of the circumference greater than 102 cm in men, and need to prevent further weight gain beyond this greater than 88 cm in women, is associated with a level, the first category of overweight included in the substantially increased risk of obesity-related meta- new WHO classification system is termed ‘pre- bolic complications (Table 1.2). The level of health obese’ (BMI 25—29.9 kg/m). risk associated with a particular waist circumfer- Caution is required when interpreting BMI ence or waist-to-hip ratio may vary across popula- measurements in certain individuals and ethnic tions. groups. The relationship between BMI and body fat content varies according to body build and body proportion, and a given BMI may not correspond THE HEALTH, SOCIAL AND to the same degree of fatness across all populations. ECONOMIC COSTS ASSOCIATED Recently, a meta-analysis among different ethnic groups showed that for the same level of body fat, WITH OVERWEIGHT AND OBESITY age and gender, American blacks have a 1.3 kg/m higher BMI and Polynesians have a 4.5 kg/m high- There is reason to be concerned about overweight er BMI compared to Caucasians. By contrast, and obesity as overwhelming evidence links both to BMIs in Chinese, Ethiopians, Indonesians and substantial health, social and economic costs. Thais were shown to be 1.9, 4.6, 3.2 and 2.9 kg/m lower than in Caucasians (3). This suggests that population-specific BMI cut-off points for obesity Overview of the Health Costs need to be developed. US figures suggest that about 61% of non-insulin- dependent diabetes mellitus (NIDDM) and 17% of both coronary heart disease (CHD) and hyperten- Measuring Central Obesity sion can be attributed to obesity. Indeed, as a per- son’s BMI creeps up through overweight into the For a comprehensive estimate of weight-related obese category and beyond, the risk of developing a health risk it is also desirable to assess the extent of number of chronic non-communicable diseases intra-abdominal or ‘central’ fat accumulation. This such as NIDDM, CHD, gallbladder disease, and can be done by simple and convenient measures certain types of cancer increases rapidly. There is such as the waist circumference or waist-to-hip also a graded increase in relative risk of premature ratio. Changes in these measures tend to reflect death (Figure 1.1). changes in risk factors for cardiovascular disease Before life-threatening chronic disease develops, and other forms of chronic illness. Some experts however, many overweight and obese patients de- OBESITY AS A GLOBAL PROBLEM 5 between intra-abdominal and subcutaneous adipose tissues which makes the former more susceptible to both hormonal stimulation and changes in lipid metabolism. People of Asian descent who live in urban societies are particularly susceptible to central obesity and tend to develop NIDDM and CHD at lower levels of overweight than other populations.

Figure 1.1 The relationship between risk of premature death Overview of the Economic Costs and BMI. The figure is based on data from professional, white US women who have never smoked and illustrates the graded Conservative estimates clearly indicate that obesity increase in relative risk of premature death as BMI increases. represents one of the largest costs in national health Adapted from WHO (1) care budgets, accounting for up to 6% of total expenditure in several developed countries (Table Table 1.3 Relative risk of health problems associated with obesity 1.4). In the USA in 1995, for example, the overall direct costs attributed to obesity (through hospital- Greatly increased Moderately izations, outpatients, medications and allied health (relative risk much increased Slightly increased professionals’ costs) were approximately the same greater than 3) (relative risk 2—3) (relative risk 1—2) as those of diabetes, 1.25 times greater than those of NIDDM CDH Certain cancers coronary heart disease, and 2.7 times greater than Gallbladder disease Hypertension Reproductive those of hypertension (5). The costs associated with hormone pre-obesity (BMI 25—30 kg/m) are also substantial abnormalities because of the large proportion of individuals in- Dyslipidaemia Osteoarthritis Polycystic ovary (knees) syndrome volved. Insulin resistance Hyperuricaemia Impaired fertility The economic impact of overweight and obesity and gout does not only relate to the direct cost of treatment in Breathlessness Low back pain due the formal health care system. It is also important to to obesity consider the cost to the individual in terms of ill Sleep apnoea Increased anaesthetic risk health and reduced quality of life (intangible costs), Fetal defects arising and the cost to the rest of society in terms of lost from maternal productivity due to sick leave and premature dis- obesity ability pensions (indirect costs). Overweight and obesity are responsible for a considerable propor- Source: WHO (1). tion of both. Thus, the cost of lost productivity attributed to obesity in the USA in 1994 was $3.9 velop at least one of a range of debilitating condi- billion and reflected 39.2 million days of lost work. tions which can drastically reduce quality of life. In addition, there were 239 million restricted-activ- These include musculoskeletal disorders, respir- ity days, 89.5 million bed-days, and 62.6 million atory difficulties, skin problems and infertility, physician visits. which are often costly in terms of absence from Estimates of the economic impact of overweight work and use of health resources. Table 1.3 lists the and obesity in less developed countries are not health problems that are most commonly asso- available. However, the relative costs of treatment if ciated with overweight and obesity. In developed available are likely to exceed those in more affluent countries, excessive body weight is also frequently countries for a number of reasons. These include the associated with psychosocial problems. accompanying rise in coronary heart disease and The risk of developing metabolic complications is other non-communicable diseases, the need to im- exaggerated in people who have central obesity. port expensive technology with scarce foreign ex- This is related to a number of structural differences change, and the need to provide specialist training 6 INTERNATIONAL TEXTBOOK OF OBESITY Table 1.4 Conservative estimates of the direct economic costs of obesity

Country Year Obesity deﬁnition Estimated direct costs % National health care costs

USA 1995 BMI P30 US$52 billion 5.7 Australia 1989/90 BMI 930 AUD$464 million 92 Netherlands 1981—89 BMI 925 Guilders 1 billion 4 France 1992 BMI P27 FF 12 billion 2

Table 1.5 Estimated world prevalence of obesity

Population aged P15 Prevalence of Approximate estimate (mid-point) of years (millions) obesity (%) number of obese individuals (millions)

Established market economies 640 15—20 96—128 (112) Former socialist economies 330 20—25 66—83 (75) India 535 0.5—1.0 3—7 (5) China 825 0.5—1.0 4—8 (6) Other Asian countries and Islands 430 1—34—12 (8) Sub-Saharan Africa 276 0.5—1.0 1—3 (2) Latin America and Caribbean 280 5—10 14—28 (21) Middle East 300 5—10 15—30 (22) World 3616 (251)

Source: Seidell (4). for health professionals. As many countries are still limited availability of suitable data for an accurate struggling with undernutrition and infectious dis- assessment of obesity prevalence and trends in dif- ease, the escalation of obesity and related health ferent countries. Although it is half a century since problems creates a double economic burden. obesity was introduced into the International Clas- sification of Diseases (ICD), overweight and obesity are rarely recognized by health professionals as a THE GLOBAL OBESITY PROBLEM distinct disease or cause of death, and so are infre- quently recorded on morbidity or mortality statis- The number of people worldwide with a BMI of 30 tics. This means that we have to rely on BMI data or above is currently thought to exceed 250 million, collected as part of specific health screening surveys i.e. 7% of the world’s adult population (Table 1.5) or scientific studies. Unfortunately, very few coun- (4). When individual countries are considered, the tries conduct national surveys on a regular basis, range of obesity prevalence covers almost the full and even fewer report obesity prevalence. This re- spectrum, from below 5% in China, Japan and flects the fact that most national nutrition surveys, at least in developing countries, are still used to certain African nations to more than 75% in urban provide information about undernutrition in Samoa. It is difficult to calculate an exact global figure because good quality and comparable data women and young children. The costs and re- are not widely available. The assessment in Table sources required to conduct regular comprehensive 1.5 is a conservative estimate. national surveys are a major barrier to implementation. The second issue relates to the need for caution when making comparisons of obesity rates between Important Issues Associated with Data studies and countries. Comparison is complicated Collation by a number of factors including differences in obesity classification systems, mismatched age groups, Discussion and comparison of overweight and inconsistent age-standardization of study popula- obesity rates throughout the world are complicated tions, discordant time periods and dates of data by a number of important issues associated with collection, and use of unreliable self-reported data collation. The first of these relates to the weight and height measurements for calculation of OBESITY AS A GLOBAL PROBLEM 7 BMI. In particular, the use of BMI cut-off points countries. Recent data from the Russian Longitudi- either above or below 30 kg/m to denote obesity nal Monitoring Survey indicate that Russia has a has a great impact on estimates of obesity preva- particularly serious obesity problem, especially lence in a given population. In the US, obesity has among women where 28% of the population was until very recently been routinely classified as a obese in 1996. Results from the Italian National BMI at or above 27.8 kg/m in men and 27.3 kg/m Health Survey indicate that Italy has one of the in women. With these cut-off points, 31.7% of men lowest levels of obesity in Europe. However, the and 34.9% of women were deemed obese in the Italian data may be underestimated due to self- period 1988—1994. These estimates fall to 19.9% of reporting of weight and height measurements. men and 24.9% of women when a BMI of 30 kg/m National figures for North America are similar to is applied. Projects such as the WHO MONICA those of Europe, with approximately 20% of males (MONItoring of trends and determinants in CAr- and 25% of females currently obese in the USA, and diovascular diseases) study (see below), where data 15% of all adults obese in Canada. Rates in the are collected from a large number of populations in general populations of Australia and New Zealand the same time periods according to identical proto- are also in the range of 15—18%. Japan, at less than cols, are particularly valuable for comparison pur- 3%, still has a very low level of obesity for an poses. industrialized country. A third issue is the need to be aware that many In the oil-exporting countries of the Middle East, countries such as Brazil and Mexico show great the adult populations appear to have a major obes- variation in wealth by region. Combining data from ity problem. Women in particular are affected, with all areas into a single country figure, or from a prevalence several fold higher than that reported for number of countries into a regional figure, is likely many industrialized countries. Bahrain (urban), to mask patterns of relationships between social Kuwait, Jordan, Saudi Arabia (urban), and the variables and obesity. United Arab Emirates all document female obesity rates well above 25%. The highest obesity rates in the world are found in the Pacific Island populations of Melanesia, Current Prevalence of Obesity Polynesia and Micronesia. In urban Samoa, for example, approximately 75% of women and 60% of Despite the limited availability and fragmentary men were classified as obese in 1991. These figures nature of suitable country-level data, it is clear that correspond with some of the highest rates in the obesity rates are already high and increasing rapid- world of diabetes and other related chronic dis- ly in all regions of the world. Table 1.6 shows the eases. With regard to obesity, it should be noted most current estimates of obesity prevalence, ac- that the prevalence figures may be slightly exag- cording to a BMI of 30 or greater, in a selection of gerated because Polynesians are generally leaner countries from around the globe. Nationally repre- than Caucasians at any given BMI. sentative data sets based on measured weight and From a nutrition perspective, research and policy height are presented where possible. in many Asian and lower-income countries have Examination of Table 1.6 reveals large variations focused on undernutrition. However, there are clear in obesity prevalence between countries, both with- indications that a number of these countries are in and between regions. In Africa, for example, now beginning, or are already experiencing, high obesity rates are extremely high among women of levels of overweight and obesity. Urban China, ur- the Cape Peninsula but very low among women in ban Thailand, Malaysia and the Central Asian Tanzania. countries that were members of the Societ Union Much of the developed world already has excep- before 1992 (such as Kyrgyzstan) are all examples. tionally high levels of overweight and obesity. In Overweight is also becoming a serious problem in Europe, obesity prevalence now ranges from about urban India, most notable in the upper-middle 6 to 20% in men and from 6 to 30% in women. class. The situation in China and India is further Rates are highest in the East (e.g. Russia, former complicated by the fact that chronic energy defi- East Germany and Czech Republic) and lowest in ciency is still a major problem for large parts of the some of the Central European and Mediterranean population. 8 INTERNATIONAL TEXTBOOK OF OBESITY Table 1.6 Prevalence of obesity (BMI P 30 kg/m) in a selection of countries

Prevalence of obesity (%)?

Country Year Age Men Women

Europe Finland 1991/93 20—75 14 11 Netherlands 1995 20—59 8.4 8.3 UK England 1997 16—64 17 20 Scotland 1995 16—64 16 17 ?Italy 1994 15; 6.5 6.3 France 1997? 15; 8.6 8.4 Czech Republic 1995 20—65 22.6 25.6 former East Germany 1992 25—69 21 27 former West Germany 1990 25—69 17 19 Russia 1996 Adults 10.8 27.9 North America Canada 1991 18—74 15 15 USA 1988—94 20—74 19.9 24.9 Central and Mexico (urban) 1995 Adults 11 23 South America Brazil 1989 25—64 5.9 13.3 Curac¸ao 1993/94 18; 19 36 Middle East Iran, Islamic 1993/94 20—74 2.5 7.7 Republic of (south) Cyprus 1989/90 35—64 19 24 Kuwait 1994 18; 32 44 Jordan (urban) 1994—96 25; 32.7 59.8 Bahrain (urban) 1991/92 20—65 9.5 30.3 Saudi Arabia 1990/93 15; 16 24 Australasia Australia (urban) 1995 25—64 18.0 18.0 and Oceania New Zealand 1989 18—64 10 13 Samoa (urban) 1991 25—69 58.4 76.8 Papua New Guinea 1991 25—69 36.6 54.3 (urban) South and East Japan 1993 20; 1.7 2.7 Asia India (urban Delhi 1997 40—60 3.19 14.28 middle class) China 1992 20—45 1.2 1.64 Malaysia 18—60 4.7 7.9 Singapore@ 1992 Adults 4 6 Kyrgyzstan 1993 18—59 4.2 10.7 Africa Mauritius 1992 25—74 5.3 15.2 Tanzania 1986/89 35—64 0.6 3.6 Rodrigues (Creoles) 1992 25—69 10 31 Cape Peninsula 1990 15—64 7.9 44.4 (Coloured)

?Data are from the Italian National Health Survey and are self-reported. @Obesity criterion: BMI P31 kg/m.

A similar picture is emerging in Central and coloured population of Cape Peninsula and the South America. Mexico and Brazil are already ex- multiethnic island nation of Mauritius. Only the periencing high levels of obesity, especially among very underdeveloped countries of Africa appear to low income and urban populations. Within the Af- be avoiding the worldwide epidemic of obesity, al- rican region too, there are clear pockets where obes- though the lack of good quality data makes it diffi- ity is already a major problem. These include the cult to judge their true weight status. OBESITY AS A GLOBAL PROBLEM 9 of the relative importance of each problem in the Recent Trends population—changed dramatically between 1974 and 1989. This reversed from a ratio of 1.5: 1 (under- Good quality data on trends in body composition weight to overweight) in 1974 to a ratio of less than are even harder to find than cross-sectional data on 0.5: 1 in 1989 (7). prevalence at one point in time, especially for coun- The level of obesity among Chinese adults re- tries outside Europe and the US. Fortunately, na- mains low, but the marked shifts in diet, activity and tionally representative or large nationwide data sets overweight suggest that major increases in over- are now available for a small number of lower and weight and obesity will occur. During the most middle income countries including Brazil, China, recent period of the national China Health and Mauritius, Western Samoa and Russia. Nutrition Survey (CHNS), an ongoing longitudinal The countries of North America and Europe survey of eight provinces in China, data show a have seen startling increases in obesity rates over consistent increase in adult obesity in both urban the last 10—20 years. In Europe, the most dramatic and rural areas. Changes in diet and activity pat- rise has been observed in England, where obesity terns are rapid in urban residents of all incomes but prevalence more than doubled from 6% to 17% in are even more rapid in middle and higher income men and from 8% to 20% in women after 1980. rural residents. Prevalence has increased by about 10—40% over the Few countries seem to have escaped the rapid last 10 years in the majority of other European escalation in obesity rates in the last two decades. countries. The Netherlands, Italy and Finland are rare excep- Obesity rates in the USA have increased from tions where population height and weight data col- 10.4% to 19.9% and from 15.1% to 24.9% in men lected over this period indicate only small increases and women, respectively, over the period or even stabilization of the rates of obesity. 1960—1962 until 1988—1994. The largest increases, however, occurred from the period 1976—1980 on- wards. In Japan, although overall rates of obesity The MONICA Study remain below 3%, prevalence increased by a factor of 2.4 in the adult male population and by a factor The WHO MONICA project provides a compre- of 1.8 in women aged 20—29 years. hensive set of obesity prevalence data from cities Russia has seen a consistent increase in adult and regions. Information was collected in two risk obesity from 8.4% to 10.8% in men and from 23.2% surveys, conducted approximately 5 years apart to 27.9% in women in only 4 years. This is despite from 38 populations. Most surveys were conducted marked shifts toward a lower fat diet in the post- in European cities but there were a few centres in reform period, during which price subsidies of meat North America, Asia and Australasia. Although and dairy products were removed. However, year- they are not national data, they were collected from to-year fluctuations underscore the fact that the over 100 000 randomly selected participants aged economy is in flux and that these changes cannot be 35 to 64 years, are age-standardized and are based used to predict trends. It is also worth noting that on weights and heights measured with identical the prevalence of pre-obesity declined slightly be- protocols. This provides a high level of confidence tween 1992 and 1994 in females but not in males. in the detailed analysis of the data, including com- Trend data from the western Pacific Islands indi- parisons between centres and observations over cate that obesity levels are not only high in these time. Such analysis is rarely possible with less rigor- populations, but that the prevalence of obesity con- ously collected data sets. tinues to increase considerably in each island (6). Analysis of the results from the first round of data Data from two comparable national surveys in collection between 1983 and 1986 showed that the Brazil conducted 15 years apart show that adult average prevalence of obesity among European obesity has increased among all groups of men and centres participating in the study was 15% in men women, especially families of lower income. Nation- and 22% in women, with the lowest in Sweden al figures increased from 3 to 6% in men and from 8 (Go¨ teborg: 7% in men, 9% in women) and the to 13% in women. It is also of interest that the ratio highest in Lithuania (Kaunas: 22% in men, 45% in between underweight and overweight—a measure women). 10 INTERNATIONAL TEXTBOOK OF OBESITY The average age-standardized absolute changes Socioeconomic Status in the prevalence of obesity over 5 years showed that rates increased in three-quarters of the popula- Socioeconomic status (SES) is a complex variable tions for men and in half of the populations for that is commonly described by one or more simple women (8). The largest increases were observed in indicators such as income, occupation, education Catalonia, where there was a 9.4% rise in absolute and place of residence. Substantial evidence sug- prevalence in men and a 6.5% rise in women. A gests that high SES is negatively correlated with small number of populations actually saw a statisti- obesity in developed countries, particularly among cally significant decrease in obesity prevalence over women, but positively correlated with obesity in the 5-year period. The most notable of these was in populations of developing countries. As developing Ticino (Switzerland), where absolute rates fell by countries undergo economic growth, the positive 11.7% in men and 9.6% in women. Charleroi in relationship between SES and obesity is slowly re- Belgium saw a 14.9% decrease in obesity prevalence placed by the negative correlation seen in modern in women but not in men. societies (see below, ‘What is Driving the Global Obesity Epidemic?’

Modern Societies Future Projections In developed countries there is usually an inverse Worldwide growth in the number of severely over- association between level of education and rates of weight adults is expected to be double that of under- obesity that is more pronounced among women. In weight adults between 1995 and 2025. Figure 1.2 the MONICA survey, a lower educational level was presents some crude projections of the expected rise associated with higher BMI in almost all female in obesity rates over the next 25 years for five of the populations (both surveys) and in about half of countries included in Table 1.6. These estimates are male populations. Between the two surveys, there based on a simple linear extrapolation of increases was a strengthening of this inverse association and observed over the period 1975—1995 and indicate the differences in relative body weight by education that by the year 2025, obesity rates could be as high increased. This suggests that socioeconomic in- as 40—45% in the USA, 30—40% in Australia, Eng- equality in health consequences associated with land and Mauritius, and over 20% in Brazil. It has obesity may actually be widening in many countries even been suggested that, if current trends persist, (10). One analysis has shown that reproductive his- the entire US population could be overweight with- tory, unhealthy dietary habits, and psychosocial in a few generations (9). stress may account for a large part of the association between low SES and obesity among middle- aged women (11). There is some evidence to suggest that there are racial differences between BMI and SES in develop- KEY FEATURES AND PATTERNS OF ed countries. Although women in the USA with low THE GLOBAL OBESITY EPIDEMIC incomes or low education are more likely to be obese than those of higher SES overall, this associ- Closer analysis of obesity prevalence and trend data ation was not found in a large survey of Mexican from around the world reveals a number of interest- American, Cuban American, and Puerto Rican ing patterns and features. These include an increase adults (12). Similar findings have been reported for in population mean BMI with socioeconomic tran- young girls where a lower prevalence of obesity was sition, a tendency for urban populations to have seen at higher levels of SES in white girls, but no higher rates of obesity than rural populations, a clear relationship was detected in black girls (13), tendency for peak rates of obesity to be reached at who tend to have much higher overall rates of obes- an earlier age in the less developed and newly indus- ity. trialized countries, and a tendency for women to have higher rates of obesity than men. These and others are considered in some detail below. OBESITY AS A GLOBAL PROBLEM 11

Figure 1.2 Projected increases in obesity prevalence. The ﬁgure illustrates the rate at which obesity prevalence is increasing in selected countries. It is based on crude projections from repeated national surveys. Source: IOTF unpublished

Developing and Transition Societies obesity than rural populations, especially in less developed nations. Urbanization causes people to New evidence from India illustrates the positive move away from their traditional way of living and association between SES and obesity in developing is associated with a wide range of factors which countries. Nearly a third of males, and more than adversely aﬀect diet and physical activity levels. half of females, belonging to the ‘upper middle class’ These include a shift to sedentary occupations, de- in urban areas are currently overweight (BMI pendency on automated transport, reliance on 9 25). This is in stark contrast to the prevalence of processed convenience foods, and exposure to overweight among slum dwellers (see Table 1.7) aggressive food marketing and advertising. Detri- (14). mental changes to family structures and value sys- In Latin American and a number of Caribbean tems may also be an important contributor to re- countries, a recent assessment of maternal and child duced physical activity and poor diet associated obesity from national surveys since 1982 also found with this shift. a tendency for higher obesity rates in poorly In most countries, urbanization has led to popu- educated women throughout the region, except in lations consuming smaller proportions of complex Haiti and Guatemala where the reverse was true. carbohydrates, greater proportions of fats and ani- mal products, more sugar, more processed foods, and more foods consumed away from home. Ur- Urban Residence banization also has eﬀects on physical activity levels. In Asian cities, bicycles are rapidly being Urban populations tend to have higher rates of displaced by motorbikes and cars with nearly 12 INTERNATIONAL TEXTBOOK OF OBESITY Table 1.7 Prevalence of overweight (BMI 925) in urban adults by socioeconomic status in Delhi, India 20 % Overweight MEN WOMEN

Socioeconomic status Males Females % 10 Middle class 1. High 32 50 2. Middle 16 30 3. Low 7 28 0 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 Slum (poor) 1 4 Age (years)

Source: Gopalan (14). Figure 1.3 Obesity prevalence across the lifespan in the Neth- erlands. There is a consistent rise in the prevalence of obesity throughout all age groups in the Dutch population, reaching a peak in the seventh decade. Source: Seidell (15) 10 000 cars being added to the automobile fleet every month in Delhi. Meanwhile the rural populations are mainly engaged in agricultural occupa- Obesity rates tend to decline in age groups older tions involving manual labour and a fairly high than this in association with the high mortality that level of physical activity. accompanies the rapidly developing diabetes and Steady urban migration has been an important cardiovascular disease (CVD). feature of the ongoing developmental transition in all developing countries. Asia’s urban population is expected to exceed 1242 million by the year 2000, a Gender Differences more than fivefold increase since 1950. This process is expected to continue in the decades to follow. By More women than men tend to be obese whereas 2025, the world’s urban population is expected to the reverse is true for overweight (BMI .25). This reach 5 billion (61% of the world’s people), of whom can be seen in countries as diverse as England, 77% will live in less developed countries. Mauritius, Japan and Saudi Arabia. There are likely to be many social influences that differentially influence male and female food intake and energy expenditure patterns. However, it is Age clear that biological and evolutionary components are also important factors underlying the differen- Figure 1.3 shows the general pattern of overweight ces in rates of obesity between the sexes. In all and obesity in the Netherlands, where a general rise populations, from contemporary hunting and in body weight and a modest increase in percentage gathering groups to those in complex industrial body fat occur over the lifespan, at least until 60—65 countries, women have more overall fat and much years of age. This is reflected by an increase in more peripheral body fat in the legs and hips than obesity prevalence with age, reaching a maximum men. In addition, there appears to be a tendency for in the 60s, and then declining steadily thereafter. females to channel extra energy into fat storage in The decline is related in part to selective survival of contrast to men who utilize a higher proportion of people with a lower BMI. The issue is further com- the energy to make protein and muscle. These gen- plicated by the fact that BMI is not as reliable a der differences are believed to be associated with the measure of adiposity in old age because a decrease need for adequate fat deposits to ensure reproduc- during this period often reflects a decrease in lean tive capacity in females. Men have, proportionally, body mass rather than fat mass. much more central body fat. They also have a high- Peak rates of obesity and the associated health er proportion of lean muscle mass which leads to a effects tend to be reached at a much earlier age in higher basal energy expenditure. developing economies. In countries such as West- ern Somoa, the maximum rates of obesity tend to be reached at around 40 years of age (Figure 1.4). OBESITY AS A GLOBAL PROBLEM 13

Figure 1.4 Obesity prevalence across the lifespan in Western Samoa. Peak rates of obesity are reached at around 40 years in communities of Western Samoa. Source: Hodge et al. (16) High-risk Groups for Weight Gain Vulnerable Periods of Life As outlined above, a general rise in body weight and Minority Populations in Industrialized a modest increase in percent body fat can be ex- Countries pected with age. However, there are certain periods In many industrialized countries, minority ethnic of life when an individual may be particularly vul- groups are especially liable to obesity and its com- nerable to weight gain (Table 1.8). plications. Some researchers believe that this is the result of a genetic predisposition to store fat which only becomes apparent when the individuals are Other Factors Promoting Weight Gain exposed to a positive energy balance promoted by A number of other groups have been identiﬁed as modern lifestyles. Central obesity, hypertension and being at risk of weight gain and obesity for genetic, NIDDM are very common in urban Australian biological, lifestyle and other reasons. These include Aborigines, but can be reduced or even eliminated family history of obesity, smoking cessation, excess- within a very short time by simply reverting to a ive alcohol intake, drug treatment for a wide range more traditional diet. of medical conditions, certain disease states, It is likely that other factors, especially those changes in social circumstance, and recent success- associated with poverty, may also have a role to ful weight loss. Major reductions in activity as a play in the far higher levels of obesity and its com- result of, for example, sports injury can also lead to plications observed in minority populations. In na- substantial weight gain when there is not a compen- tive American and African American populations, satory decrease in habitual food intake. for instance, where poverty is common, low levels of activity stem from unemployment and poor diets reﬂect dependence on cheap high-fat processed foods. Rates of hypertension among African Ameri- WHAT IS DRIVING THE GLOBAL can females below the poverty level are 40% com- OBESITY EPIDEMIC? pared with 30% of those at or above the poverty level. The particularly high levels of obesity among The Changing Environment minority groups living in the USA are illustrated clearly in Figure 1.5. Although research advances have highlighted the importance of leptin and other molecular genetic 14 INTERNATIONAL TEXTBOOK OF OBESITY

Figure 1.5 Obesity prevalence among ethnic groups in the USA, illustrating the disparity that exists between different ethnic groups, particularly amongst women, in the level of overweight and obesity in the USA. Source: Flegal et al. (17) factors in determining individual susceptibility to placement of traditional foodstuffs by Westernized obesity, these cannot explain the current obesity high-fat products and other negative changes have epidemic. The rapid rise in global obesity rates has also been a product of this process. The end result is occurred in too short a time for there to have been often a move to weight-gain-promoting dietary any significant genetic modifications within popu- habits and physical activity patterns. lations. This suggests that changes to the environment—physical, socio-cultural, economic and political—are primarily responsible for the epidemic and that genetics, age, sex, hormonal effects and other such factors influence the susceptibility of Economic Growth and Modernization individuals to weight gain who are living in that environment. A key factor in the global coverage of the obesity There are a number of societal forces which epidemic, particularly with respect to developing underlie the environmental changes implicated in and transition countries, is economic growth. Rapid the obesity epidemic. These include modernization, urbanization, changing occupational structures and economic restructuring and transition to market shifts in dietary structure related to socioeconomic economies, increasing urbanization, changing occu- transition all affect population mean BMI. Demo- pational structures, technical and scientific develop- graphic shifts associated with higher life expectancy ments, political change, and globalization of food and reduced fertility rates, as well as shifts in pat- markets. Many of these factors are associated with terns of disease away from infection and nutrient improved standards of living and other societal ad- deficiency towards higher rates of non-communi- vances but urban crowding, increasing unemploy- cable diseases, are other components of this so- ment, family and community breakdown, and dis- called ‘transition’. OBESITY AS A GLOBAL PROBLEM 15 Table 1.8 Vulnerable periods of life for weight gain and the mean population BMI is 23 or below, there are very development of future obesity few individuals with a value of 30 kg/m or greater. Prenatal Poor growth and development of the unborn However, when mean BMI rises above 23 kg/m , baby can increase the risk of abdominal fatness, there is a corresponding increase in the prevalence obesity and related illness in later life. of obesity. An analysis by Rose (20) of 52 communi- Adiposity ‘Adiposity rebound’ describes a period, usually ties in the large multi-country INTERSALT Study rebound between the ages of 5 and 7, when BMI begins found that there is a 4.66% increase in the preva- (5—7 years) to increase rapidly. This period coincides with increased autonomy and socialization and so lence of obesity for every single unit increase in may represent a stage when the child is population BMI above 23 kg/m (Figure 1.7). particularly vulnerable to the adoption of behaviours that both influence and predispose to the development of obesity. Early adiposity The ‘Nutrition Transition’ rebound may be associated with an increased risk of obesity later in life. Generally, as incomes rise and populations become Adolescence This is a period of increased autonomy which is more urban, diets high in complex carbohydrates often associated with irregular meals, changed and fibre give way to varied diets with a higher food habits and periods of inactivity during proportion of fats, saturated fats and sugars. Recent leisure combined with physiological changes. These promote increased fat deposition, analyses of economic and food availability data, particularly in females. however, reveal a major shift in the structure of the Early Early adulthood is often associated with a global diet over the last 30 years. Innate preferences adulthood marked reduction in physical activity. This for palatable diets coupled with the greater avail- usually occurs between the ages of 15 and 19 ability of cheap vegetable oils in the global econ- years in women but as late as the early 30s in men. omic have resulted in greatly increased fat con- Pregnancy The average weight gain after pregnancy is less sumption and greater dietary diversity among low than 1 kg although the range is wide. In many income nations. As a result, the classic relationship developing countries, consecutive pregnancies between incomes and fat intakes has been lost, with with short spacing often result in weight loss the so-called ‘nutrition transition’ now occurring in rather than weight gain. Menopause Menopausal women are particularly prone to nations with much lower levels of gross national rapid weight gain. This is primarily due to product than previously. The process is accelerated reductions in activity although loss of the by rapid urbanization (21). menstrual cycle also affects food intake and reduces metabolic rate slightly. The Relationship Between Undernutrition Source: Gill (18). and Later Obesity In countries undergoing transition where overnutrition coexists with undernutrition, the shift in Effect on BMI Distribution population weight status has been linked to exag- Improvement in the socioeconomic conditions of a gerated problems of obesity and associated non- country tends to be accompanied by a population- communicable diseases in adults. wide shift in BMI so that problems of overweight Recent studies have shown that infants who were eventually replace those of underweight (Figure undernourished in utero and then born small have a 1.6). In the early stages of transition, undernutrition greater risk of becoming obese adults (22,23). In remains the principal concern in the poor whilst the particular, poor intrauterine nutrition appears to more affluent tend to show an increase in the pro- predispose some groups to abdominal obesity and portion of people with a high BMI. This often leads results in an earlier and more severe development of to a situation where overweight coexists with comorbid conditions such as hypertension, CHD underweight in the same country. As transition pro- and diabetes (24—26). The apparent impact of in- ceeds, overweight and obesity also begin to increase trauterine nutrition on the later structure and func- among the poor. tioning of the body has become known as ‘program- Even in affluent countries, the distribution of ming’ and is often referred to as the ‘Barker body fatness within a population ranges from hypothesis’, after one of the key researchers in- underweight through normal to obese. When the volved in developing this concept. 16 INTERNATIONAL TEXTBOOK OF OBESITY

Figure 1.6 BMI distribution for various adult populations worldwide (both sexes). As the proportion of the population with a low BMI decreases there is a consequent increase in the proportion of the population with an abnormally high BMI. Many countries have a situation of unacceptably high proportions of both under- and overweight. Source: WHO (19)

Figure 1.7 The relationship between population mean BMI and the prevalence of obesity, illustrating the direct association between population mean BMI and the prevalence of deviant (high) BMI values across 52 population samples from 32 countries (men and women aged 20—59 years). r : 0.94; b : 4.66% per unit BMI. Source: Rose (20) OBESITY AS A GLOBAL PROBLEM 17

Figure 1.8 Shifts in distribution of occupations for lower income countries, 1972—1993. There has been a steady decline in employment in labour intensive agricultural occupations and a concomitant increase in employment within the less physical demanding service sector. Source: Popkin and Doak (7)

The ramifications of programming are immense occupations that has been occurring in lower in- for countries such as India and China where a large come countries during the past several decades. proportion of infants are still born undernourished. There has been a move towards more capital inten- If these children are later exposed to high-fat diets sive and knowledge based employment that relies and sedentary lifestyles associated with economic far less on physical activity. In China, the rapid transition, and develop into obese adults, then it is decline in physical activity at work in urban areas likely that they will suffer severe consequences in has been associated with increased levels of adult the form of early heart disease, hypertension and obesity (27). Large shifts towards less physically diabetes. demanding work have also been observed on a Central obesity is already emerging as a serious worldwide basis, both in the proportion of people problem in India, even at low relative weight; working in agriculture, industry and services, and in among non-overweight urban middle-class resi- the type of work within most occupations. dents with BMI less than 25 kg/m, nearly 20% of males and 22% of females had a high waist-to-hip ratio. In overweight subjects with a BMI over 25 kg/ Other Possible Explanations m, abdominal obesity was found in a striking 68% of males and 58% of females. Changing Demographic Structure of In many populations undergoing rapid modern- Populations ization and economic growth, high levels of obesity are associated with high rates of NIDDM, hyper- Obesity, like many other non-communicable dis- tension, dyslipidaemia and CVD as well as alcohol eases, is age dependent and the highest rates are abuse and cigarette smoking. This has been de- generally found in older age groups. The recent scibed as the ‘New World syndrome’ and is respon- decline in fertility rates and increase in proportion sible for the disproportionately high rates of of the population surviving into adulthood has led mortality in developing nations and among the dis- to a shift in the age structure of most populations advantaged ethnic minority groups in developed with the result that they are generally older than a countries. few decades ago. This is particularly evident in developing countries. It has been suggested that such Occupational Structure changes in the demographics of societies could Figure 1.8 shows the shift in the distribution of make a significant contribution to inflating the 18 INTERNATIONAL TEXTBOOK OF OBESITY measured increase over time in a number of chronic OBESITY IN CHILDREN AND diseases such as obesity (28). However, the finding ADOLESCENTS that the greatest increases over the last few decades in mean body weight and rates of obesity have Obesity is also emerging as a serious global health occurred in younger age groups does not support problem among children and adolescents. Although this explanation for the recent obesity epidemic. good quality nationally representative data are still lacking, studies have generally reported a substantial rise in prevalence in both industrialized and Smoking Cessation and Increasing Obesity developing countries. Rates It has been suggested that the fall in smoking rates observed over recent years in many industrialized Defining Obesity in Children and countries has made a significant contribution to the Adolescents rises in mean body weight and rates of obesity. Studies have shown that smokers have significantly The major factor limiting our understanding of the lower mean BMI than those who have never true extent of the childhood obesity problem is the smoked and that male ex-smokers tend to have the lack of a standard population-level methodology highest level of BMI (29). Mean weight gain attribu- for measuring overweight and obesity in children table to smoking cessation in a nationally represen- and adolescents. Presently a number of different tative cohort of smokers and non-smokers in the methods or indices are in use with a variety of USA was 2.8 kg in men and 3.8 kg in women, with 9 cut-off points for designating a child as obese. The heavy smokers ( 15 cigarettes per day) and US National Center for Health Statistics (NCHS) younger people at higher risk of weight gain 9 growth reference charts have been recommended by ( 13 kg) on cessation (30). However, analyses of the the WHO for international use since the late 1970s contribution of smoking cessation to population but a number of serious technical and biological weight gain have been equivocal. One study sugges- problems have been identified with their develop- ted that smoking cessation may account for up to ment and application. 20% of the increase in overweight adults in the USA An expert working group of the International but other studies have indicated that the contribu- Obesity Task Force investigated this issue and con- tion may be much lower. Declines in self-reported cluded that BMI-for-age, based on a redefined in- cigarette smoking accounted for only 7% of change ternational reference population from 5 to 18 years, in BMI among males and 10% in females in a New was a reasonable index of adiposity and could be Zealand Study (31). Studies from Australia (32) and used for population studies. They identified a novel Finland (33) did not find significant differences in approach to determine cut-off values that classify the rates of weight increase over time between children as overweight or obese using percentiles smokers, non-smokers and ex-smokers. that correlate to the standard cut-off points for BMI in adults (34). WHO is also in the process of developing a new growth reference for infants and Cultural Body Shape Ideals children from birth to 5 years. Culturally defined standards of a beautiful body vary between societies and across historical periods of time. ‘Fatness’ is still viewed as a sign of health and prosperity in many developing countries, es- The Scale of the Childhood Obesity pecially where conditions make it easy to remain Problem lean. ‘Bigness’ (large structure and muscularity but not necessarily fatness) also tends to be viewed as Despite the lack of agreement over childhood obes- the male body ideal in most developed countries. ity classification, there is ample evidence to illus- Such views can inhibit patients from seeking treat- trate the scale of the problem across the world. ment and support the continuing upwards trend in Using the existing WHO standards, the 1998 obesity rates. World Health Report indicated that about 22 mil- OBESITY AS A GLOBAL PROBLEM 19 Table 1.9 Prevalence of overweight? in 6- to 8-year-old children

USA China Russia South Africa Brazil (1988—1991) (1993) (1994—1995) (1994) (1989)

Girls 24.2 12.2 17.8 20.3 10.5 Boys 21.3 14.1 25.6 25.0 12.8

?Defined as BMI higher than the US reference NHES 85th percentile. Source: Popkin et al. (35). lion children under 5 years are overweight across Health Impact of Obesity in Childhood the world (2). This was based on weight-for-height data from 79 developing countries and a number of Obese children and adolescents are at increased risk industrialized countries. Once the new growth refer- of developing a number of health problems. The ence is available a more realistic estimate should be most significant long-term consequence is the per- possible. sistence of obesity and its associated health risks Another comparison performed using the US into adulthood. Some 30% of obese children be- NHES criteria also revealed the alarmingly high come obese adults. This is more likely when the levels of overweight that exist in older children in onset of obesity is in late childhood or adolescence both developed and developing countries. In some and when the obesity is severe. Other obesity- countries, up to a quarter of the school age child related symptoms include psychosocial problems, population is already overweight (Table 1.9). raised blood pressure and serum triglycerides, ab- Trend data suggest that the childhood obesity normal glucose metabolism, hepatic gastrointes- problem is increasing rapidly in many parts of the tinal disturbances, sleep apnoea and orthopaedic world. In the US, the percentage of young people complications. aged 5—14 who are overweight has more than doubled in the past 30 years. Prevalence has risen from 15% in 1973—1974 to 32% in 1992—1994. Mean- while, in England, triceps skinfold measurement in- Stunting and Obesity creased by almost 8% in 7-year-old English boys and by 7% in 7-year-old girls between 1972 and A number of studies have indicated that there is an 1994. In Scotland over the same period, triceps important association between stunting and over- skinfold measurement increased by nearly 10% weight or obesity in a variety of ethnic, environ- in 7-year-old boys and by 11% in 7-year-old mental and social backgrounds. Popkin et al. (35), girls. Weight for height index followed a similar for example, found that the income-adjusted risk pattern. ratios of being overweight for a stunted child in four Childhood obesity is also increasing in Asia. In nations undergoing transition ranged from 1.7 to Thailand, the prevalence of obesity in 6- to 12-year- 7.8. Obesity associated with stunting was also more old children rose from 12.2% in 1991 to 15.6% in common than obesity without stunting in a shanty- 1993. In Izumiohtsu city in Japan, the percentage of town population in the city of Sao Paulo, in both obese children aged 6—14 years doubled from 5 to younger children and adolescents (36). 10% between 1974 and 1993. The association between stunting and obesity has Data from developing countries in Latin America serious public health implications, particularly for show that urban residency, high SES and higher lower income countries, but the underlying mech- maternal education are associated with greater risk anisms remain relatively unexplored. Recently, of overweight in children and that obesity is more Sawaya et al. (37) suggested that stunting may in- common in girls than in boys. In developed coun- crease the susceptibility to excess body fat gain in tries an opposite association between SES and children who consume a high fat diet. A significant obesity is often found, with children from poorer association was found between the baseline percen- educated parents with lower occupations more like- tage of dietary energy supplied by fat and the gain in ly to be overweight. weight-for-height during follow-up in girls with 20 INTERNATIONAL TEXTBOOK OF OBESITY mild stunting (P : 0.048), but not in the non- terns of children are changing particularly rapidly stunted control girls (P : 0.245). Despite clear in- in developing countries where high energy-dense, dications that catch-up growth cannot be achieved manufactured food is replacing less energy-dense outside critical growth windows, many countries traditional food and snacks based on cereals, fruits continue with poorly targeted nutrition supple- and vegetables. mentation programmes based on energy-dense foods.

THE NEED FOR GLOBAL ACTION

Key Factors Underlying the Increase in Obesity is a serious international public health Childhood Obesity Rates problem which urgently needs action on a global scale. Governments, international agencies, indus- The fact that obesity is emerging as the most preva- try/trade, the media, health professionals and con- lent nutritional disease among children and adoles- sumers, among others, all have important roles to cents in the developed world is hardly surprising. As play in arresting this epidemic. outlined earlier, the highly technological societies of Strategies aimed at preventing weight gain and today have created an environment where it is in- obesity are likely to be more cost effective, and to creasingly convenient to remain sedentary whilst all have a greater positive impact on long-term control forms of physical activity and active recreation are of body weight, than treating obesity once it has discouraged. Children are particularly susceptible developed. The majority of treatment therapies fail to such changes as many of the decisions about diet to keep weight off in the long-term and health care and physical activity patterns are beyond their per- resources are no longer sufficient to offer treatment sonal control. Parents are becoming increasingly to all. In countries still struggling with high levels of concerned about the safety of their children and are undernutrition, tackling the problem of overweight preventing them from walking or cycling to school and obesity poses even more of a challenge as many or playing in public spaces. In addition, lack of are not prepared institutionally to deal with prob- resources, space and staff for supervision has led to lems of diet and chronic disease. a reduction in the time spent in active play or sports In the face of the current environment character- when children are at school. As a result, the physical ized by sedentary occupations and persistent temp- activity levels of children are dropping drastically tation of high fat/energy-dense food, action to pre- and more sedentary pursuits such as television vent obesity must include measures to reduce the watching are replacing time once spent in active obesity-promoting aspects of the environment. Pre- play. This is a trend that is spreading through- vious attempts to improve community diet and out many newly industrialized and developing physical activity habits have shown that efforts countries as safety becomes a serious issue in over- cannot rely solely on health education strategies crowded urban areas and consumer goods such as aimed at changing individual behaviour. Living televisions become more accessible. environments need to be improved so that they Television advertising and the rapid spread of both promote and support healthy eating and ready-prepared foods directly marketed at children physical activity habits throughout the life cycle for appear to have greatly influenced children’s food the entire population. Strategies are needed which preferences. There is a great deal of concern that the address the underlying societal causes of obesity majority of food and drink advertisements screened through action in sectors such as transport, envi- during children’s television programmes are for ronment, employment conditions, education, products high in fat and/or sugar, which clearly health and food policies, social and economic poli- undermine messages for healthy eating. Only a very cies (Table 1.10). few countries such as Norway and Sweden have For those individuals and subgroups of the popu- sought to restrict the level of television advertising lation who have already developed, or are at in- directed towards children under 12 and during creased risk of developing, obesity and the asso- children’s programmes. The globalization of world ciated health complications, obesity management food markets has meant that traditional eating pat- programmes within health care and community ser- OBESITY AS A GLOBAL PROBLEM 21 Table 1.10 Potential public health interventions to prevent epidemic, this has a key role in developing, improv- obesity ing and refining strategies to deal with it. Predominantly food related 1. Increase food industry development, production, distribution and promotion of products low in dietary fat and energy REFERENCES 2. Use pricing strategies to promote purchase of healthy foods 3. Improve quality of food labelling 4. Increase mass media promotion of healthy foods 1. World Health Organization. Obesity: Preventingand 5. Promote water as the main daily drink Managing the Global Epidemic. Report of a WHO Consulta- 6. Promote development and implementation of appropriate tion on Obesity. Geneva, 3—5 June 1997. Geneva: World nutrition standards and guidelines for catering Health Organization, 1998 WHO/NUT/NCD/98.1. establishments (public and private) 2. World Health Organization. The World Health Report 7. Regulate food advertising and marketing practices aimed at 1998—Life in the 21st century: a vision for all. Geneva: World children Health Organization, 1998. 8. Provide land in towns and cities for ‘family’ growing of 3. Deurenberg P, Yap M, van Staveren WA. Body mass index vegetables, legumes and other healthy produce and percent body fat: a meta analysis among different ethnic groups. Int J Obes 1998; 22: 1164—1171. Predominantly activity related 4. Seidell JC. Effects of obesity. Medicine 1998; 4—8. 1. Improve public transport to reduce dependence on the 5. Wolf AM, Colditz GA. Current estimates of the economic motor car cost of obesity in the United States. Obes Res 1998; 6:97—106. 2. Implement measures to promote walking and cycling as 6. Hodge AM, Dowse GK, Zimmet PZ, Collins VR. Preva- means of transport lence and secular trends in obesity in Pacific and Indian 3. Change building codes to promote use of stairs instead of Ocean island populations. Obes Res 1995; 3 (Suppl 2): elevators and escalators 77s—87s. 4. Increase provision of affordable local exercise/recreational 7. Popkin BM, Doak CM. The obesity epidemic is a worldwide facilities and programmes phenomenon. Nutr Rev 1998; 56: 106—114. 5. Provide flexible working arrangements to allow time for 8. Dobson AJ, Evans A, Ferrario M, Kuulasmaa KA, Molt- exercise (and to decrease reliance on convenience processed chanovVA, Sans S, Tunstall-Pedoe H, Tuomilehto JO, foods) Wedel H, Yarnell J. Changes in estimated coronary risk in 6. Provide exercise and changing facilities at work the 1980s: data from 38 populations in the WHO MONICA 7. Promote learning and practice of healthy physical activity Project. World Health Organization. Monitoring trends and (and nutrition) habits through schools determinants in cardiovascular diseases. Ann Med 1998; 30: 199—205. Health sector related 9. Foreyt J, Goodrick K. The ultimate triumph of obesity. 1. Mass media public awareness campaign on the need to Lancet 1995; 346: 134—135. maintain a healthy weight throughout life 10. Molarius A, Seidell JC, Sans S, Tuomilehto J, Kuulasmaa K. 2. Build economic incentives into health insurance plans Educational level and relative body weight, and changes in 3. Provide adequate training in obesity prevention and the association over 10 years—an international perspective management for physicians and other health care workers from the WHO Monica project. Int J Obes 1998; 22: S43. 11. Wamala SP, Wolk A, Orth-Gomer K. Determinants of obes- Source: IOTF unpublished. ity in relation to socioeconomic status among middle-aged Swedish women. Prev Med 1997; 26: 734—744. 12. Khan LK, Sobal J, Martorell R. Acculturation, socioeconomic status, and obesity in Mexican Americans, Cuban vices are essential. The effectiveness of such pro- Americans, and Puerto Ricans. Int J Obes 1997; 21:91—96. 13. Kimm SY, Obarzanek E, Barton BA, Aston CE, Similo SL, grammes is likely to be enhanced if improved and Morrison JA, Sabry ZI, Schreiber GB, McMahon RP. Race, extended training of all relevant health care workers socioeconomic status, and obesity in 9- to 10-year-old girls: is provided. Obesity needs to be viewed as a disease the NHLBI Growth and Health Study. Ann Epidemiol 1996; in its own right and one which warrants interven- 6: 266—275. tion even when comorbidities are not present. 14. Gopalan C. Obesity in the Indian urban ‘Middle Class’. Nutrition Foundation of India Bulletin 1998; 19:1—5. Negative attitudes of health care professionals to- 15. Seidell JC. Obesity in Europe. Obes Res 1995; 3 (Suppl 2): wards the condition are not helpful. 89s—93s. Finally, in all interventions aimed at preventing 16. Hodge AM, Dowse GK, Toelupe P, Collins VR, Imo T, and managing overweight and obesity, systematic Zimmet PZ. Dramatic increase in the prevalence of obesity in Western Samoa over the 13-year period 1978—1991. Int J assessment and evaluation should be a routine el- Obes 1994; 18: 419—428. ement. Together with research into the develop- 17. Flegal KM, Carroll MD, Kuczmarski RJ, Johnson ment, consequences and scale of the global obesity CL. Overweight and obesity in the United States: prevalence 22 INTERNATIONAL TEXTBOOK OF OBESITY and trends, 1960—1994. Int J Obes 1998; 22:39—47. 29. Molarius A, Seidell JC, Kuulasmaa K, Dobson AJ, Sans S. 18. Gill TP. Key issues in the prevention of obesity. Br Med Bull Smoking and relative body weight: an international perspec- 1997; 53: 359—388. tive from the WHO MONICA Project. J Epidemiol Commu- 19. World Health Organization. Physical Status: the use and nity Health 1997; 51: 252—260. interpretation of anthropometry. Report of a WHO Expert 30. Williamson DF. Smoking cessation and severity of weight Committee. Geneva: World Health Organization, 1995 gain in a national cohort. N EngJ Med 1991; 324: 729—745. (WHO Technical Report Series, No. 854). 31. Simmons G, Jackson R, Swinburn B, Yee RL. The increasing 20. Rose G. Population distributions of risk and disease. Nutr prevalence of obesity in New Zealand: is it related to recent Metab Cardiovasc Dis 1991; 1:37—40. trends in smoking and physical activity. NZ Med J 1996; 21. Drewnowski A, Popkin BM. The nutrition transition: new 109:90—92. trends in the global diet. Nutr Rev 1997; 55:31—43. 32. Boyle CA, Dobson AJ, Egger G, Magnus P. Can the increas- 22. Ravelli GP, Stein ZA, Susser M. Obesity in young men after ing weight of Australians be explained by the decreasing famine exposure in utero and early infancy. N EngJ Med prevalence of smoking? Int J Obes 1994; 18:55—60. 1976; 295: 349—353. 33. Laaksonen M, Rahkonen O, Prattala R. Smoking status and 23. Jackson AA, Langley-Evans SC. McCarthy HD. Nutritional relative weight by educational level in Finland, 1978—1995. influences on early life upon obesity and body proportions. Prev Med 1998; 27: 431—437. In: Chadwick DJ, Cardew GC (eds) The Origins and Conse- 34. Dietz WH, Robinson TN. Use of the body mass index (BMI) quences of Obesity. Chichester: Wiley, 1996: 118—137 (Ciba as a measure of overweight in children and adolescents. J Foundation Symposium 201). Pediatr 1998; 132: 191—193. 24. Law CM, Barker DJ, Osmond C, Fall CH, Simmonds SJ. 35. Popkin BM, Richards MK, Montiero CA. Stunting is asso- Early growth and abdominal fatness in adult life. J Epidemiol ciated with overweight in children of four nations that are Community Health 1992; 46: 184—186. undergoing the nutrition transition. J Nutr 1996; 126: 25. Schroeder DG, Martorell R, Flores R. Infant and child 3009—3016. growth and fatness and fat distribution in Guatemalan 36. Sawaya AL, Dallal G, Solymos G, de Sousa MH, Ventura adults. Am J Epidemiol 1999; 149: 177—185. ML, Roberts SB, Sigulem DM. Obesity and malnutrition in 26. Barker DJ. Maternal nutrition, fetal nutrition, and disease in a Shantytown population in the city of Sao Paulo, Brazil. later life. Nutrition 1997; 13: 807—813. Obes Res 1995; 3 (Suppl 2): 107s—115s. 27. Popkin BM, Paeratakul S, Zhai F, Ge K. Dietary and envi- 37. Sawaya AL, Grillo LP, Verreschi I, da Silva AC, Roberts SB. ronmental correlates of obesity in a population study in Mild stunting is associated with higher susceptibility to the China. Obes Res 1995; 3 (Suppl 2): 135s—143s. effects of high fat diets: studies in a shantytown population in 28. Pellitier DL, Rahn, M. Trends in body mass index in devel- Sao Paulo, Brazil. J Nutr 1998; 128 (2 Suppl): 415S—420S. oping countries. Food Nutr Bull 1998; 19: 223—239.