Assessing Adiposity

Abstract

HomeCirculationVol. 124, No. 18Assessing Adiposity Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessResearch ArticlePDF/EPUBAssessing AdiposityA Scientific Statement From the American Heart Association Marc-Andre Cornier, MD, Chair, Jean-Pierre Després, PhD, FAHA, Nichola Davis, MD, MS, Daurice A. Grossniklaus, RN, MEd, PhD, Samuel Klein, MD, FAHA, Benoit Lamarche, PhD, FAHA, Francisco Lopez-Jimenez, MD, MSc, Goutham Rao, MD, Marie-Pierre St-Onge, PhD, Amytis Towfighi, MD and Paul Poirier, MD, PhD, FAHA Marc-Andre CornierMarc-Andre Cornier Search for more papers by this author , Jean-Pierre DesprésJean-Pierre Després Search for more papers by this author , Nichola DavisNichola Davis Search for more papers by this author , Daurice A. GrossniklausDaurice A. Grossniklaus Search for more papers by this author , Samuel KleinSamuel Klein Search for more papers by this author , Benoit LamarcheBenoit Lamarche Search for more papers by this author , Francisco Lopez-JimenezFrancisco Lopez-Jimenez Search for more papers by this author , Goutham RaoGoutham Rao Search for more papers by this author , Marie-Pierre St-OngeMarie-Pierre St-Onge Search for more papers by this author , Amytis TowfighiAmytis Towfighi Search for more papers by this author and Paul PoirierPaul Poirier Search for more papers by this author and on behalf of the American Heart Association Obesity Committee of the Council on Nutritionand Physical Activity and Metabolismand Council on Arteriosclerosisand Thrombosis and Vascular Biologyand Council on Cardiovascular Disease in the Youngand Council on Cardiovascular Radiology and Interventionand Council on Cardiovascular Nursing, Council on Epidemiology and Preventionand Council on the Kidney in Cardiovascular Disease, and Stroke Council Originally published26 Sep 2011https://doi.org/10.1161/CIR.0b013e318233bc6aCirculation. 2011;124:1996–2019Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions: January 1, 2011: Previous Version 1 IntroductionThe prevalence of obesity in the United States and the world has risen to epidemic/pandemic proportions. This increase has occurred despite great efforts by healthcare providers and consumers alike to improve the health-related behaviors of the population and a tremendous push from the scientific community to better understand the pathophysiology of obesity. This epidemic is all the more concerning given the clear association between excess adiposity and adverse health consequences such as cardiovascular disease (CVD) and type 2 diabetes mellitus (T2DM). The risks associated with overweight/obesity are primarily related to the deposition of adipose tissue, which leads to excess adiposity or body fatness. Furthermore, weight loss, specifically loss of body fat, is associated with improvement in obesity-related comorbidities. Before weight loss interventions can be recommended, however, patients must be assessed for their adiposity-related risk. Unfortunately, healthcare providers and systems have not done a good job of assessing for excess adiposity even in its simplest form, such as measuring body mass index (BMI). It is for these reasons that we must emphasize the importance of assessing adiposity in clinical practices. Although it can be argued that the entire population should be targeted as an important public health issue with a goal of prevention of weight gain and obesity, there are currently so many “at risk” individuals that simple strategies to identify and treat those individuals are necessary. We must identify those individuals at highest risk of comorbidities in order to identify those who might benefit the most from aggressive weight management.This scientific statement will first briefly review the epidemiology of obesity and its related comorbidities, supporting the need for improved assessment of adiposity in daily clinical practice. This will be followed by a discussion of some of the challenges and issues associated with assessing adiposity and then by a review of the methods available for assessing adiposity in adults. Finally, practical recommendations for the clinician in practice will be given with a goal of identifying more at-risk overweight/obese individuals.Excess Adiposity: The Scope of the ProblemClassification of Overweight and ObesityThe Centers for Disease Control and Prevention classify obesity according to BMI1 as summarized in Table 1.2,2a Among adults, a BMI between 18.5 and 24.9 kg/m2 corresponds to a healthy weight, BMI between 25.0 and 29.9 kg/m2 is overweight, and BMI of ≥30.0 kg/m2 is obese. The degree of obesity is classified separately. A BMI of 30.0 to 34.9 kg/m2 is class 1 or mild obesity, 35.0 to 39.9 kg/m2 is class 2 or moderate obesity, and ≥40.0 kg/m2 is class 3 or severe obesity. The absolute value of BMI is not used to classify weight status in children because change in BMI is normal and expected as children grow and develop. Instead, BMI percentiles adjusted for age and sex and calculated based on a compilation of national survey data collected over a 30-year period are used. In children 2 to 19 years of age, a BMI between the 5th and <85th percentiles is healthy, between the 85th and <95th percentiles is overweight, and at or above the 95th percentile is obese.3Table 1. Classification of Body Weight According to BMI in Adults and in Children2,2aAdults Underweight: BMI <18.5 kg/m2 Normal or acceptable weight: BMI 18.5–24.9 kg/m2 Overweight: BMI 25–29.9 kg/m2 Obese: BMI ≥30 kg/m2 Class 1: BMI 30–34.9 kg/m2 Class 2: BMI 35.0–39.9 kg/m2 Class 3: BMI ≥40 kg/m2 (severe, extreme, or morbid obesity)Children (youths between 2 and 18 y of age) Overweight: BMI of 85th to 94th percentile Obese: BMI of 95th percentile or BMI of ≥30 kg/m2, whichever is lower Severe obesity: 99th percentile BMI ≥30 to 32 kg/m2 for youths 10–12 y of age ≥34 kg/m2 for youths 14–16 y of ageBMI indicates body mass index.Adapted from Reference 2a with kind permission from Springer Science+Business Media. Copyright © 1997, Springer Science+Business Media.Epidemiology of Overweight and ObesityOn the basis of data collected as part of the 2007 to 2008 National Health and Nutrition Examination Survey, in the United States 72.3% of men, 64.1% of women, and 68.0% of adults overall were either overweight or obese, with 32.2% of men, 35.5% of women, and 33.8% of adults overall being obese. The rates were stable for women for the 10 years preceding the survey and showed a slight increase for men during that period.4 Data from the same period revealed a prevalence of obesity of 16.9% and a combined overweight and obesity prevalence of 31.7% among children ages 2 to 19 years.5 As among adults, there are encouraging data to suggest that these rates have stabilized, with the exception of an increase in the number of boys ages 6 to 19 years with a BMI percentile at or above the 97th.There are significant racial and regional differences in the prevalence of obesity. Non-Hispanic white adults have an obesity prevalence of 32.8%, compared with 44.1% for non-Hispanic blacks and 38.7% for Hispanics. Racial differences are especially pronounced among women: 33.0% of non-Hispanic white women are obese compared with 49.6% and 43.0% of non-Hispanic black and Hispanic women, respectively. Similar racial differences are present among children: 15.3% of white children are obese, compared with 20.0% of non-Hispanic black children and 20.9% of Hispanic children. Regional data obtained from the Behavioral Risk Factor Surveillance System, which relies on self-reported height and weight, reveal a range of adult obesity prevalence by county of 12.4% to 43.7%. The highest rates of obesity are in the South, the western Appalachians, and coastal North and South Carolina. The lowest rates are in the West, the northern Plains, and New England.6 Among states, according to the Behavioral Risk Factor Surveillance System, Mississippi has the highest adult obesity prevalence (32.8%), and Colorado has the lowest, with a rate of 18.5%.7Complications of Excess AdiposityA substantial body of evidence demonstrates a harmful effect of obesity and excess adiposity on cardiovascular health. Both abdominal obesity and general obesity are independently associated with cerebrovascular disease (odds ratio [OR] range 1.22–2.37)8–14 and coronary heart disease (OR range 1.21–3.25).14–18 Furthermore, obesity is associated with increased overall mortality19–21 (OR range 1.9–2.42) and mortality after cardiovascular events (OR range 1.07–1.94).22–25 Although some studies have shown a J-shaped curve between BMI and mortality, with higher mortality rates in individuals in both the highest and lowest BMI categories, often referred to as the “obesity paradox,”12,26 comorbidities associated with excess adiposity appear to increase across the continuum of overweight and obesity. Furthermore, abdominal obesity, an important component of the metabolic or the cardiometabolic syndrome, has been shown to be associated with stroke,27 coronary heart disease,28 and overall mortality29,30 independent of other cardiac risk factors. Overweight and obesity are also associated with increased risk of a number of other comorbid conditions, such as T2DM, systemic hypertension, dyslipidemia, obstructive sleep apnea, osteoarthritis, depression, gout, nonalcoholic liver disease, reproductive-endocrine disorders, and several cancers, to name a few.Assessing Excess Adiposity: The ProblemsTotal Body Fat Versus Distribution of Body Fat Versus Body Composition Versus Ectopic FatHeterogeneity of ObesityAlthough numerous population-based studies have shown that there is a clear relationship between BMI (the most common index of adiposity used in clinical practice) and the documented comorbidities associated with excess body fatness,31–34 obesity has remained a puzzling condition for clinicians because of its remarkable heterogeneity. For instance, although obese patients are as a group at greater risk of comorbidities than normal-weight individuals, some obese patients may nevertheless show trivial or even no metabolic complications, the so-called metabolically healthy obese,35–42 whereas others with the same level of obesity (on the basis of similar BMI values) could show numerous metabolic abnormalities, including insulin resistance, glucose intolerance, dyslipidemia, systemic hypertension, and a prothrombotic-inflammatory profile.43–56 Thus, although BMI has been useful to describe secular changes in adiposity at the population level, BMI cannot always properly discriminate the risk of chronic disease at the individual level.Body Shape Matters: The PioneerNumerous epidemiological and metabolic studies published over the past 3 decades have provided support to Jean Vague's early seminal observations57,58 that the common complications of obesity, such as insulin resistance, atherogenic dyslipidemia, T2DM, and CVD, were more closely related to the distribution of body fat than to the absolute degree of fatness per se.19,32,43–;56,59–64 Vague coined the term “android” obesity (more frequently found in men) to describe the high-risk form of obesity, whereas he introduced the term “gynoid” obesity to describe the low risk typical of lower-body adiposity more frequently found in premenopausal women.57The Renaissance of Regional Adipose Tissue DistributionIn the early 1980s, Björntorp and colleagues59,60,65,66 in Gothenburg, Sweden, and Kissebah and collaborators64,67 in Milwaukee, WI, reported that when the ratio of waist to hip circumferences (waist-hip ratio [WHR]) was used as an index of the relative accumulation of abdominal fat, this variable was related both to the risk of coronary heart disease and T2DM and to a diabetogenic/atherogenic metabolic risk profile. The rationale for this ratio was simple: The greater the relative accumulation of abdominal fat, the greater the waist circumference (WC) relative to the hip girth. This early work has had a tremendous impact on the field of body fat distribution and health, because it provided evidence that body fat distribution deserved more attention as a predictor of the comorbidities than had been, in the past, attributed to excess body fatness per se.Imaging Techniques: A Major Advancement in the Study of Body Fat DistributionIn the mid-1980s, the introduction of imaging techniques such as computed tomography (CT) gave investigators interested in body fat topography a more sophisticated tool that allowed for more precise measurements of regional fat accumulation. CT was found to be particularly helpful in distinguishing the abdominal fat stored subcutaneously (ie, subcutaneous adipose tissue [SAT]) from the adipose tissue located in the abdominal cavity, including omental, mesenteric, and retroperitoneal adipose tissue, which has commonly been described as intra-abdominal or visceral adipose tissue (VAT). Studies that have measured SAT and VAT areas with CT have shown that although the size of both adipose depots is associated with a progressive deterioration in cardiometabolic risk profile, when matched for levels of SAT, individuals with excess VAT and deep SAT were characterized by a more diabetogenic/atherogenic risk factor profile.43–56,68 These results have provided robust evidence that although excess fatness is associated with metabolic abnormalities, high levels specifically of VAT are characterized by the most severe metabolic abnormalities. More recent epidemiological studies that have used imaging techniques such as CT or magnetic resonance imaging (MRI) have been able to identify the respective contributions of SAT and VAT in very large study samples and have clearly shown that visceral adiposity is associated with more severe metabolic disturbances than subcutaneous adiposity.53,69,70Factors Associated With Individual Differences in Visceral AdiposityThe factors that regulate regional body fat deposition have been investigated extensively (Table 2). Several factors are associated with differences in visceral adiposity, such as sex, age, genetic factors, hormonal profile, smoking, and nutritional factors, as well as vigorous endurance exercise.71–73 Major sex differences are observed in visceral adiposity before menopause, with premenopausal women having on average 50% less VAT than men and with significantly more gluteal-femoral adipose tissue in women, which may be metabolically protective.74,75 Such a sex difference in visceral adiposity has been shown to largely but not entirely explain the gender gap in cardiometabolic risk variables.76 With age, there is also a selective deposition of VAT that is predictive of the age-related deterioration in the cardiometabolic risk profile,75,77–80 particularly among those who have a family history of visceral obesity.81Table 2. Factors Associated With Increased Visceral AdiposityIncreasing ageSex (men>women)Menopause in womenSmokingNutritional factors (high-caloric diet)Sedentary behaviorRace ↑ Asians ↓ in blacksEthnicity and race are also associated with differences in susceptibility to the selective deposition of VAT.82–87 For instance, blacks are more prone to subcutaneous adiposity than whites or Hispanics, whereas evidence available suggests that Asians may be more prone to visceral fat deposition.82–87 Ethnic and racial differences in visceral body fat deposition are currently an area of intense study.Visceral Adiposity and Metabolic ComplicationsAn important question with considerable clinical implications is whether excess visceral adiposity is causally related to metabolic abnormalities. An extensive discussion of this issue is beyond the scope of this scientific statement, and the reader is referred to several comprehensive reviews on the topic.71,88–91 Currently, 3 main theories have been proposed to explain the relationship between visceral adiposity and metabolic complications: The portal free fatty acid model: Björntorp put forward the hypothesis that in visceral obesity, an uninterrupted overflow of free fatty acid from intra-abdominal or visceral adipocytes would expose the liver to high concentrations, leading to several impairments in hepatic metabolism.92–94 These include reduced extraction and degradation of insulin that exacerbates systemic hyperinsulinemia, reduced degradation of apolipoprotein B that leads to hypertriglyceridemia, and increased hepatic glucose production that leads to impaired glucose tolerance and eventually to T2DM.94,95 Therefore, under this model, one can explain the relationship between excess visceral adiposity and hypertriglyceridemia, hyperapolipoprotein B, hyperinsulinemia, and glucose intolerance that is found in at-risk overweight/obese patients. Although elegant work conducted in dogs supports this model,96 the hypothesis has been under criticism since Jensen et al97–99 provided evidence that most of the free fatty acid found in the portal circulation originates from SAT. Despite the fact that these investigators also found a relationship between visceral adiposity and portal free fatty acid levels coming from the visceral fat depot,98 other scenarios may be involved in the full explanation of the dysmetabolic state of visceral obesity.The “endocrine” function of VAT: Another advance in our understanding of adipose tissue biology was the discovery that adipose tissue is more than a triglyceride storage/mobilization organ. Indeed, numerous potentially important adipose tissue cytokines, commonly referred to as adipokines, could play a role in the dysmetabolic state associated with total/visceral adiposity.100 For instance, leptin, which is produced by adipose cells, has been shown to be better correlated with total and subcutaneous adiposity than with visceral adiposity.101–103 This is why circulating leptin levels are higher in women, who have on average more subcutaneous fat than men.102,104,105 Another adipokine, adiponectin, appears to better reflect visceral than total adiposity.49,52,106 Accordingly, adiponectin levels are generally lower in men than in women, and they are low in viscerally obese individuals and in patients with T2DM.49,107,108 However, a key finding was the observation that hypertrophied adipose tissue is characterized by an infiltration of macrophages, some of which are a major source of inflammatory cytokines such as tumor necrosis factor-α and interleukin-6.109,110 The cytokine interleukin-6 is a major driver of the production of C-reactive protein by the liver.111 Therefore, in viscerally obese patients, the increased production of interleukin-6 by the expanded visceral adipose depot could contribute to expose the liver to high interleukin-6 levels, which could in turn stimulate hepatic C-reactive protein production and impair liver metabolism. Of course, the model is more complicated than the above oversimplification, but several adipokines and the role of the “inflamed” hypertrophied VAT are certainly under the radar screen and are the subject of considerable investigations.Visceral obesity, a marker of dysfunctional adipose tissue leading to ectopic fat deposition: Finally, although visceral adiposity is clearly related to the metabolic abnormalities of overweight/obesity, whether there is a causal relationship between excess visceral adiposity and metabolic complications has been debated. In numerous recent papers and review articles, it has been proposed that excess visceral adiposity may not necessarily impair carbohydrate and lipid metabolism directly but rather may reflect the relative inability of SAT to properly adapt to positive energy balance and to expand by hyperplasia (multiplication of preadipocytes to an increase in the number of adipose cells), creating a “protective metabolic sink.”41,90,91,112 Under this model, a sedentary individual exposed to a surplus of calories would store this extra energy in SAT. To do so, the subcutaneous fat depot would undergo hyperplasia, if need be, to allow the safe storage of this extra energy. However, in situations in which subcutaneous fat could not undergo hyperplasia and therefore would have a limited ability to expand to store the caloric excess, as might occur in the setting of adipose tissue hypoxia,113 these excess triglyceride molecules would accumulate at undesired sites such as liver, heart, pancreas, or skeletal muscle, a phenomenon referred to as “ectopic fat deposition.” Substantial experimental evidence supports the view that excess visceral adiposity is a marker of dysfunctional adipose tissue and of ectopic fat. For instance, women, who have a lot more subcutaneous fat than men, are characterized by lower postprandial lipemia than men because their SAT can better handle the dietary fat load than men.114 In addition, individuals with partial lipodystrophies have more visceral/ectopic fat because of their dysfunctional SAT.115,116 Thiazolidinediones, which improve insulin sensitivity and decrease liver fat, have been shown to induce hyperplasia of SAT, and this is probably a key mechanism explaining how this class of drugs improves glycemia and the cardiometabolic risk profile.117–119 Finally, a negative energy balance induced by diet or by endurance exercise has been shown not only to induce weight loss but also to induce a rapid reduction of liver fat and VAT.120–123 Thus, under circumstances in which the “pressure” for storage of excess triglyceride molecules in SAT is decreased, there will no longer be a need to deposit triglyceride at undesired sites, and ectopic fat depot will be mobilized more readily than subcutaneous fat.Liver Fat as a Key Feature of Ectopic Fat Associated With Dysfunctional Adipose Tissue and Visceral ObesityThe liver plays a central role in the regulation of carbohydrate and lipid/lipoprotein metabolism. Thus, any impairment in liver function is likely to have a major impact on risk factors/markers for prevalent complications such as T2DM and CVD. As for the study of visceral adiposity, the development of imaging techniques such as CT, MRI, and proton magnetic resonance spectroscopy (MRS) has allowed the study of individual differences in liver fat content and its relationship with cardiometabolic risk variables.124–127 First, it has been found that the growing prevalence of obesity has had a major impact on the prevalence of nonalcoholic fatty liver disease,128–130 a condition that could evolve to nonalcoholic steatohepatitis and cirrhosis. Studies that have examined the relationships between body composition, adipose tissue distribution, and liver fat content assessed by MRS have clearly shown that excess visceral adiposity is related to liver fat content even after controlling for total body fat.70 However, liver fat content has generally been found to be more strongly related to insulin resistance and hypertriglyceridemia than visceral adiposity.131 Thus, liver fat is closely related to features of the metabolic syndrome,132 but visceral adiposity is the best adiposity predictor of liver fat content.133 In the landmark Dallas Heart Study conducted on >2000 subjects, ethnic and racial differences (among Hispanic, whites, and blacks) were observed in liver fat content, with blacks having less liver fat than whites and Hispanics.70 However, differences in visceral adiposity were also noted, with blacks having less VAT than the 2 other ethnic and racial groups. A major sex-based difference in the relationship of total adiposity to liver fat content has also been observed: Compared with men, women appear to be relatively protected from the liver fat accumulation expected from excess adiposity. However, this sex difference was entirely attributable to the fact that women had less visceral fat than men. On the other hand, men have greater liver fat content than women, a phenomenon that can be explained entirely by their greater accumulation of VAT compared with women.70Another recently reported international study involving >4500 patients from 29 countries also provides evidence of a strong correlation between visceral adiposity and liver fat content.69 This study also found that the greater accumulation of liver fat in men than in women was entirely accounted for by the greater visceral adiposity of men compared with women. Although such robust and consistent associations cannot be taken as evidence of a causal relationship between visceral adiposity and liver fat (as previously discussed), these observations provide highly concordant evidence that excess liver fat is commonly accompanied by excess visceral adiposity. Other ectopic fat depots (epicardial fat, skeletal muscle, pancreas) are also related to cardiometabolic risk,90,91,134–137 but their specific contribution beyond visceral adiposity and liver fat is not clear. The evidence currently available suggests that excess liver fat is a key central feature predictive of cardiometabolic abnormalities,131 which makes it a priority target for management of complications of overweight/obesity.In summary, excess VAT may be related to cardiometabolic risk in part through a direct mechanism, but we need to keep in mind that another likely scenario is that excess visceral adiposity is a marker of dysfunctional SAT and of ectopic fat deposition. Under this model, 2 key features of ectopic fat may be excess visceral adiposity and liver fat.“Normal-Weight” ObesityRecent studies have suggested that individuals with normal body weight as defined by BMI might still be at risk for metabolic syndrome, insulin resistance, and increased mortality if they have a high body fat content.138,139 A recent report from a sample of individuals representative of the adult US population showed that men of normal weight in the upper tertile of body fat percentage (>23% fat), as measured with electric bioimpedance, were 4 times more likely to have metabolic syndrome and had a higher prevalence of dyslipidemia, T2DM, systemic hypertension, and CVD than those in the lowest tertile.139 Women in the highest tertile of body fat (>33% fat) were 7 times more likely to have metabolic syndrome. Interestingly, women with normal-weight obesity were almost twice as likely to die at follow-up as women in the lowest tertile of body fat. The prevalence of central obesity was low in this group of normal-weight individuals, so these associations were not explained by differences in measures of central obesity between those with normal-weight obesity and control subjects. Studies have also shown that people with normal BMI but enlarged WC have a higher rate of cardiovascular events and death (discussed further below). Although further research is needed to clarify these interesting results, it is clear that subjects with normal weight as defined by BMI may need more detailed classification to better define their adiposity-related risk.Assessing Excess Adiposity: MethodsThis section will review the most accepted methods available both to clinicians and researchers for assessing excess body fat. These methods include those for assessing total body fat mass, distribution of body fat, body composition (percent body fat), and ectopic fat.Assessing Total Body AdiposityBody WeightBefore the use of formulas and tables to adjust body weight for height, the diagnosis of obesity relied on the subjective interpretation of physical appearance and the absolute body weight. The use of weight alone to estimate adiposity, however, is inappropriate, because it fails to consider the fact that body weight is proportional to height, an observation first documented in the 19th century by a Belgian mathematician.140 This relationship, originally known as the Quetelet Index, is now known as BMI. The first attempt to formally diagnose obesity on the basis of body weight indexed to height in modern times was the use of actuarial tables from the Metropolitan Life Insurance Company. These tables were used to estimate ideal weight and then determine the percentage of excess weight.141,142 Because these tables were not based on a simple formula and required the subjective interpretation of an individual's constitution according to normal, thin, and big frame, their use is not practical or reproducible. Thus, a simple body weight is not sufficient in and of itself for the clinical assessment of body fatness.Body Mass IndexBMI, calculated as body weight in kilograms divided by height in meters squared (kg/m2), is one of the most commonly used anthropometric measures to assess for total body adiposity. Because of its simplicity as a measure, it has been used in epidemiological studies and is recommended as a screening tool in the initial clinical assessment of obesity.143,144 Multiple epidemiological studies have demonstrated increased morbidity and mortality with BMI >30 kg/m2.145 Data from the Prospective Studies Collaboration, which analyzed 900 000 adults, demonstrated a 30% increase in all-cause mortality for every increase of 5 U in BMI above a BMI of 25 kg/m2.20Although the utility of BMI has been borne out in epidemiological data, there are limitations to the use of BMI alone to assess for adiposity in clinical practice, particularly among adults with BMI ≤30 kg/m2.146 The numerator in the BMI calculation is “total” body weight and does not distinguish between lean and fat mass. Thus, individuals with normal weight but excess body fat may not be diagnosed as overweight or obese. Conversely, adults with high levels of lean body mass may be misclassified as overweight or obese. Data from the National Health and Nutrition Examination Survey III were analyzed to compare BMI with the World Health Organization criteria for obesity (body fat >25% in men and >35% in