Waist Circumference and Cardiometabolic Risk

Reviews/Commentaries/ADA Statements CONSENSUS STATEMENT

Waist Circumference and Cardiometabolic Risk A Consensus Statement from Shaping America’s Health: Association for Weight Management and Obesity Prevention; NAASO, The Obesity Society; the American Society for Nutrition; and the American Diabetes Association

1 5 SAMUEL KLEIN, MD RUDOLPH L. LEIBEL, MD tant risk factor for obesity-related dis- 2 6 DAVID B. ALLISON, PHD CATHY NONAS, MS, RD, CDE eases. Excess abdominal fat (also known 3 7 STEVEN B. HEYMSFIELD, MD RICHARD KAHN, PHD 4 as central or upper-body fat) is associated DAVID E. KELLEY, MD with an increased risk of cardiometabolic disease. However, precise measurement of abdominal fat content requires the use besity is an important risk factor for plications and mortality rate (e.g., 3,4). of expensive radiological imaging tech- cardiometabolic diseases, including Men and women who have a BMI Ն30 niques. Therefore, waist circumference O diabetes, hypertension, dyslipide- kg/m2 are considered obese and are gen- (WC) is often used as a surrogate marker mia, and coronary heart disease (CHD). erally at higher risk for adverse health of abdominal fat mass, because WC cor- Several leading national and international events than are those who are considered relates with abdominal fat mass (subcuta- institutions, including the World Health overweight (BMI between 25.0 and 29.9 neous and intra-abdominal) (5) and is Organization (WHO) and the National kg/m2) or lean (BMI between 18.5 and associated with cardiometabolic disease Institutes of Health, have provided guide- 24.9 kg/m2). Therefore, BMI has become risk (6). Men and women who have waist lines for classifying weight status based on the “gold standard” for identifying pa- circumferences greater than 40 inches BMI (1,2). Data from epidemiological tients at increased risk for adiposity- (102 cm) and 35 inches (88 cm), respec- studies demonstrate a direct correlation related adverse health outcomes. tively, are considered to be at increased between BMI and the risk of medical com- Body fat distribution is also an impor- risk for cardiometabolic disease (7). These cut points were derived from a regression curve that identified the waist ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● circumference values associated with a BMI Ն30 kg/m2 in primarily Caucasian From the 1Division of Geriatrics and Nutritional Science, Center for Human Nutrition, Washington Uni- versity School of Medicine, St. Louis, Missouri; the 2Clinical Nutrition Research Unit, University of Alabama men and women living in north Glasgow at Birmingham, Birmingham, Alabama; the 3Clinical Research Department, Metabolism, Merck Pharmaceu- (8). tical Company, Rahway, New Jersey; the 4Obesity and Nutrition Research Center, University of Pittsburgh, 5 An expert panel, organized by the Na- Pittsburgh, Pennsylvania; the Naomi Berrie Diabetes Center, Columbia University, New York, New York; tional Heart, Lung and Blood Institute, the 6Obesity and Diabetes Programs, North General Hospital, New York, New York; and the 7American Diabetes Association, Alexandria, Virginia. has recommended that WC be measured Address correspondence and reprint requests to Samuel Klein, MD, Washington University School of as part of the initial assessment and be Medicine, 660 South Euclid Ave., Campus Box 8031, St. Louis, MO 63110. E-mail: [email protected]. used to monitor the efficacy of weight loss Approved for publication 7 March 2007. therapy in overweight and obese patients D.B.A. has received research grants from Frito-Lay and OMP; has served as a consultant to Kraft Foods, Ͻ 2 Pfizer, Bristol-Myers Squibb, and Bio Era; and has received financial support from Lilly, Pfizer, Merck who have a BMI 35 kg/m (7). How- Pharmaceutical Company, Unilever, Coca-Cola, General Mills, International Life Sciences Institute, Glaxo- ever, measurement of WC has not been SmithKline, OMP, Jansen Pharmaceuticals, and Frito-Lay. S.K. has received research grants from Sanofi- widely adopted in clinical practice, and Aventis, Merck, and Takeda for clinical trials; has served as a consultant to Sanofi-Aventis, Amylin the anatomical, metabolic, and clinical Pharmaceuticals, EnteroMedics, Dannon-Yakult, and Merck Pharmaceutical Company. S.B.H. is an em- ployee of Merck Pharmaceutical Company. D.E.K. has received research grants from Novartis Pharmaceu- implications of WC data can be confus- ticals, Sanofi-Aventis, and Pfizer; has served as a consultant/advisor to Novartis Pharmaceuticals, Sanofi- ing. Therefore, Shaping America’s Health: Aventis, Pfizer, Merck Pharmaceutical Company, and GlaxoSmithKline; and has been on speaker’s bureaus Association for Weight Management and for Novartis Pharmaceuticals, Sanofi-Aventis, and Merck Pharmaceutical Company. R.L.L. has received Obesity Prevention; NAASO: The Obesity research grants from GlaxoSmithKline and has been a consultant/advisor to Amylin Pharmaceuticals, Merck Pharmaceutical Company, Arisaph Pharmaceuticals, and Genaera Corporation. C.N. has been a consultant/ Society; and the American Diabetes Asso- advisor to Amylin Pharmaceuticals, GlaxoSmithKline, and Slim Fast. ciation convened a panel, comprised of Abbreviations: CHD, coronary heart disease; CT, computed tomography; IAAT, intra-abdominal adi- members with expertise in obesity man- pose tissue; MRI, magnetic resonance imaging; NHANES III, National Health and Nutrition Examination agement, obesity-related epidemiology, Survey III; SAAT, subcutaneous abdominal adipose tissue; WC, waist circumference; WHO, World Health Organization. adipose tissue metabolic pathophysiol- A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion ogy, statistics, and nutrition science to re- factors for many substances. view the published scientific literature DOI: 10.2337/dc07-9921 and hear presentations from other experts © 2007 by NAASO and the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby in these fields. The Consensus Panel met marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. from December 17 to 20, 2006, in Wash-

DIABETES CARE, VOLUME 30, NUMBER 6, JUNE 2007 1647 Consensus Statement ington, DC, and was charged to provide Table 1—Distribution of adipose tissue mass commonly used to describe intra- answers to the following four questions: in lean and obese men abdominal fat and includes both intra- 1. What does waist circumference peritoneal fat (mesenteric and omental measure? Lean Obese fat), which drains directly into the portal 2. What are the biological mecha- men men circulation, and retroperitoneal fat, which nisms responsible for the associa- drains into the systemic circulation. tion between waist circumference BMI (kg/m2) 23 37 Magnetic resonance imaging (MRI) and cardiometabolic risk? Body weight (kg) 71 116 and computed tomography (CT) are con- 3. What is the power of waist cir- Body fat (%) 15 32 sidered the gold-standard methods for cumference to predict adverse car- Total body fat (kg) 10 37 determining the quantity of subcutaneous diometabolic outcomes? How Total subcutaneous fat (kg) 9 32 abdominal adipose tissue (SAAT) and in- does the predictive power of waist Abdominal fat (kg) 4.3 12.3 tra-abdominal adipose tissue (IAAT) (17). circumference compare with that Subcutaneous (kg) 2.4 7.2 Most MRI and CT methods involve acqui- of BMI? Does measuring waist cir- Intra-abdominal (kg) 1.9 5.1 sition of cross-sectional abdominal im- cumference in addition to BMI im- Intraperitoneal (kg) 1.1 3.5 ages, which are then analyzed for fat prove predictability? Retroperitoneal (kg) 0.8 1.6 content. A single slice is often acquired at 4. Should waist circumference be Adapted from reference 16. the L4-L5 intervertebral level to estimate measured in clinical practice? SAAT and IAAT volume, expressed as 3 cm . However, L4-L5 imaging does not surements are typically taken three times provide the best estimate of total IAAT QUESTION 1: What does waist and recorded to the nearest 0.1 cm. Al- mass, which is more reliably estimated circumference measure? though specific techniques have been rec- several centimeters cephalad of the L4-L5 Measurement technique. Waist cir- ommended for measuring WC in the intervertebral space (17,18). In addition, cumference is actually a perimeter, which clinical setting (2,10), there is no uni- measurement site influences the relation- provides an estimate of body girth at the formly accepted approach. Training tech- ship between IAAT volume and cardio- level of the abdomen. Different anatomi- nicians and even patients to use an metabolic risk; the association between cal landmarks have been used to deter- appropriate technique for measuring WC IAAT volume and presence of the meta- mine the exact location for measuring WC is essential to obtain reliable data; special bolic syndrome is greater when IAAT vol- in different clinical studies, including: 1) tape measures, instructional manuals, ume is determined at the L1-L2 than at the midpoint between the lowest rib and the and videotapes are available for this pur- L4-L5 level (19). Currently, there is no iliac crest; 2) the umbilicus; 3) narrowest pose (11). universally accepted site for measuring (minimum) or widest (maximum) waist The reproducibility of WC measure- IAAT and SAAT. circumference; 4) just below the lowest ments at all sites is high for both men and The relationship between WC, rib; and 5) just above the iliac crest. The women (e.g., iliac crest site, intra-class weight, and BMI can be conceptualized specific site used to measure WC influ- correlation coefficient, r ϭ 0.998 and r ϭ by using simple geometric relationships ences the absolute WC value that is ob- 0.999, respectively) (9,12,13). The corre- that consider the body as a cylinder; WC tained (9). The most commonly used sites lation between technician- and self- is the cylinder’s circumference, height is reported in studies that evaluated the re- measured WC after proper training can its length, and weight is a measure of lationship between morbidity or mortal- also be high for both men (r ϭ 0.95) and mass. Therefore, BMI provides informa- ity rate and WC were at the midpoint women (0.89), respectively (14). How- tion about body volume and mass, and between the lowest rib and the iliac crest ever, self-reported measurements are WC provides information about body (29%), umbilicus (28%), and narrowest prone to a systematic bias, and there is a shape. In general, BMI and WC are highly waist circumference (22%). Although nontrivial underestimate of self-measured correlated, typically with r values in the sites that use an easily identifiable and re- WC at all anatomic sites (15). range of 0.80–0.95 (20), and WC reflects producible landmark (e.g., just above the Anatomical relationships. Adipose tis- both SAAT and IAAT volumes (21). The bony landmark of the lilac crest) might be sue consists of adipocytes, inflammatory relationships among WC, BMI, and adi- more precise and easier to use than other cells, and vascular, connective, and neu- pose tissue compartments in primarily sites, we are not aware of data from any ral tissues. Adipose tissue is distributed Caucasian and African-American men studies that demonstrate an advantage of throughout the body, as large homoge- and women are shown in Table 2 (18). one measurement site over others. neous discrete compartments and as These data demonstrate that both BMI Waist circumference measurements small numbers of cells “marbling” or ad- and WC are strongly correlated with total should be made around a patient’s bare jacent to other tissues. Most adipose tis- body adipose tissue mass but that WC is a midriff, after the patient exhales while sue (ϳ85% of total adipose tissue mass) is better predictor of IAAT than is BMI. standing without shoes, both feet touch- located under the skin (subcutaneous Assessment of WC provides a mea- ing, and arms hanging freely. The measur- fat), and a smaller amount (ϳ15%) is lo- sure of fat distribution that cannot be ob- ing tape should be made of a material that cated within the abdomen (intra- tained by measuring BMI. However, there is not easily stretched, such as fiberglass. abdominal fat) in lean and obese persons is no standardized approach for measur- The tape should be placed perpendicular (Table 1) (16). The relative contribution ing WC and different anatomical land- to the long axis of the body and horizontal of intra-abdominal fat mass to total body marks have been used to measure WC in to the floor and applied with sufficient fat is influenced by sex, age, race/ different studies. Moreover, the measure- tension to conform to the measurement ethnicity, physical activity, and total adi- ment site that provides the best correla- surface. In a research setting, WC mea- posity. The term “visceral fat” is tion with disease risk and best reflects

1648 DIABETES CARE, VOLUME 30, NUMBER 6, JUNE 2007 Klein and Associates

Table 2—Relationships among waist circumference, BMI, and adipose tissue compartments in volved in the association between abdom- men and women inal fat mass and adverse metabolic consequences. Men Women QUESTION 3: What is the power of Waist Waist waist circumference to predict BMI circumference BMI circumference adverse cardiometabolic outcomes? Total adipose tissue 0.82 0.87 0.91 0.87 How does the predictive power of Percent body fat 0.70 0.79 0.86 0.82 waist circumference compare with Total subcutaneous 0.82 0.83 0.91 0.86 that of BMI? Does waist adipose tissue circumference measurement in Total intra-abdominal 0.59 0.79 0.69 0.77 addition to BMI improve adipose tissue predictability? The importance of WC in predicting car- Data are correlation coefficients. Adapted from reference 18. diometabolic risk factors (e.g., elevated blood pressure, dyslipidemia, and hyper- changes in abdominal adipose tissue mass limited ability of subcutaneous fat depots glycemia) and adverse outcomes (e.g., di- has not been established. Nonetheless, to store excess energy results in “over- abetes, CHD, and death rate) has been the precision of WC measurement is high flow” of chemical energy to IAAT and “ec- examined in many large epidemiological at any given landmark. Even self- topic” sites, such as liver and skeletal studies (7,24–33). Specific relative risks measurement can be highly reproducible muscle. Excessive ectopic fat accumula- between WC and these outcomes vary, when performed by properly trained sub- tion then causes metabolic dysfunction in depending on the population sampled jects, although self-measurement results those organs. In fact, increased intrahe- and the outcome measured. The relation- in an underestimation of true WC. Mea- patic fat is associated with dyslipidemia ship between WC and clinical outcome is surement of WC cannot determine the in- and hepatic insulin resistance (23), and consistently strong for diabetes risk, and dividual contributions of SAAT and IAAT increased intramyocellular fat is associ- WC is a stronger predictor of diabetes to abdominal girth, which require imag- ated with skeletal muscle insulin resis- than is BMI. The relative risk of developing by MRI or CT. The value of these scan- tance (24). In this paradigm, IAAT is ing diabetes between subjects in the high- ning techniques in clinical practice has primarily a marker of the magnitude of est and lowest categories of reported WC not been determined. overflow of fatty acids from subcutaneous often exceeds 10 and remains statistically depots. Therefore, increased WC could be a significant after adjusting for BMI. These QUESTION 2: What are the discernible marker of a system-wide im- data demonstrate that WC can identify biological mechanisms responsible pairment in energy storage regulation, in persons who are at greater cardiometa- for the association between waist which an increase in IAAT reflects a re- bolic risk than those identified by BMI circumference and metabolic and duced capacity for energy storage in other alone. Values for WC are also consistently cardiometabolic risk? adipose tissues. A third hypothesis pro- related to the risk of developing CHD, It is not known whether the storage of an poses a direct effect of omental and mesen- and the relative risk of developing CHD absolute or relative excess amount of tri- teric adipose tissue depots on insulin between subjects in the highest and low- glycerides in abdominal fat depots is di- resistance, lipoprotein metabolism, and est categories of WC ranges from 1.5 to rectly responsible for increased disease blood pressure. Metabolic products of 2.5 and remains statistically significant af- risk or whether such deposition is simply omental and mesenteric adipose tissue deter adjusting for BMI. Values for WC are associated with other processes that cause pots are released into the portal vein, which usually strongly associated with all-cause risk, or both. In addition, WC values pro- provides direct delivery to the liver. Lipoly- and selected cause-specific mortality vide a measure of both SAAT and IAAT sis of omental and mesenteric adipose tissue rates. Data from several studies support masses. Therefore, the relationship be- triglycerides release free fatty acids that can the notion that WC is an important pre- tween WC and cardiometabolic risk can- induce hepatic insulin resistance and dictor of diabetes, CHD, and mortality not determine whether risk is associated provide substrate for lipoprotein synthesis rate, independent of traditional clinical with SAAT, IAAT, or both. and neutral lipid storage in hepatocytes. In tests, such as blood pressure, blood glu- The mechanism(s) responsible for the addition, specific proteins and hormones cose, and lipoproteins (7,26). However, relationship between excess abdominal produced by omental and mesenteric adi- there is not yet a compelling body of evi- fat distribution and cardiometabolic dis- pose tissue, such as inflammatory adipo- dence demonstrating that WC provides ease is not known, but several hypotheses kines, angiotensinogen, and cortisol clinically meaningful information that is have been proposed. One of the earliest (generated by local activity of 11 ␤-hydrox- independent of well-known cardiometa- hypotheses that implicated IAAT as a met- ysteroid dehydrogenase), can also contrib- bolic risk factors. abolic risk factor suggested that activation ute to cardiometabolic disease. A fourth The relationships between WC and of the central nervous system–adrenal hypothesis is that genes that predispose to health outcomes are affected by demo- axis by environmental stressors caused preferential deposition of fat in abdominal graphic variables, including sex, race/ both the preferential deposition of adi- depots independently cause cardiometa- ethnicity, and age. WC is an important pose tissue in the trunk and the cardio- bolic disease. predictor of health outcomes in men and vascular and metabolic disorders These hypotheses are not mutually women; Caucasians, African Americans, associated with that deposition (22). exclusive, and it is possible that all, and Asians, and Hispanics; and adults of all More recently, it has been suggested that a other unknown mechanisms, are in- age-groups. In fact, the relationship be-

DIABETES CARE, VOLUME 30, NUMBER 6, JUNE 2007 1649 Consensus Statement tween WC and health outcome changes incremental value in these predictions clinical management if NHLBI obesity much less with increasing age than does above and beyond that offered by BMI treatment guidelines are followed? An- the relationship between BMI and health and commonly evaluated cardiometa- swer: Probably not. outcome (31). However, it is not known bolic risk factors, such as blood glucose Measurement of WC in clinical prac- whether WC can provide a better assess- concentration, lipid profile and blood tice is not trivial, because providing this ment of health risk in one sex, racial/ pressure? Answer: Uncertain. assessment competes for the limited time ethnic group, or age category than Data from many large population available in a busy office practice and re- another. studies have found waist circumference to quires specific training to ensure that re- The shape of the relationship between be a strong correlate of clinical outcome, liable data are obtained. Therefore, waist WC and health outcomes (e.g., linear, particularly diabetes, and to be indepen- circumference should only be measured if monotonic, step-function, or U-shaped) dent of BMI. In addition, data from a lim- it can provide additional information that influences the WC value that can most ited number of studies demonstrates that influences patient management. Based on efficiently distinguish between “normal” WC remains a predictor of diabetes, NHANES III data, 99.9% of men and and “abnormal” and serve as a basis for CHD, and mortality rate, even after ad- 98.4% of women would have received the considering clinical action. Data from justing for BMI and several other cardio- same treatment recommendations pro- most studies suggest that the shape of the metabolic risk factors. Additional studies posed by the NHLBI Expert Panel by eval- relationship between WC and health out- are needed to confirm that WC remains uating BMI and other cardiovascular risk come lends itself to identifying clinically an independent predictor of risk. factors, without an assessment of WC meaningful cut point values because risk 3. Do the current definitions used to (37). However, it is likely that different often accelerates monotonically above, determine a high WC identify a nontrivial WC cut point values could provide more and can be relatively flat below, a specific number of patients who are at increased useful clinical information. For example, WC value. Optimum WC cut points will cardiometabolic risk, but who would not an analysis of data obtained from the likely vary according to the population otherwise be identified by having a BMI NHANES III and the Canadian Heart studied, the health outcome of interest, Ն25 kg/m2 and an assessment of com- Health Surveys found that BMI-specific and demographic factors. monly evaluated cardiometabolic risk fac- WC cut points provided a better indicator Data from most clinical weight loss tors? Answer: Yes. of cardiometabolic risk than the recom- and exercise training trials have shown The recommended WC thresholds mended WC thresholds (35). For nor- that reductions in WC occur concomi- for increased cardiometabolic risk in men mal-weight (BMI 18.5–24.9 kg/m2), tantly with reductions in obesity-related (Ͼ40 inches [102 cm]) and women (Ͼ35 overweight (BMI 25.0–29.9 kg/m2), class cardiometabolic risk factors and disease. inches [88 cm]) were derived from WC I obesity (BMI 30.0–34.9 kg/m2), and However, these results do not prove that values that correlated with a BMI Ն30 class II/class III obesity (BMI Ն35.0 kg/ the reduction in WC was responsible for kg/m2 (2). The National Health and Nu- m2), the optimal WC cut points were 87, the beneficial effect on health outcome. trition Examination Survey III (NHANES 98, 109, and 124 cm in men and 79, 92, Additional studies are needed to evaluate III) found that about 14% of women and 103, and 115 cm in women, respectively. the effect of decreasing WC on cardio- about 1% of men had a “high” WC but a Therefore, it is possible that WC measure- metabolic outcomes. normal BMI (18.5–24.9 kg/m2) (36). In ment could be an effective clinical tool for addition, ϳ70% of women who were identifying “metabolically obese, lean” QUESTION 4: Should waist overweight (BMI 25.0–29.9 kg/m2) had a patients who might benefit from lifestyle circumference be measured in WC Ͼ35 inches and ϳ25% of men who therapy but would not have been consid- clinical practice? were overweight had a WC Ͼ40 inches. ered for treatment because of a normal The panel concluded that determining An estimate based on data available from BMI. Waist circumference could also whether waist circumference should be the WHO Monica Project, conducted in identify “metabolically normal, obese” measured in clinical practice depends on more than 32,000 men and women from subjects who do not require aggressive the responses to the following four key Europe, Australia, and New Zealand, sug- obesity therapy because they do not have questions: gest that about 10% of participants who a marked increase in cardiometabolic 1. Can waist circumference be reli- had BMI values Ͻ30 kg/m2 hadaWC risk. ably measured? Answer: Yes. above the recommended cut points for in- Health care personnel and even pa- creased risk (36). It is not known what CONCLUSIONS tients themselves, who are given appro- portion of subjects who had a large WC Waist circumference provides a unique priate training in technique, can provide would have been identified as having in- indicator of body fat distribution, which highly reproducible measurements of creased cardiometabolic risk based on can identify patients who are at increased WC in men and women. However, it is findings from a standard medical evalua- risk for obesity-related cardiometabolic not know whether measurement of one tion. Therefore, the optimal WC criteria disease, above and beyond the measure- anatomical site is a better indicator of car- needed to identify patients at increased ment of BMI. However, the current WC diometabolic risk than measurement at risk of metabolic disease, who would oth- cut points recommended to determine other sites. erwise not be identified by evaluating BMI health risk (2) were derived by regression 2. Does waist circumference provide: and/or other standard cardiometabolic from an “obese” BMI and are unlikely to a) good prediction of diabetes, CHD, and risk factors, is not known and will likely affect clinical management when BMI and mortality rate? Answer: Yes; b) incremen- require adjustments based on BMI, sex, other obesity-related cardiometabolic risk tal value in predicting diabetes, CHD, and age, and race/ethnicity. factors are already being determined. mortality rate above and beyond that pro- 4. Would assessment of WC in pa- Therefore, the clinical usefulness of mea- vided by BMI? Answer: Yes; c) sufficient tients who have a BMI Ն25 kg/m2 affect suring WC, when risk is based on the cur-

1650 DIABETES CARE, VOLUME 30, NUMBER 6, JUNE 2007 Klein and Associates rently accepted guidelines, is limited. Evans DJ, Hartz AJ, Kalkhoff RK, Adams inal subcutaneous adipose tissue alter as- However, WC measurement can some- PW: Relation of body fat distribution to sociations with the metabolic syndrome? times provide additional information to metabolic complications of obesity. J Clin Diabetes Care 29:679–684, 2006 help the clinician determine which pa- Endocrinol Metab 54:254–260, 1982 20. Ford ES, Mokdad AH, Giles WH: Trends tients should be evaluated for the pres- 7. Wang Y, Rimm EB, Stampfer MJ, Willett in waist circumference among U.S. adults. WC, Hu FB: Comparison of abdominal Obes Res 11:1223–1231, 2003 ence of cardiometabolic risk factors, such adiposity and overall obesity in predicting 21. Chan DC, Watts GF, Barrett PH, Burke V: as dyslipidemia, and hyperglycemia. In risk of type 2 diabetes among men. Am J Waist circumference, waist-to-hip ratio addition, measuring WC can be useful in Clin Nutr 81:555–563, 2005 and body mass index as predictors of ad- monitoring a patient’s response to diet 8. Lean ME, Han TS, Morrison CE: Waist ipose tissue compartments in men. QJM and exercise treatment because regular circumference as a measure for indicating 96:441–447, 2003 aerobic exercise can cause a reduction in need for weight management. BMJ 311: 22. Bjorntorp P: Body fat distribution, insulin both WC and cardiometabolic risk, with- 158–161, 1995 resistance, and metabolic diseases. Nutri- out a change in BMI (38). Further studies 9. Wang J, Thornton JC, Bari S, Williamson tion 13:795–803, 1997 are needed to establish WC cut points that B, Gallagher D, Heymsfield SB, Horlick 23. Seppala-Lindroos A, Vehkavaara S, Hak- can assess cardiometabolic risk, not ade- M, Kolter D, Laferrere B, Mayer L, Pi-Su- kinen AM, Goto T, Westerbacka J, Sovi- nyer FX, Pierson RN Jr: Comparison of jarvi A, Halavaara J, Yki-Jarvinen H: Fat quately captured by BMI and routine clin- waist circumferences measured at 4 sites. accumulation in the liver is associated ical assessments. Selection of the most Am J Clin Nutr 77:379–384, 2003 with defects in insulin suppression of glu- appropriate WC values will be complex 10. Douketis JD, Paradis G, Keller H, Martin- cose production and serum free fatty acids because they are likely influenced by sex, eau C: Canadian guidelines for body independent of obesity in normal men. race/ethnicity, age, BMI, and other fac- weight classification in adults: application J Clin Endocrinol Metab 87:3023–3028, tors. Nonetheless, it should be possible to in clinical practice to screen for over- 2002 determine more useful WC cut points weight and obesity and to assess disease 24. Sinha R, Dufour S, Petersen KF, LeBon V, than are currently recommended, by risk. CMAJ 172:995–998, 2005 Enoksson S, Ma YZ, Savoye M, Rothman carefully reviewing published data and 11. NHANES III Anthropometric Procedures DL, Shulman GI, Caprio S: Assessment of reevaluating datasets available from exist- Video. Available from http://www.cdc. skeletal muscle triglyceride content by ing population studies. These additional gov/nchs/about/major/nhanes/avideo.htm. (1)H nuclear magnetic resonance spec- 12. Nordhamn K, Sodergren E, Olsson E, troscopy in lean and obese adolescents: analyses will define the future role of WC Karlstrom B, Vessby B, Berglund L: Reli- relationships to insulin sensitivity, total measurement in clinical practice. ability of anthropometric measurements body fat, and central adiposity. Diabetes in overweight and lean subjects: conse- 51:1022–1027, 2002 quences for correlations between anthro- 25. Katzmarzyk PT, Craig CL: Independent Acknowledgments— This conference was pometric and other variables. Int J Obes effects of waist circumference and physi- supported in part by an educational grant from Relat Metab Disord 24:652–657, 2000 cal activity on all-cause mortality in Cana- the Campbell Soup Company. 13. Chen MM, Lear SA, Gao M, Frohlich JJ, dian women. Appl Physiol Nut Metal 31: Birmingham CL: Intraobserver and inter- 271–276, 2006 observer reliability of waist circumference 26. Yusuf S, Hawken S, Ounpuu S, Bautista L, References and the waist-to-hip ratio. Obes Res 9:651, Franzosi MG, Commerford P, Lang CC, 1. World Health Organization: Obesity: Pre- 2001 Rumboldt Z, Onen CL, Lisheng L, Tan- venting and Managing the Global Epidemic: 14. Rimm EB, Stampfer MJ, Colditz GA, omsup S, Wangai P Jr, Razak F, Sharma Report of a WHO Consultation on Obesity. Chute CG, Litin LB, Willett WC: Validity AM, Anand SS; INTERHEART Study In- Geneva, World Health Organization, 1998 of self-reported waist and hip circumfer- vestigators: Obesity and the risk of myo- 2. National Institutes of Health, National ences in men and women. Epidemiology cardial infarction in 27,000 participants Heart, Lung, and Blood Institute: Clinical 1:466–473, 1990 from 52 countries: a case-control study. guidelines on the identification, evalua- 15. Bigaard J, Spanggaard I, Thomsen BL, Lancet 366:1640–1649, 2005 tion, and treatment of overweight and Overvad K, Tjonneland A: Self-reported 27. Hu G, Tuomilehto J, Silventoinen K, obesity in adults—the evidence report. and technician-measured waist circum- Barengo N, Jousilahti P: Joint effects of Obes Res 6 (Suppl. 2):51S–209S, 1998 ferences differ in middle-aged men and physical activity, body mass index, waist 3. Colditz GA, Willett WC, Rotnitzky A, women. J Nutr 135:2263–2270, 2005 circumference and waist-to-hip ratio with Manson JE: Weight gain as a risk factor for 16. Abate N, Garg A, Peshock RM, Stray-Gun- the risk of cardiovascular disease among clinical diabetes mellitus in women. Ann dersen J, Grundy SM: Relationships of middle-aged Finnish men and women. Intern Med 122:481–486, 1995 generalized and regional adiposity to in- Eur Heart J 24:2212–2219, 2004 4. Calle EE, Thun MJ, Petrelli JM, Rodriguez sulin sensitivity in men. J Clin Invest 96: 28. Lofgren I, Herron K, Zern T, West K, Pata- C, Heath CW Jr: Body-mass index and 88–98, 1995 lay M, Shachter NS, Koo SI, Fernandez mortality in a prospective cohort of U.S. 17. Shen W, Wang Z, Punyanyita M, Lei J, ML: Waist circumference is a better pre- adults. N Engl J Med 341:1097–1105, Sinav A, Kral JG, Imielinska C, Ross R, dictor than body mass index of coronary 1999 Heymsfield SB: Adipose tissue quantifica- heart disease risk in overweight premeno- 5. Pouliot MC, Despres JP, Lemieux S, tion by imaging methods: a proposed pausal women. J Nutr 134:1071–1076, Moorjani S, Bouchard C, Tremblay A, classification. Obes Res 11:5–16, 2003 2004 Nadeau A, Lupien PJ: Waist circumfer- 18. Shen W, Punyanitya M, Wang Z, Gal- 29. Suk SH, Sacco RL, Boden-Albala B, ence and abdominal saggital diameter: lagher D, St-Onge MP, Albu J, Heymsfield Cheun JF, Pittman JG, Elkind MS, Paik best simple anthropometric indices of ab- SB: Total body skeletal muscle and adi- MC; Northern Manhattan Stroke Study: dominal visceral adipose tissue accumu- pose tissue volumes: estimation from a Abdominal obesity and risk of ischemic lation and related cardiovascular risk in single abdominal cross-sectional image. stroke: the Northern Manhattan Stroke men and women. Am J Cardiol 73:460– J Appl Physiol 97:2333–2338, 2004 Study. Stroke 34:1586–1592, 2003 468, 1994 19. Kuk JL, Church TS, Blair SN, Ross R: Does 30. Carey VJ, Walters EE, Colditz GA, So- 6. Kissebah AH, Vydelingum N, Murray R, measurement site for visceral and abdom- lomon CG, Willett WC, Rosner BA,

DIABETES CARE, VOLUME 30, NUMBER 6, JUNE 2007 1651 Consensus Statement

Speizer FE, Manson JE: Body fat distribu- the Iowa Women’s Health Study. Arch In- of type 2 diabetes in the general population and risk of non-insulin-dependent tern Med 160:2117–2128, 2000 tion: are there differences between men diabetes mellitus in women: the Nurses’ 33. Hu FB, Willett WC, Li T, Stampfer MJ, and women? The MONICA/KORA Augs- Health Study. Am J Epidemiol 145:614– Colditz GA, Manson JE: Adiposity as burg cohort study. Am J Clin Nutr 84:483– 619, 1997 compared with physical activity in pre- 489, 2006 31. Visscher TL, Seidell JC, Molarius A, van dicting mortality among women. N Engl 37. Kiernan M, Winkleby MA: Identifying pa- der Kuip D, Hofman A, Witteman JC: A J Med 351:2694–2703, 2004 tients for weight-loss treatment: an empir- comparison of body mass index, waist- 34. Baik I, Ascherio A, Rimm EB, Giovan- ical evaluation of the NHLBI Obesity hip ratio and waist circumference as pre- nucci E, Spiegelman D, Stampfer MJ, Wil- Education Initiative Expert Panel treat- dictors of all-cause mortality among the lett WC: Adiposity and mortality in men. ment recommendations. Arch Intern Med elderly: the Rotterdam study. Int J Obes Am J Epidemiol 152:264–271, 2000 160:2169–2176, 2000 Relat Metab Disord 25:1730–1735, 2001 35. Ardern CI, Janssen I, Ross R, Katzmarzyk 38. Dekker MJ, Lee S, Hudson R, Kilpatrick K, 32. Folsom AR, Kushi LH, Anderson KE, PT: Development of health-related waist Graham TE, Ross R, Robinson LE: An exer- Mink PJ, Olson JE, Hong CP, Sellers TA, circumference thresholds within BMI cat- cise intervention without weight loss de- Lazovich D, Prineas RJ: Associations of egories. Obes Res 12:1094–1103, 2004 creases circulating interleukin-6 in lean and general and abdominal obesity with mul- 36. Meisinger C, Doring A, Thorand B, Heier obese men with and without type 2 diabetes tiple health outcomes in older women: M, Lowel H: Body fat distribution and risk mellitus. Metabolism 56:332–338, 2007

1652 DIABETES CARE, VOLUME 30, NUMBER 6, JUNE 2007