SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Although obesity is a risk factor for mortality, it is unclear whether waist circumference (WC) is a better predictor of mortality than BMI in a clinical setting of patients at high risk for coronary artery disease (CAD). Thus, we compared the association between WC and BMI with all-cause mortality in relation to traditional CAD risk factors in a high-risk cohort. Study population included 5,453 consecutive new patients seen between 1996 and 2005 for management of CAD risk factors in a preventive cardiology clinic. Mortality was determined from the Social Security Death Index. There were 359 deaths over a median follow-up of 5.2 years. Mortality was greater in high (>102 cm in men and >88 cm in women) vs. normal WC in both genders (P < 0.01). The unadjusted Cox proportional hazard ratio (HR) for continuous WC (per cm) was 1.02 (P < 0.001) in both genders and remained significant after adjustment for CAD risk factors (HR = 1.01 in men, HR = 1.03 in women, both P < 0.05). BMI did not associate statistically with mortality. WC associated with diabetes mellitus (DM) and CAD prevalence (P < 0.001). BMI associated only with DM (P < 0.001) and this association disappeared when WC was added to the model. We conclude that WC is an independent predictor of all-cause mortality in a preventive cardiology population. These data affirm the clinical importance of WC measurements for mortality, DM, and CAD risk prediction and suggest that obesity-specific interventions targeting WC in addition to traditional risk factor management may favorably impact these outcomes.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Obesity is a dominant risk factor for type 2 diabetes mellitus (DM) and coronary artery disease (CAD) (1,2) and a major contributor to mortality in the United States (3,4,5). The following BMI categorization—18.5–24.9 kg/m2 = normal weight, 25–29.9 kg/m2 = overweight, and ≥30 kg/m2 = obese—has virtually become the “gold standard” for evaluation and management of obesity. However, some population studies examining the relationship between BMI and mortality have observed a curvilinear (J-shaped) relationship suggesting decreased mortality in the overweight BMI range compared to underweight as well as severely high BMI ranges (4,6,7). Underweight BMI has been associated with increased mortality in the elderly, and with a greater incidence of DM and CAD in some populations (7,8). These observations suggest that obesity is not a homogenous entity; BMI, a convenient clinical measure of overall adiposity, may be inadequate in predicting health risk and mortality.

Unlike BMI, waist circumference (WC) reflects body fat distribution and intra-abdominal adiposity (9,10,11). In the National Health and Nutrition Examination Survey (NHANES) (12), WC has been demonstrated to add to BMIs association with adverse health risk. Population studies (9,10,11,13,14) have consistently found a strong association between abdominal adiposity and features of metabolic syndrome, insulin resistance and inflammation—which play a critical role in the development and progression of DM and CAD (1,15,16,17). Thus, obese individuals are often suboptimally controlled for CAD risk factors despite specific medical therapy (2). It is unclear whether measures of adiposity impact all-cause mortality in a high-risk clinic population treated for traditional CAD risk factors.

The National Heart, Lung, and Blood Institute (NHLBI) recognized that obtaining BMI is only the first step for making a risk assessment in a clinical setting. It has proposed a gender-specific subcategorization combining BMI and WC (using a single cutoff of 102 cm in men and 88 cm in women) for evaluation and management of obesity in clinics (18). However, a recent consensus statement from Shaping America's Health (19) stopped short of recommending WC measurement for routine clinical practice citing lack of evidence into two areas: (i) do WC measurements provide significant incremental value in addition to measuring BMI and other commonly measured clinical risk characteristics? (ii) what are the optimal WC cutoffs that identify increased cardiometabolic risk in the population that would not be otherwise identified by BMI?

We address the above questions by examining data from over 5,000 new patients seen in a preventive cardiology program. The purpose of this investigation was threefold: (i) to compare WC to BMI for prediction of mortality and the prevalence of DM and CAD, (ii) to determine the independent value of BMI and WC in predicting mortality when added to commonly used CAD risk indicators, and (iii) to compare single WC cutoffs (to subcategorize BMI) adopted by NHLBI vs. gender and BMI-specific WC cutoffs proposed by Ardern et al. from NHANES data (20).

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Data collection

The PreCIS database comprised information from >5,000 patients referred to a preventive cardiology clinic for primary and secondary prevention of cardiovascular disease (CVD). Patients referred to this clinic receive standardized, goal-directed, aggressive, pharmacological-based risk reduction interventions. Demographic information, medical history, physical exam, and fasting laboratory data are obtained and entered into an electronic medical record and corresponding database (PreCIS) at each patient's baseline visit. The Institutional Review Board of Cleveland Clinic approved this study.

Subjects and outcomes

For these analyses, we included patients who were seen for their initial visit between 1 January 1995 and 28 August 2006. WC measurements were obtained at the level of the iliac crest. Data with missing WC or BMI were excluded from the analyses. None of the subjects had missing BMI, whereas 57 subjects had missing WC. Patients who were actively smoking at baseline visit were designated as smokers. Patients reporting a history of diabetes and/or use of glycemic medications at the baseline visit or with fasting glucose ≥126 mg/dl were identified as having DM. CAD was defined by a history of coronary artery bypass grafting, percutaneous coronary intervention or myocardial infarction at the baseline visit. History of hypertension and hypercholesterolemia were based on patient's recall at the baseline clinic visit. Mortality data were obtained using Social Security Death Index queries in all patients who had baseline entry visit data.

Statistical analyses

BMI and WC were categorized according to the NHLBI definition (BMI 18.5–24.9 kg/m2 = normal weight, BMI 25–29.9 kg/m2 = overweight and BMI ≥30 kg/m2 = obese; normal WC ≤102 cm in men and ≤88 cm in women). Analyses are presented separately for men and women. Demographic and baseline characteristics are presented as mean ± s.d. (or median ± interquartile range for non-normally distributed data) for continuous variables and as percentages for categorical variables. Comparisons between the WC categories were made using a Student's t-test for continuous variables (or Wilcoxon rank-sum for non-normally distributed data) and χ2-test for categorical data. The Pearson's correlation coefficient was calculated to summarize correlations between BMI and WC. Kaplan–Meier (KM) estimates were generated and compared using the log-rank test to assess time to all-cause mortality across BMI and WC categories. Hazard ratios (HRs) and their corresponding 95% confidence intervals (CIs) were calculated using Cox proportional hazards models. The Cox proportional model was constructed to evaluate the association of continuous BMI and WC (separately and then added simultaneously in the same model) and adjusted for age, race, smoking history, DM, CAD, systolic blood pressure, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol and log triglycerides. Logistic regression was used to generate odds ratios (ORs) with 95% CI to summarize associations with prevalent CAD and DM. Additional comparisons in mortality were made across the gender–BMI specific subclassification of WC proposed by Ardern et al. using NHANES III data (20). They proposed the following cutoffs: Men—(BMI 18.5–24.9 = 87 cm), (BMI 25–29.9 = 98 cm), (BMI 30–34.9 = 109 cm), and (BMI > 35 = 124 cm); Women—(BMI 18.5–24.9 = 79 cm), (BMI 25–29.9 = 92 cm), (BMI 30–34.9 = 103 cm), and (BMI > 35 = 115 cm). To compare which of the two classifications is better associated with mortality, we calculated the net reclassification improvement (NRI) proportion for each gender (21). NRI is the improvement of reclassification using the Ardern et al.'s cutoffs compared to the National Institutes of Health (NIH) standard cutoffs. In addition, model fit characteristics were compared. The likelihood ratio χ2 statistic, which measures goodness of fit, and Akaike information criteria, was assessed for models comparing Ardern et al. and NIH classifications.

All analyses were performed using SAS 8.2 (SAS, Cary, NC). All P values <0.05 were considered to be statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Baseline characteristics

Our population includes 5,453 patients characterized by male predominance and white ethnicity (Tables 1 and 2). Gender-specific characteristics are presented in Tables 1 and 2. Risk factors for CAD, including hypercholesterolemia, hypertension, DM and prior CAD, were prevalent in the population. The mean BMI in both genders was consistent with excess weight. The correlation coefficient between BMI and WC was 0.87 in men and 0.84 in women; 55% men and 43% of women met normal WC criteria (≤102 cm in men and ≤88 cm in women). Baseline cardiovascular (CV) laboratory risk indicators include higher cholesterol, triglyceride, glucose, ratio of apolipoprotein B to A, C-reactive protein, and fibrinogen levels in high vs. normal WC in both genders (Tables 1 and 2). However, LDL levels were not different between WC categories. Additionally, the frequency of DM and hypertension medications usage assessed at the baseline visit increased with increasing categorical BMI. Any DM medication use at baseline was 8% among BMI < 25, 13% among BMI 25–30, 20% among BMI 30–35, and 29% among BMI > 35 (P < 0.001). Similarly any hypertension medication increased with increasing BMI: 47% among BMI < 25, 59% among BMI 25–30, 67% among BMI 30–35, and 74% among BMI > 35 (P < 0.001).

Table 1.  Demographic and baseline characteristics by gender
inline image
Table 2.  Cardiovascular risk factors by gender and waist circumference category
inline image

Prospective observational analysis of adiposity measures and mortality

The observed mortality rate was 7.3% over a median follow-up period of 5.5 years in men and 5.7% over 4.8 years in women.

Categorical adiposity measures and all-cause mortality for men and for women. KM estimates for all-cause mortality were similar among BMI categories in both genders (P = NS). In contrast, WC categorization predicted all-cause mortality in both men (Figure 1a) (KM estimate of mortality = 1.44, CI = 1.12–1.85, P = 0.005) and women (Figure 1b) (KM estimate of mortality = 1.69, CI = 1.13–2.53, P = 0.01). The KM curves for high vs. low WC groups separated statistically within 3 years in women and 6 years in men.

image

Figure 1. Kaplan–Meier curves for all-cause mortality in men (a) and women (b). NS, not significant; WC, waist circumference.

Download figure to PowerPoint

Continuous adiposity measures and all-cause mortality. When WC was used as a continuous measurement in a Cox proportional hazards model, the HR (for 1 cm increase) was 1.02 in both genders (men: CI 1.01–1.03, women: CI 1.01–1.03, both P < 0.001). WC remained associated with excess mortality in both men (HR 1.01, CI 1.00–1.02, P = 0.04) and women (HR 1.01, CI 1.00–1.03, P = 0.04) after adjustment for CV risk factors of age, race, smoking history, DM, CAD, systolic blood pressure, LDL-cholesterol, HDL-cholesterol, and log triglycerides. BMI, as a continuous variable, did not predict a statistically significant mortality risk in men or women (Table 3).

Table 3.  Hazard ratio (confidence interval) for all-cause mortality and odds ratios (confidence intervals) for DM and CAD with continuous BMI alone (per kg/m2), WC alone (per cm), or both BMI and WC concurrently in the model
inline image

BMI and WC included concurrently in mortality models. When BMI and WC were included concurrently in the Cox proportional hazards model, all-cause mortality became inversely related to BMI (Table 3). BMI was also inversely related to adjusted mortality in women (HR 0.89, CI 0.84–0.94, P < 0.001). Conversely, the unadjusted HR for mortality with WC increased to 1.06 in both genders (both P < 0.001). After adjustment for CV risk factors, the HR of WC with mortality persisted in women (HR 1.05, CI 1.03–1.07, P < 0.001) but not in men (HR 1.02, CI 1.00–1.05, P = 0.06).

Comparison of NHLBI and Ardern et al. classifications. We compared NHLBI (18) vs. Ardern et al. (20) WC cutoffs (to subclassify BMI groups) in their prediction of mortality (Table 4). Ardern et al.–proposed gender and BMI-specific WC subclassification resulted in a more uniform distribution of sample sizes within BMI groups than NHLBI's gender-specific subclassification. Within the overweight (BMI 25–29.9) group, KM estimates of mortality were higher in the NHLBI high WC category compared to the normal WC (22.0% vs. 12.1%, P < 0.001, in men and 17.0% vs. 9.6%, P = 0.009, in women). However, the other KM estimates in NHLBI's high WC subclassification did not reach statistical significance. The KM estimates of mortality for Arden et al. high WC subclass were higher in BMI 25–29.9 group as well (KM estimate of mortality for high vs. normal WC = 18.7% vs. 10.6%, P < 0.001, in men and 22.7% vs. 8.7%, P < 0.001, in women). In addition, men in severely obese (BMI ≥ 35) group (22.0% vs. 6.7%, P < 0.001) and women in normal BMI (18.5–24.9) group (19.2% vs. 9.2%, P = 0.027) had significantly higher mortality if their WC was high based on Ardern et al.'s cutoffs. KM estimates of mortality for the rest of Ardern et al.'s high WC subclasses were higher than normal WC as well. However, these did not reach statistical significance.

Table 4.  Kaplan–Meier estimates of mortality (sample size) comparing two BMI and WC classifications
inline image

A NRI for mortality rates was observed with Ardern et al.'s classification compared to NHLBI classification. The NRI with Arden et al. classification was statistically significant for both men (NRI = 9.4%, P = 0.02) and women (NRI = 12.8%, P = 0.02). The likelihood ratio χ2 statistics were also considerably higher for Ardern et al.'s classification in both men and women (both 19.5) compared to a model using NHLBI classification (7.9 in men, 7.0 in women). Akaike information criteria estimates were lower for the models using Ardern et al.'s classification.

Cross-sectional analyses of adiposity measures with prevalence of DM and CAD

Adiposity measures and DM. DM prevalence increased significantly with increasing categorical BMI (comparator group BMI ≤ 25: OR = 1.5 in men, 1.9 in women for BMI 25–29.9; 2.8 in men, 3.5 in women for BMI 30–34.9; 4.3 in men, 5 in women for BMI ≥ 35). Similarly categorical high WC was associated with a higher prevalence of DM (OR = 2.6 in men, 4.3 in women, both P < 0.001). Correspondingly, continuous BMI and WC were both associated with higher prevalence of age-adjusted DM (Table 3). However, when BMI and WC were included in the same age-adjusted DM model, BMI became negatively associated (OR = 0.97, P = 0.05) and WC remained positively associated (OR = 1.06, P < 0.001) with prevalence of DM in women.

Adiposity measures and CAD. Prevalence of CAD was not statistically different among BMI categories in both genders. High WC category had a significant association with CAD (OR = 1.3 in men and 1.5 in women, both P < 0.001). Continuous BMI and WC (Table 3) had a statistically significant OR of 1.02 for age-adjusted prevalence of CAD in women. However, when BMI and WC were included concurrently in the age-adjusted CAD model for women, negative association trend with BMI (OR = 0.97, P = 0.07) and positive association with WC (OR = 1.03, P < 0.0001) was found. CAD did not associate statistically with continuous BMI or WC in men.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

In a cohort of treated preventive cardiology clinic patients, WC predicted all-cause mortality after adjustment for commonly measured CV risk factors. Extrapolating from data presented in Table 3, a 10 cm difference in WC was associated with an increase in all-cause mortality of 20% for both genders, in DM prevalence of 40% and 50%, and CAD prevalence of 10% and 20%, for men and women, respectively. In contrast, BMI did not associate with all-cause mortality and CAD but did associate with DM prevalence. These data support the clinical utility of WC over BMI for mortality as well as DM and CAD risk prediction and support the use of obesity-specific interventions that target WC reduction and abdominal obesity rather than weight loss alone (e.g., exercise training) to improve CV outcomes in high risk populations.

The apparent lack of association of BMI with CAD and mortality demonstrated in this study is consistent with some reports (22,23,24,25), but it conflicts with population studies that have found a curvilinear (J-shaped) relationship between BMI and mortality (7,26). We propose that risk factors associated with CAD that are unrelated to generalized adiposity (e.g., age, smoking, LDL levels) may have negated the effect of BMI on mortality in this population. In addition, treatment of traditional cardiovascular risk factors in our clinic population may have obscured the BMI-mortality association. Flegal et al. proposed that risk factor modification might explain the decreased mortality with increasing excess weight seen in their study comparing mortality rates in NHANES II and III surveys to NHANES I (5,7) and suggested that improvements in public health and medical care over time may have accounted for the decreasing mortality.

Although several organizations include WC in risk assessment algorithms (18,27), a recent consensus statement opposed WC measurement in clinical practice (19). Our analyses extend work by others (12,15,28,29) to support the clinical utility of WC as an independent predictor of mortality as well as show an association with DM and CAD prevalence in a high-risk preventive cardiology clinic population. Various CAD risk factors such as C-reactive protein, interleukin-6, and fibrinogen have been associated with abdominal adiposity in multiple studies (13,14) and likely mediate cardiometabolic risks and mortality associated with WC. A potential clinical implication of these findings is that WC reduction in such high risk patients, as an adjunct to the medical management of risk factors, would further lower cardiovascular events and death. The long-term health impact of an intensive lifestyle intervention in overweight or obese adults with type 2 diabetes are being evaluated in the ongoing LOOK AHEAD trial(30). This trial may highlight the effects of improvement in anthropometric measurements, achieved by intensive lifestyle intervention, on the incidence of major cardiovascular disease events.

An intriguing observation from our analysis was the negative prediction of mortality by BMI, when WC was held constant. Similarly, despite a strong correlation between BMI and WC in their cohort as well, Bigaard et al. (6) observed a negative correlation of BMI with mortality when WC and BMI were included simultaneously in their mortality model. They hypothesized that for persons with the same BMI, WC is a reflection of abdominal fat; whereas for persons with the same WC, BMI is a reflection of lean mass. Additionally, we and others (10,15,28,31) suggest that this negative association of BMI with mortality (when WC is held constant) supports the theoretical belief that an anthropometric “pear-shape” carries a lower cardiometabolic risk than “apple shape” especially in women. Conversely, in our analyses, the HR for WC with mortality increased when BMI was held constant. This interaction of BMI and WC supports the principle of anthropometric subcategorization using BMI and WC together as suggested by NHLBI (18) and Ardern et al. (27).

In our analyses, gender and BMI specific WC cutoffs proposed by Ardern et al. (20) appeared to be a better predictor of mortality than the NIH gender-specific WC categorization (27). Use of Ardern et al.'s subclassification of BMI groups identified men with less mortality risk in the severely obese (BMI ≥ 35) group and women with a higher mortality risk among the normal BMI (18.5–24.9) group. Ardern et al.'s subclassification showed approximately a 10% increase in correct risk classifications in both men and women. This is clinically useful information as subjects may be placed in more correct risk categories and thereby should get the treatment best suited for the actual risk category. We postulate that such a BMI and gender specific WC subcategorization may be a preferable risk assessment tool for clinical use than the gender-only subcategorization adopted by NIH. A monotonic relationship between WC and mortality in our analyses also supports the utility of such an approach.

Some limitations of this study should be acknowledged. First, given its observational nature, the present findings are hypothesis generating. Although the cause of mortality is not established in this population, our assumption is that most of the mortality is related to cardiovascular disease. Indeed, in our analyses, the contribution of WC to mortality and CAD tested in multiple models indicated a strong relationship of WC to both. Secondly, ours is a highly selected self/physician referral population. So, our observations may not be generalizable to all clinical situations. However, this population is quite typical of patients seen in many lipid and cardiology prevention clinics. Hence, our data have clinical relevance to such high-risk patient groups. White patients comprise the majority of patients in our clinic population and hence the results may not be generalizable to other race/ethnic groups. Due to incomplete data, we could not adjust for follow-up confounding variables (systolic blood pressure, cholesterol values) in our mortality models. Instead, we used single-point baseline measurements of these variables in the Cox proportional hazard models. Finally, some of the subgroups using gender–BMI specific WC cutoffs (Table 3) do not reach statistical significance, even though they appear to predict mortality better than gender-specific cutoffs (NIH). We suspect that these analyses may be limited by power in the subgroups of WC within BMI.

We conclude that obesity is an independent risk factor for all-cause mortality in patients treated for traditional risk factors related to CAD; WC is clinically superior to BMI in predicting outcomes. Further studies are needed to assess whether interventions that target WC reduction as a part of a comprehensive intensive risk factor management would lead to improvement in CAD risk and mortality.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

We are grateful to Dr Adi Mehta (Department of Endocrinology, Cleveland Clinic, Cleveland, Ohio) for his thoughtful comments and helpful criticisms of the manuscript at several stages in its development and Drs Charles Faiman and Angelo Licata for their review of the final manuscript. We thank Cindy Stevenson and Brian Hoar for their assistance with data collection and management. S.R.K was supported in part by the National Institutes of Health, National Center for Research Resources (NCRR), Multidisciplinary Clinical Research Career Development Programs Grant 5K12RR023264.

REFERENCES

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES
  • 1
    Mokdad AH, Ford ES, Bowman BA et al. Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA 2003; 289: 7679.
  • 2
    Molenaar EA, Hwang SJ, Vasan RS et al. Burden and rates of treatment and control of cardiovascular disease risk factors in obesity: the Framingham Heart Study Diabetes Care 2008; 31: 13671372.
  • 3
    Allison DB, Fontaine KR, Manson JE, Stevens J, VanItallie TB. Annual deaths attributable to obesity in the United States. JAMA 1999; 282: 15301538.
  • 4
    Calle EE, Teras LR, Thun MJ. Obesity and mortality. N Engl J Med 2005; 353: 21972199.
  • 5
    Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL. Overweight and obesity in the United States: prevalence and trends, 1960–1994. Int J Obes Relat Metab Disord 1998; 22: 3947.
  • 6
    Bigaard J, Tjonneland A, Thomsen BL et al. Waist circumference, BMI, smoking, and mortality in middle-aged men and women. Obes Res 2003; 11: 895903.
  • 7
    Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005; 293: 18611867.
  • 8
    Sairenchi T, Iso H, Irie F et al. Underweight as a predictor of diabetes in older adults: a large cohort study. Diabetes Care 2008; 31: 583584.
  • 9
    Fox CS, Massaro JM, Hoffmann U et al. Abdominal visceral and subcutaneous adipose tissue compartments: association with metabolic risk factors in the Framingham Heart Study. Circulation 2007; 116: 3948.
  • 10
    Kuk JL, Katzmarzyk PT, Nichaman MZ et al. Visceral fat is an independent predictor of all-cause mortality in men. Obesity (Silver Spring) 2006; 14: 336341.
  • 11
    Rosito GA, Massaro JM, Hoffmann U et al. Pericardial fat, visceral abdominal fat, cardiovascular disease risk factors, and vascular calcification in a community-based sample: the Framingham Heart Study. Circulation 2008; 117: 605613.
  • 12
    Janssen I, Katzmarzyk PT, Ross R. Waist circumference and not body mass index explains obesity-related health risk. Am J Clin Nutr 2004; 79: 379384.
  • 13
    Piche ME, Lemieux S, Weisnagel SJ et al. Relation of high-sensitivity C-reactive protein, interleukin-6, tumor necrosis factor-alpha, and fibrinogen to abdominal adipose tissue, blood pressure, and cholesterol and triglyceride levels in healthy postmenopausal women. Am J Cardiol 2005; 96: 9297.
  • 14
    Pou KM, Massaro JM, Hoffmann U et al. Visceral and subcutaneous adipose tissue volumes are cross-sectionally related to markers of inflammation and oxidative stress: the Framingham Heart Study. Circulation 2007; 116: 12341241.
  • 15
    Carr DB, Utzschneider KM, Hull RL et al. Intra-abdominal fat is a major determinant of the National Cholesterol Education Program Adult Treatment Panel III criteria for the metabolic syndrome. Diabetes 2004; 53: 20872094.
  • 16
    Kashyap SR, Defronzo RA. The insulin resistance syndrome: physiological considerations. Diab Vasc Dis Res 2007; 4: 1319.
  • 17
    Vega GL, Adams-Huet B, Peshock R et al. Influence of body fat content and distribution on variation in metabolic risk. J Clin Endocrinol Metab 2006; 91: 44594466.
  • 18
    Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults—the Evidence Report. National Institutes of Health. Obes Res 1998; 6 (Suppl 2): 51S209S.
  • 19
    Klein S, Allison DB, Heymsfield SB et al. 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. Diabetes Care 2007; 30: 16471652.
  • 20
    Ardern CI, Janssen I, Ross R, Katzmarzyk PT. Development of health related waist circumference thresholds within BMI categories. Obes Res 2004; 12: 10941103.
  • 21
    Pencina MJ, D'Agostino RB Sr, D'Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008; 27: 157172.
  • 22
    Baik I, Ascherio A, Rimm EB et al. Adiposity and mortality in men. Am J Epidemiol 2000; 152: 264271.
  • 23
    Diehr P, Bild DE, Harris TB et al. Body mass index and mortality in nonsmoking older adults: the Cardiovascular Health Study. Am J Public Health 1998; 88: 623629.
  • 24
    Price GM, Uauy R, Breeze E, Bulpitt CJ, Fletcher AE. Weight, shape, and mortality risk in older persons: elevated waist–hip ratio, not high body mass index, is associated with a greater risk of death. Am J Clin Nutr 2006; 84: 449460.
  • 25
    Widlansky ME, Sesso HD, Rexrode KM, Manson JE, Gaziano JM. Body mass index and total and cardiovascular mortality in men with a history of cardiovascular disease. Arch Intern Med 2004; 164: 23262332.
  • 26
    Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW Jr. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999; 341: 10971105.
  • 27
    Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2002; 106: 31433421.
  • 28
    Calle EE, Teras LR, Thun MJ. Adiposity and physical activity as predictors of mortality. N Engl J Med 2005; 352: 13811384.
  • 29
    Koster A, Leitzmann MF, Schatzkin A et al. Waist circumference and mortality. Am J Epidemiol 2008; 167: 14651475.
  • 30
    Pi-Sunyer X, Blackburn G, Brancati FL et al. Reduction in weight and cardiovascular disease risk factors in individuals with type 2 diabetes: one-year results of the look AHEAD trial. Diabetes Care 2007; 30: 13741383.
  • 31
    Bigaard J, Frederiksen K, Tjonneland A et al. Body fat and fat-free mass and all-cause mortality. Obes Res 2004; 12: 10421049.