| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta
2 Address reprint requests to HS Kahn, Mail Stop K-10, CDC, 4770
Buford Highway NE, Atlanta, GA 30341-3717. E-mail: hkahn{at}csiro.au
See corresponding editorial on page 902
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: Using thresholds for EW and ET observed among the youngest adults, we estimated for all adults the prevalence of combined EW and ET (EWET) and described the metabolic risks associated with EWET.
Design: In a cross-sectional, weighted sample of 9183 adults, we
used two-dimensional displays to provide thresholds for EW (men:
95 cm; women: 88 cm) and fasting ET ( 1.45 mmol/L) and
estimated the characteristics of EWET among adults of all ages.
Results: The population prevalence of EWET in 18-24-y-olds was 6%; it rose with age until age 55-74 y (prevalence: 43%) and then was lower among the elderly. Persons with EWET were more likely (P < 0.0001) to have adverse mean (± SEE) concentrations of risk variables in adjusted analyses (fasting insulin: 43 ± 3 pmol/L; HDL cholesterol: -0.27 ± 0.02 mmol/L; apolipoprotein B: 0.20 ± 0.01 g/L; fasting glucose: 0.71 ± 0.07 mmol/L; uric acid: 50 ± 2 µmol/L) and to have diabetes (relative risk: 3.2) than were persons without EWET. Compared with a similar-size subpopulation with high body mass index, persons with EWET were older and had more dyslipidemia, hyperglycemia, and hyperuricemia. Compared with "metabolic syndrome," EWET identified more persons who were younger and had greater LDL-cholesterol and apolipoprotein B concentrations. Compared with "prediabetes," EWET identified more persons with hyperinsulinemia, dyslipidemia, and hyperuricemia.
Conclusions: EWET identifies a syndrome of lipid overaccumulation associated with metabolic risk and accelerated mortality after middle age. Prospective studies should evaluate this simple indicator.
Key Words: Adult • anthropometry • body mass index • glucose intolerance • hyperinsulinemia • hyperlipidemia • hypertension • hyperuricemia • insulin resistance • metabolic syndrome X • metabolic diseases • obesity • risk assessment
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
For this population-based study, we postulated that a state of lipid overaccumulation could be recognized as a combination of abdominal enlargement and elevated concentration of circulating triacylglycerols (12). We established normative threshold values for an enlarged waist (EW) circumference and elevated triacylglycerols (ET) by assessing only the youngest adults, because abdominal size and triacylglycerol concentration increase with adult age and metabolic risk (13-16). Accordingly, establishing the thresholds with values from any subpopulation but the youngest adults would be inappropriate. We defined our category of lipid overaccumulation as the occurrence in one person of both EW and ET (EWET). Using the thresholds derived from the youngest adults, we examined the associations of EWET with hyperinsulinemia, insulin resistance, dyslipidemia, hyperglycemia, hyperuricemia, and hypertension among adults through age 90 y.
We report here the testing of cross-sectional relations in
adulthood between our marker of lipid overaccumulation
(EWET) and continuous metabolic risk variables. We also
compare the estimated subpopulation identified by EWET with
other estimated subpopulations identified by 3 alternative, dichotomous risk markers: 1) high body mass index (BMI; in
kg/m2): a BMI above a threshold chosen to yield an estimated
subpopulation similar in size to the subpopulation with EWET;
2) metabolic syndrome: 3 of 5 risk factors defined by the
Third Report of the Adult Treatment Panel (MS-ATPIII) of the
National Cholesterol Education Program (5, 17); and 3) prediabetes: impaired fasting glucose (IFG) or impaired glucose
tolerance (IGT), or both, shown by a 2-h oral-glucose-tolerance
test (18).
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Serum triacylglycerol concentrations were measured enzymatically after hydrolysis to glycerol (Hitachi 704 Analyzer;
Boehringer Mannheim, Indianapolis); the CV was 3-5% over
the study and across the clinical range. Additional details of all
laboratory procedures are available elsewhere (22). Calculations of LDL-cholesterol concentrations were limited to participants with triacylglycerol concentration < 4.5 mmol/L (a
requirement of the Friedewald equation; 23) who fasted 9 h.
Analyses for serum apolipoprotein B were conducted only
during phase 1 of NHANES III (1988-1991).
Laboratory values for insulin, glucose, and glycated hemoglobin were omitted for 63 participants who described themselves as current insulin users. For participants with valid
fasting values of insulin and glucose, we estimated insulin
resistance by using the homeostasis model assessment of insulin resistance (24), defined as fasting insulin (pmol/L) x fasting glucose (mmol/L)/22.5. Participants 40-74 y old, with
the exception of insulin users, underwent a 2-h oral-glucose-tolerance test with 75 g glucose. On the basis of self-reported
histories of diabetes and the oral-glucose-tolerance test, we
classified these participants as having diabetes (nongestational,
unspecified type), isolated IFG (6.1 
fasting glucose < 7.0
mmol/L), isolated IGT (7.8 2-h glucose < 11.1 mmol/L),
combined IFG and IGT, or normal glucose tolerance (25).
Statistical analysis
We used NHANES III sampling weights along with SAS/Graph (release 8.2; SAS Institute Inc, Cary, NC), and SUDAAN
(release 8.0; Research Triangle Institute, Research Triangle, NC)
software to estimate the represented population sizes, the prevalences of EWET, and the distributions of associated risk variables.
Our analyses thereby incorporated weights that accounted for
unequal selection probabilities (clustered design, planned oversampling, and differential nonresponse) (26). On the basis of the
weights assigned, we estimated that our analytic sample represented a total of 100 113 964 US adults aged 18-90 y, 50.5% of
whom were women.
We prepared sex-specific, age-stratified, bubble plots of
population distribution by the values for waist circumference
(to nearest cm) and triacylglycerol concentration (to nearest
mmol/L) (Figure 1
). The area of each circle on these plots is
proportional to the estimated number of US men or women
represented by those intersections. We established normative
thresholds for waist circumference and triacylglycerol concentration by visual inspection of the plots restricted to subjects
aged 18-24 y (Figure 1, A and B). For both sexes we noted
predominant, rounded clusters in the lower left region that
appeared to end just below waist threshold values of 95 cm
(37.4 inches) for men and 88 cm (34.6 inches) for women and
just below a triacylglycerol threshold of 1.45 mmol/L (128
mg/dL) for both sexes. We applied these threshold boundaries
to all of our sex-specific, two-dimensional, graphic displays,
thus establishing consistent quadrants for the plots of estimated
persons in each age range. Subpopulations of persons whose
values were above both thresholds (ie, located in the upper
right quadrant) were defined as having EWET.
|
For the comparison of EWET with each alternative dichotomous marker, we estimated the 2 subpopulations that were discordant for marker status (ie, EWET-positive with negative alternative marker; EWET-negative with positive alternative marker). In the 2 discordant subpopulations, we compared prevalences (unadjusted) of adverse risk levels and then compared the mean values of risk variables with adjustments for sex, age, and race or ethnicity.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). For both sexes combined,
at age 18-90 y, the estimated prevalence of EWET was 24.8 ±
0.9%. We also estimated prevalences of 19.6 ± 0.5% for EW with
low triacylglycerol concentration, 12.3 ± 0.6% for low waist
circumference with ET, and 43.3 ± 1.1% for low waist circumference with low triacylglycerol concentration.
|
). These adverse differences were most pronounced
for insulin, homeostasis model assessment of insulin resistance
(Figure 3), and uric acid irrespective of age or sex. The risk
variables related to lipids, glucose, and blood pressure also had
strong associations with EWET, although they were less consistent by age and sex (data not shown).
|
|
29.25 identified a subpopulation (high
BMI) similar in size (24.9 million) to that identified by EWET
(24.8 million). A majority within the estimated population was
concordant for EWET and high BMI: 14.3 ± 0.7% met the
criteria for both conditions, and 64.6 ± 0.9% were concordant
by meeting neither criterion. The discordant subpopulations
were of equal size (Table 2), but the subpopulation with
EWET alone was older (
age: 54.7 y compared with 42.6 y;
P < 0.0001).
|
). For all lipid
variables (including LDL cholesterol; P = 0.017), fasting
glucose (P = 0.011), and glycated hemoglobin (P = 0.022),
the adjusted mean values were higher for persons with EWET
alone. For fasting insulin and systolic blood pressure (P =
0.014), the adjusted mean values were higher for persons with
high BMI alone.
EWET compared with the metabolic syndrome
There were 8730 adult NHANES III participants with sufficient data (fasting triacylglycerol, fasting glucose, blood pressure, sex-specific waist circumference, and sex-specific HDL
cholesterol) to allow the identification of either EWET or
MS-ATPIII (5), and this sample represented an adult population of 95.5 million. The prevalence of MS-ATPIII was 23.1 ±
1.0%. Most of this population was concordant for EWET and
MS-ATPIII, including 16.0 ± 0.8% who met the criteria for
both conditions and 67.5 ± 1.1% who met neither criterion. Of
those who were discordant for EWET and MS-ATPIII, the
subpopulation with EWET alone was larger (Table 2) and
younger ( age: 48.4 y compared with 53.1 y; P = 0.0016) than
was the group with MS-ATPIII alone.
EWET alone identified more persons with high-risk concentrations of LDL cholesterol and apolipoprotein B than did
MS-ATPIII alone (Table 2), and the adjusted mean concentrations were higher for these 2 variables (for LDL cholesterol:
P = 0.024). However, HDL cholesterol, fasting glucose, blood
pressure (3 of the 5 risk variables that define MS-ATPIII),
fasting insulin, insulin resistance, and glycated hemoglobin
were more likely to be adverse in the group with MS-ATPIII
alone. The adjusted mean value for the ratio of total to HDL
cholesterol was similar (P = 0.10) in these 2 subpopulations.
EWET in relation to diabetes and prediabetes
Participants 40-74 y old who either reported having diabetes
or underwent an oral-glucose-tolerance test (n = 3800) were
estimated to total 42.4 million adults. Among those with
EWET, an estimated 25.4 ± 1.7% had diabetes; among those
without EWET, the estimated diabetes prevalence was 8.0 ±
0.8% (relative risk: 3.2; 95% CI: 2.5, 4.0). The relative risk for
diabetes was lower for men (2.6; 95% CI: 1.9, 3.6) than for
women (4.0; 95% CI: 3.0, 5.3). Among the persons with
diabetes, 64.7 ± 2.7% were estimated to have EWET; of the
nondiabetic adults in the same age range, an estimated 31.9 ±
1.2% had EWET.
After excluding the 6.1 million persons with diabetes, we
estimated that 8.9 million had prediabetes, a dysglycemic state
defined by IFG, IGT, or combined IFG and IGT (18). Of those
with prediabetes, an estimated 4.0 million also had EWET (a
concordant subpopulation). Of the 27.4 million with normal
glucose tolerance (ie, without prediabetes), 19.8 million also
did not have EWET (concordant). The discordant subpopulation with EWET alone was larger than that with prediabetes
alone (Table 2), but these discordant subpopulations were of
similar mean age (54.3 y compared with 55.7 y; P = 0.19).
The estimated subpopulation with EWET alone had higher
unadjusted prevalences (Table 2) of adverse values related to
hyperinsulinemia, insulin resistance (P = 0.062), all lipid variables, and uric acid than did the subpopulation with prediabetes
alone. Persons with prediabetes alone had higher prevalences
of fasting hyperglycemia (a criterion of prediabetes) and systolic hypertension (P = 0.017). With the exception of fasting
glucose, glycated hemoglobin, and blood pressures, the adjusted mean values of risk variables were more adverse for the
subpopulation with EWET alone.
Potential efficiency of screening for EWET
On the assumption that measuring waist circumferences is
simpler and less expensive than obtaining a fasting blood
specimen or blood pressure reading, we used these NHANES
III data to illustrate how the community-based screening of
adults could benefit logistically from screening for EWET. Of
the noninstitutionalized US population aged 18-90 y, we estimated that 44.4 ± 0.8% (slightly higher for the men and lower
for the women) would exceed the threshold for waist circumference (Table 3). Thus, 56% of adults initially screened in this
manner might be immediately determined not to have EWET
and would thereby not have any need for a blood test. Table 3
further illustrates how the logistical simplicity of EWET is
more advantageous for younger adults than for older adults.
Among younger adults (aged 18-34 y), 75% would avoid a
blood test, whereas, among older adults (aged 55-74 y), only
31% would do so. At age 75-90 y, the proportion avoiding a
blood test would rise to 38% (37% for men and 39% for
women).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) suggest that
persons who have experienced several decades of excessive
lipid burden may eventually also experience higher mortality
rates. We might ask: through what mechanisms might abdominal obesity and elevated circulating triacylglycerols become associated with more familiar risk factors such as conventional dyslipidemia, hyperinsulinemia, glucose intolerance, hyperuricemia, and hypertension? Lipid fuels are primarily the long-chain fatty acids that circulate through blood in their esterified form (relatively stable triacylglycerols inside lipoproteins) or in their nonesterified form (relatively unstable free fatty acids). The increased delivery of lipolytic products from enlarged intraabdominal fat depots to the liver could result in a higher concentration of circulating VLDL and LDL particles (27, 28), whereas enhanced clearance of selected triacylglycerol-rich lipoprotein particles could depress the concentration of HDL cholesterol (29). At the same time, excessive fluxes of lipid fuels would lead to the ectopic accumulation of intracellular or subfascial fatty acid metabolites in nonadipose tissues (eg, liver, skeletal muscle), which would cause them to become insulin-resistant (10, 11, 30). Compensatory hyperinsulinemia would result (31) and would contribute through renal or other mechanisms to hyperuricemia (32, 33) and hypertension (32, 34). The increased delivery of lipolytic products to the liver would also promote hepatic gluconeogenesis (35), whereas the long-term accumulation of fatty acid metabolites in the pancreas could impair insulin secretion (36). This combination of processes would lead to hyperglycemia and an increased risk of type 2 diabetes (37).
The use of abdominal size and circulating triacylglycerol
concentration to estimate risk is not new. Investigators in
Quebec have pointed out that a "hypertriglyceridemic waist"
could serve to identify men with hyperinsulinemia, elevated
apolipoprotein B, and small, dense LDL particles (38). Without
reference to the age dependence of abdominal size and triacylglycerol concentration, they assigned cutoffs for waist circumference ( 90 cm) and triacylglycerol concentration
( 2.0 mmol/L) that were rounded for simplicity. In cross-sectional studies limited to several hundred men, they found
that the hypertriglyceridemic waist was positively associated
with angiographically assessed coronary artery disease, dyslipidemia, hyperglycemia, and hyperinsulinemia (38-40).
Our study has built on the reports from Quebec by using population-based survey data that include both sexes. In contrast with the approach taken in Canada, we assigned threshold values for waist circumference and triacylglycerol concentration that were based on the examination of two-dimensional plots obtained only from young adults. The thresholds we obtained in this manner are consistent with other research observations. An analysis of white adults participating in NHANES III found that obesity-associated risk factors were most efficiently predicted at waist thresholds of 96 cm for men and 86 cm for women (41). Our proposed value of 95 cm for men is similar, and it also falls between the mean baseline waist circumference for incident cases and noncases among middle-aged Finnish men followed prospectively for the onset of diabetes (42) or coronary heart disease (43). A longitudinal description of non-African American adolescents from Texas found that, by age 18 y, the waist circumference of males was 6 cm larger than that of females (44); this sex difference was similar to the 7-cm difference between our proposed thresholds of 95 cm for men and 88 cm for women. Our threshold triacylglycerol concentration estimate of 1.45 mmol/L falls midway between the mean baseline concentration in adults who later acquired type 2 diabetes and that of those who did not, as reported from Texas (45) and Finland (42).
Although waist circumference is not commonly measured in clinical practice, the excellent reproducibility (46) and simplicity of this measurement should facilitate its adoption. Obtaining a standardized waist measurement is no more complicated than properly obtaining the standardized weight and height needed for calculating a BMI, and a waist circumference conveys a concept of obesity that may be easier for patients to understand. In addition, a tape measure is less expensive and more portable than are a good-quality stadiometer and scale. Cross-sectional studies within North American populations of European ancestry showed that waist circumference, even without triacylglycerol measurement, is more closely associated with multiple cardiovascular disease risk factors than is BMI (14, 41).
In our era of apparently increasing obesity and diabetes (47), it becomes more important than ever for clinicians to identify efficiently those persons who will most likely benefit from preventive or therapeutic interventions. Public health programs also need efficient markers for monitoring the prevalence of metabolic risks in communities (surveillance) and for evaluating the effectiveness of community-based attempts to prevent chronic disease. Traditional approaches to these requirements have depended on the assessment of many biochemical, physiologic, and anthropometric variables. However, the various campaigns against high cholesterol, hypertension, diabetes, and overweight may be missing an opportunity to focus on a more primary pathophysiologic pathway—possibly a key component of the "common soil" (48) from which grow major portions of type 2 diabetes and cardiovascular disease.
We have presented evidence that EWET could serve as a simple marker for a syndrome of lipid overaccumulation that would help to identify either persons in whom or communities in which "lipotoxicity" (9) might be doing the most harm. Used in community-based trials, EWET might also reveal which interventions could most effectively prevent or reverse this process of accumulation. However, our data are only cross-sectional. Missing from our evidence is any indication that EWET does, over time, predict the most important ultimate outcomes such as cardiovascular events, major disability, or premature mortality. Population-based, prospective studies already showed that triacylglycerol concentration provides significant prediction of cardiovascular risk (49, 50). However, the use of EWET instead of triacylglycerol concentration would have the advantage of substantially reducing the number of venipunctures and blood samples required from the population being assessed. Nevertheless, prospective studies will be necessary to prove the utility of EWET as a simple indicator of metabolic risks.
|
ACKNOWLEDGMENTS |
|---|
HSK and RV contributed to the study design and data analysis; HSK drafted the manuscript. Neither author had any corporate affiliations or financial interests related to this research topic.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in AJCN:
This article has been cited by other articles:
|
|
 |
|
  D. Wiltgen, I.G. Benedetto, L.S. Mastella, and P.M. Spritzer Lipid accumulation product index: a reliable marker of cardiovascular risk in polycystic ovary syndrome Hum. Reprod., July 1, 2009; 24(7): 1726 - 1731. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-P. Despres, P. Poirier, J. Bergeron, A. Tremblay, I. Lemieux, and N. Almeras From individual risk factors and the metabolic syndrome to global cardiometabolic risk Eur. Heart J. Suppl., March 1, 2008; 10(suppl_B): B24 - B33. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Blackburn, I. Lemieux, B. Lamarche, J. Bergeron, P. Perron, G. Tremblay, D. Gaudet, and J.-P. Despres Type 2 Diabetes Without the Atherogenic Metabolic Triad Does Not Predict Angiographically Assessed Coronary Artery Disease in Women Diabetes Care, January 1, 2008; 31(1): 170 - 172. [Full Text] [PDF]
|
|
|
|
|
|
A. J.G. Hanley, L. E. Wagenknecht, A. Festa, R. B. D'Agostino Jr., and S. M. Haffner Alanine Aminotransferase and Directly Measured Insulin Sensitivity in a Multiethnic Cohort: The Insulin Resistance Atherosclerosis Study Diabetes Care, July 1, 2007; 30(7): 1819 - 1827. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wang Standardization of waist circumference reference data Am. J. Clinical Nutrition, January 1, 2006; 83(1): 3 - 4. [Full Text] [PDF]
|
|
|
|
|
|
A. Esmaillzadeh, P. Mirmiran, and F. Azizi Clustering of metabolic abnormalities in adolescents with the hypertriglyceridemic waist phenotype Am. J. Clinical Nutrition, January 1, 2006; 83(1): 36 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. S. Kahn The Lipid Accumulation Product Is Better Than BMI for Identifying Diabetes: A population-based comparison Diabetes Care, January 1, 2006; 29(1): 151 - 153. [Full Text] [PDF]
|
|
|
|
|
|
Z. T. Bloomgarden 2nd International Symposium on Triglycerides and HDL: Metabolic syndrome Diabetes Care, October 1, 2005; 28(10): 2577 - 2584. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Bigaard, I. Spanggaard, B. L. Thomsen, K. Overvad, and A. Tjonneland; Self-Reported and Technician-Measured Waist Circumferences Differ in Middle-Aged Men and Women J. Nutr., September 1, 2005; 135(9): 2263 - 2270. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. B. Tanko, Y. Z. Bagger, G. Qin, P. Alexandersen, P. J. Larsen, and C. Christiansen Enlarged Waist Combined With Elevated Triglycerides Is a Strong Predictor of Accelerated Atherogenesis and Related Cardiovascular Mortality in Postmenopausal Women Circulation, April 19, 2005; 111(15): 1883 - 1890. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Esmaillzadeh, P. Mirmiran, and F. Azizi Whole-grain intake and the prevalence of hypertriglyceridemic waist phenotype in Tehranian adults Am. J. Clinical Nutrition, January 1, 2005; 81(1): 55 - 63. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Zhao, S. Yakar, O. Gavrilova, H. Sun, Y. Zhang, H. Kim, J. Setser, W. Jou, and D. LeRoith Phloridzin Improves Hyperglycemia But Not Hepatic Insulin Resistance in a Transgenic Mouse Model of Type 2 Diabetes Diabetes, November 1, 2004; 53(11): 2901 - 2909. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Vinicor and B. Bowman The Metabolic Syndrome: The Emperor Needs Some Consistent Clothes: Response to Davidson and Alexander Diabetes Care, May 1, 2004; 27(5): 1243 - 1243. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and 480-990 women (
).
, enlarged waist and low triacylglycerol; 
, low waist circumference
and low triacylglycerol), by sex and 2 age strata. Insulin users were
excluded. The numbers of survey participants contributing to each mean
value was 126-812 men and 103-856 women. The numbers beneath
columns indicate the estimated percentage of participants reporting the use
of medication to lower blood sugar.