VOLUME 20 NUMBER 4 | apRiL 2012 | www.obesityjournal.org
Methods and techniques
nature publishing group
High levels of visceral adipose tissue (VAT) content have been
shown to be associated with cardiometabolic complications
such as insulin resistance, dyslipidemia, and hypertension (1–4),
which leads to an increased risk of cardiovascular diseases and
type 2 diabetes (2,5–7). Furthermore, excessive accumulation of
VAT may be considered as the fat depot with the greatest risk of
metabolic abnormalities (3,4,8). VAT content can be measured
using imageries techniques such as computed tomography (CT)
scanning. However, this technique is expensive, is time consum-
ing, and requires highly specialized personnel, and its avail-
ability is limited for clinical research. In addition, subjects may
be exposed to high amounts of radiation, which could increase
the risk of cancer (9–13), raising significant concerns when this
technique is used in a research context. Therefore, more simple
methods are needed to estimate VAT for clinical research.
Relationship Between the Bertin index
to Estimate Visceral adipose Tissue From
Dual-Energy X-Ray absorptiometry
and Cardiometabolic Risk Factors Before
and after Weight Loss
Antony D. Karelis1–3, Rémi Rabasa-Lhoret3–5, Roseline Pompilus3, Virginie Messier4, Irene Strychar3,5,6,
Martin Brochu7,8 and Mylene Aubertin-Leheudre1,2
The purpose of this study was to investigate the relationship between visceral adipose tissue (VAT), estimated with
the Bertin index obtained from dual-energy X-ray absorptiometry (DXA), with cardiometabolic risk factors before and
after a weight loss program and compare it with VAT measured with computed tomography (CT) scan. The study
population for this analysis included 92 nondiabetic overweight and obese sedentary postmenopausal women (age:
58.1 ± 4.7 years, BMI: 31.8 ± 4.2 kg/m2) participating in a weight loss intervention that consisted of a caloric restricted
diet with and without resistance training (RT). We measured (i) VAT using CT scan, (ii) body composition (using DXA)
from which the Bertin index was calculated, (iii) cardiometabolic risk factors such as insulin sensitivity (using the
hyperinsulinenic–euglycemic clamp technique), peak oxygen consumption, blood pressure, plasma lipids, C-reactive
protein as well as fasting glucose and insulin. VAT levels for both methods significantly decreased after the weight
loss intervention. Furthermore, no differences in VAT levels between both methods were observed before (88.0 ± 25.5
vs. 83.8 ± 22.0 cm2) and after (76.8 ± 27.8 vs. 73.6 ± 23.2 cm2) the weight loss intervention. In addition, the percent
change in VAT levels after the weight loss intervention was similar between both methods (−13.0 ± 16.5 vs. −12.5 ±
12.6%). Moreover, similar relationships were observed between both measures of VAT with cardiometabolic risk
factors before and after the weight loss intervention. Finally, results from the logistic regression analysis consistently
showed that fat mass and lean body mass were independent predictors of pre- and post-VAT levels for both methods
in our cohort. In conclusion, estimated visceral fat levels using the Bertin index may be able to trace variations of VAT
after weight loss. This index also shows comparable relationships with cardiometabolic risk factors when compared to
VAT measured using CT scan.
Obesity (2012) 20, 886–890. doi:10.1038/oby.2011.273
1Department of Kinanthropology, Université du Québec à Montréal, Montreal, Quebec, Canada; 2Institut Universitaire de Gériatrie de Montréal, Montreal, Quebec,
Canada; 3Department of Nutrition, Université de Montréal, Montreal, Quebec, Canada; 4Institut de Recherches Cliniques de Montréal (IRCM), Montreal, Quebec,
Canada; 5Montreal Diabetes Research Center (MDRC), Montreal, Quebec, Canada; 6Centre de Recherche du Centre Hospitalier de l’Université de Montréal
(CRCHUM), Montreal, Quebec, Canada; 7Faculty of Physical Education and Sports, Université de Sherbrooke, Sherbrooke, Quebec, Canada; 8Research Center on
Aging, Health and Social Services Centre, University Institute of Geriatrics of Sherbrooke, Sherbrooke, Quebec, Canada. Correspondence: Antony D. Karelis
Received 21 March 2011; accepted 24 July 2011; published online 25 August 2011. doi:10.1038/oby.2011.273
obesity | VOLUME 20 NUMBER 4 | apRiL 2012 887
Methods and techniques
Dual-energy X-ray absorptiometry (DXA) is widely used in
clinical studies and could be considered as the gold standard
for body composition measurement in clinical research (14).
This noninvasive technique can accurately determine total and
segmental (i.e., trunk fat) body composition; however, it is
unable to dissociate subcutaneous fat from VAT. Interestingly,
a potential method in estimating VAT from DXA has been
proposed by the study of Bertin et al. (15). This method uses
specific regions and diameters obtained from DXA to estimate
VAT content. In this cross-sectional study, the authors devel-
oped a simple index that accurately estimates VAT levels in
obese men and women (15). However, to our knowledge, the
possible association between VAT, estimated using the Bertin
index, with cardiometabolic risk factors and the ability to detect
changes in VAT by way of weight loss has not been investi-
gated. Therefore, to provide additional essential elements sup-
porting the use of this surrogate measure of VAT, the purpose
of this study was (i) to examine the ability of the Bertin index
to detect changes in VAT levels before and after a weight loss
intervention when compared to VAT levels measured using a
CT scan and (ii) to determine if both measures of VAT have
comparable associations with cardiometabolic risk factors in a
population of sedentary, overweight and obese postmenopau-
sal women, a group at increased risk for developing metabolic
Methods and Procedures
The MONET (Montreal Ottawa New Emerging Team) weight loss
intervention was designed to reduce body weight by 10% and con-
sisted of a 6-month randomized caloric restriction (CR) with or with-
out resistance training (RT). This study is a secondary analysis of the
weight loss study from the MONET group (16,17). Of 137 subjects, in
whom 89 were randomized in the CR group and 48 in the CR + RT
group, data for pre- and postintervention CT scans were only available
for 92 subjects. Sixty subjects were in the CR group, and 32 subjects
were in the CR + RT group. Thus, the study sample consisted of 92
overweight and obese postmenopausal women aged between 49 and
70 years. The study was approved by the Université de Montréal ethics
committee. After reading and signing the consent form, each partici-
pant was invited to the Metabolic Unit for a series of tests. Methods for
body composition, body fat distribution, anthropometrics, blood sam-
ples, insulin sensitivity, and cardiorespiratory fitness were determined
as previously described (18,19). Briefly, fat mass and lean body mass
were evaluated using DXA. Insulin sensitivity was measured using the
hyperinsulinemic-euglycemic clamp technique. Cardiorespiratory fit-
ness was assessed using indirect calorimetry on a ergocycle. In addi-
tion, weight loss intervention protocols were performed as previously
described (16). Women were included in the study if they met the
following criteria: (i) body mass index of ≥27 kg/m2, (ii) cessation of
menstruation for more than 1 year and a follicle-stimulating hormone
level ≥30 U/l, and (iii) free of known inflammatory disease. On physical
examination or biological testing, all participants had no history or evi-
dence of the following: (i) cardiovascular disease, peripheral vascular
disease, or stroke, (ii) diabetes (fasting glucose <7.0 mmol/l and 2-hour
post 75 g oral glucose-tolerance test <11.1 mmol/l), and (iii) medica-
tions that could affect cardiovascular function and/or metabolism.
Visceral fat measurement and estimation
CT. A GE High Speed Advantage CT scanner (General Electric Medical
Systems, Milwaukee, WI) was used to measure visceral fat content. The
subjects were examined in the supine position with both arms stretched
above their head and were asked to stop breathing momentarily during
the scan. The position of the scan was established at the L4-L5 verte-
bral disc using a scout image of the body. VAT area was quantified by
delineating the intra-abdominal cavity at the internal most aspect of
the abdominal and oblique muscle walls surrounding the cavity and
the posterior aspect of the vertebral body. The software used to calcu-
late the area was advantage workstation from GE 4.1; VOXta 3.0.64u
volume viewer 3D soft. The techniques factors are 120 KV and 400 UA.
An eight-slice scan of 0.625 mm each at the level of L4-L5 was used.
Values of visceral fat were calculated with the consideration of all eight
slices, i.e., a thickness of 5 mm. The test–retest measures of the different
body fat distribution indices on 10 CT scans yielded a mean absolute
difference of ±1%.
Bertin index. VAT levels were also estimated using the index of Bertin
et al. (15). This method uses regions and diameters obtained from DXA
to estimate visceral fat content. That is, this index uses two abdominal
transverse diameters on the DXA scans. The first is the transverse exter-
nal diameter (TED) corresponding to the abdominal width, which is
measured above the upper end of the iliac crests (umbilical level), and
the second is the transverse internal diameter (TID) which is the lean
core of the abdomen stripped of the subcutaneous adipose tissue at the
same level of the TED. Therefore, TID corresponds to TED without the
subcutaneous fat width (SFW) present on each side of the abdomen.
SFW is equal to half of the difference between these two diameters.
The sagittal diameter (SD) was also measured in the supine position at
the umbilical level. Thus, the following equation: (79.6) (SD − SFW) ×
(TID) / height (cm) − 149 was used to determine estimated VAT levels
in cm2 (15). The Bertin index has been shown to be strongly associated
with VAT levels using CT scan (r = 0.94 for women and r = 0.88 for
men) in relatively healthy obese individuals (15). It should be noted that
TED and TID in this study were measured using a ruler on a total body
image printout. Only one coauthor performed all of the measurements.
The intraclass coefficient correlation for test–retest for TED and TID
was 0.99 (n = 15).
Data are expressed as the mean ± standard deviation. Pearson corre-
lations were performed to examine the relationship between visceral
fat levels and metabolic risk factors. A paired t-test was performed to
compare pre- and post-VAT levels as well as percent change between
the CT scan and the Bertin index. Finally, a stepwise multilinear regres-
sion analysis was performed to identify predictors of pre- and post-VAT
levels for both methods. Independent variables considered in the final
model for pre- and post-VAT for both methods were fat mass, lean body
mass, high-density lipoprotein cholesterol, triglycerides, hsC-reactive
protein, insulin sensitivity, and cardiorespiratory fitness. Statistical
analysis was performed using SPSS for Windows version 17 (Chicago,
IL). Significance was accepted at P < 0.05.
Physical and metabolic characteristics of the 92 overweight
and obese postmenopausal women are presented in Table 1.
It should be noted that we observed a significant decrease
in body weight (pre: 82.0 ± 12.7 vs. post: 76.8 ± 12.4 kg), BMI
(pre: 31.8 ± 4.2 vs. post: 29.8 ± 4.2 kg/m2), and waist circum-
ference (pre: 95.1 ± 8.9 vs. post: 90.2 ± 9.1 cm) after the weight
There were no differences in VAT levels between the CR +
RT group and the CR group using both measures of VAT
before and after the weight loss intervention (data not shown).
Therefore, we pooled all data from both groups. Table 2 shows
VAT values of both methods before and after the weight
VOLUME 20 NUMBER 4 | apRiL 2012 | www.obesityjournal.org
Methods and techniques
loss intervention. VAT levels for both methods significantly
decreased after the weight loss intervention. Furthermore, no
differences in VAT levels between both methods were observed
before (88.0 ± 25.5 vs. 83.8 ± 22.0 cm2) and after (76.8 ± 27.8 vs.
73.6 ± 23.2 cm2) the weight loss intervention. In addition, the
percent change in VAT levels after the weight loss intervention
was similar between both methods (−13.0 ± 16.5 vs. −12.5 ±
Pearson correlation coefficients between both measures
of VAT and cardiometabolic characteristics are presented in
Table 3. A significant relationship was found between VAT
measured using the CT scan and VAT estimated by the Bertin
index at baseline (r = 0.59, P < 0.01) at after the intervention
(r = 0.65, P < 0.01). In addition, the percent change of VAT
between both methods was significantly correlated (r = 0.52,
P < 0.01). Both measures of VAT were significantly correlated
with BMI, fat mass, waist circumference, waist/hip ratio, lean
body mass, fasting insulin, insulin sensitivity, and hsC-reactive
protein at baseline and after the intervention. Furthermore,
no associations were observed between both measures of VAT
with total cholesterol, low-density lipoprotein cholesterol, free
fatty acids, and diastolic blood pressure at baseline and after the
intervention. Moreover, correlations between both measures
of VAT varied at pre and post with high-density lipoprotein
cholesterol, only at pre for triglycerides, 2-hour glucose, and
glucose as well as only at post for systolic blood pressure and
cardiorespiratory fitness. Finally, a significant correlation was
only found between VAT estimated with the Bertin index and
Finally, we performed a stepwise regression analysis to iden-
tify independent predictors of VAT. Table 4 illustrates the
summary of the model. Our results show that fat mass and lean
body mass were independent predictors of pre- and post-VAT
for both methods.
table 1 Physical and metabolic characteristics of the
VariablesMean ± s.d.Range
Age (years)58.1 ± 4.7 48.8–70.5
Body mass index (kg/m2) 31.8 ± 4.226.1–45.8
Lean body mass (kg)42.2 ± 5.8 32.6–61.0
Fat mass (kg) 37.4 ± 8.4 24.0–63.8
Waist circumference (cm)95.1 ± 8.8 77.5–117
Waist/hip 0.83 ± 0.050.70–0.93
Visceral adipose tissue (cm2) (CT scan) 88.0 ± 25.539.7–165.9
Visceral adipose tissue (cm2)
83.8 ± 22.0 30.1–148.1
Insulin sensitivity (mg/min/kg LBM) 11.2 ± 2.84.7–22.2
Total cholesterol (mmol/l) 5.4 ± 0.83.3–7.5
LDL-cholesterol (mmol/l) 3.2 ± 0.81.4–5.1
HDL-cholesterol (mmol/l)1.5 ± 0.3 1.0–2.7
Triglycerides (mmol/l) 1.7 ± 0.80.6–4.5
Free fatty acids (mmol/l)0.63 ± 0.250.3–1.7
Fasting insulin (µU/ml)15.3 ± 6.5 4.9–44.5
Fasting glucose (mmol/l)5.2 ± 0.53.9–6.6
2-hour glucose (mmol/l)6.4 ± 1.83.7–11.0
hsC-reactive protein (mg/l)3.0 ± 2.20.3–10.1
Systolic blood pressure (mm Hg)122 ± 14 92–160
Diastolic blood pressure (mm Hg)76 ± 8.060–100
VO2 peak (ml/kg/min)
Values are mean ± s.d.
CT, computer tomography; HDL, high-density lipoprotein; LBM, lean body mass;
LDL, low-density lipoprotein; VO2 peak, peak oxygen consumption.
18.1 ± 3.18.8–25.3
table 2 Visceral fat values before and after the weight loss
intervention (n = 92)
P value (between
both methods)(CT scan) (Bertin index)
Pre (cm2)88.0 ± 25.583.8 ± 22.0 0.06
Post (cm2) 76.8 ± 27.8a
73.6 ± 23.2a
−13.0 ± 16.5−12.5 ± 12.60.74
Values are mean ± s.d.
aSignificantly different between pre values.
table 3 Bivariate correlations between visceral fat and
cardiometabolic risk factors (n = 92)
Previsceral fatPostvisceral fat
Visceral fat (Bertin index)
0.50** 0.60** 0.62**0.65**
Lean body mass
Free fatty acids0.110.210.110.21
Systolic blood pressure0.120.18
Diastolic blood pressure0.060.070.170.11
CT, computer tomography; HDL, high-density lipoprotein; LDL, low-density lipo-
protein; VO2 peak, peak oxygen consumption.
*P < 0.05; **P < 0.01.
obesity | VOLUME 20 NUMBER 4 | apRiL 2012 889
Methods and techniques
It is important in clinical research to develop simple and accu-
rate methods for the measurement of VAT because of its critical
role in the pathophysiology of cardiometabolic abnormalities.
Thus, the purpose of this study was to investigate the ability
of the Bertin index to detect changes in VAT levels before and
after a weight loss intervention when compared to VAT levels
measured using a CT scan. We also examined the relationship
between measured or estimated VAT with cardiometabolic risk
factors. Such data could be essential to establish the validity of
this surrogate measure for clinical use in research protocols.
Results of this study extend the findings of Bertin et al. (15)
by examining the ability of this index to detect changes in VAT
before and after a weight loss intervention and by exploring
its relationship with cardiometabolic risk factors. That is, we
showed no significant differences in VAT levels between both
methods before and after the weight loss intervention as well
as in the percent change of VAT content. Furthermore, similar
relationships were observed between both measures of VAT
with cardiometabolic risk factors before and after the weight
loss intervention. For example, we found comparable cor-
relations (pre and post) between both measures of VAT with
body composition, insulin sensitivity, fasting insulin, and hsC-
reactive protein. However, high-density lipoprotein cholesterol
seemed to be the only metabolic risk factor that did not cor-
relate similarly (pre and post) when comparing both methods.
Finally, results from the logistic regression analysis consist-
ently showed that fat mass and lean body mass were primary
and secondary independent predictors of pre- and post-VAT
levels for both methods in our cohort.
Based on these findings, we would suggest that estimating
VAT from DXA using the Bertin index provides an accept-
able surrogate marker of VAT for clinical research intervention
studies. In addition, the Bertin index offers additional advan-
tages since the DXA is widely available and allows the meas-
urement of body composition within minutes and exposes
patients to far lower radiation than a CT scan. In light of recent
concerns about cancer risk related to CT scan radiation expo-
sure (9–13), the use of such a method to measure VAT in a
research context raises significant ethical concerns. Thus, the
DXA could be used as an alternative for the estimation of VAT.
Accordingly, this index has been used in a previous study (20).
Finally, it should be noted that other noninvasive imaging
methods may be used for the determination of VAT such as
echocardiographic epicardial fat (21,22). Evidence has shown
that echocardiographic epicardial fat thickness significantly
decreased after a weight loss intervention in obese individuals
This study has several limitations. Our findings are limited to
a cohort composed of nondiabetic sedentary overweight and
obese postmenopausal women who participated in a univer-
sity-based research weight loss program. However, our results
are strengthened by the use of pre- and post-weight loss data
as well as the use of gold-standard techniques to measure body
composition, VAT, insulin sensitivity, and blood profile in a
relatively large sample size of well-characterized overweight
and obese postmenopausal women.
In conclusion, results of this study indicate that estimated
VAT levels using the Bertin index shows comparable relation-
ships with cardiometabolic risk factors when compared to
VAT measured using CT scan in overweight/obese postmeno-
pausal women. Furthermore, changes in VAT after a weight
loss intervention appeared to be well detected with this index.
Thus, this index may be an acceptable alternative in assessing
VAT in clinical research.
This manuscript was supported by CiHR (Canadian institute for Health
Research) grants: 63279 MONET study (Montreal Ottawa New Emerging
Team) and 88590 SOMET study (Sherbrooke Montreal Ottawa Emerging
Team) as well as the J-a DeSève chair for clinical research to R.R.-L.
R.R.-L. and a.D.K. hold scholarships from the Fonds de Recherche en Santé
du Québec. R.p. is a scholar from international atomic agency.
The authors declared no conflict of interest.
© 2011 The Obesity Society
1. Arsenault BJ, Lachance D, Lemieux I et al. Visceral adipose tissue
accumulation, cardiorespiratory fitness, and features of the metabolic
syndrome. Arch Intern Med 2007;167:1518–1525.
2. Balkau B, Deanfield JE, Després JP et al. International Day for the Evaluation
of Abdominal Obesity (IDEA): a study of waist circumference, cardiovascular
disease, and diabetes mellitus in 168,000 primary care patients in 63
countries. Circulation 2007;116:1942–1951.
table 4 stepwise regression analysis regarding independent predictors of Vat
Dependent variable Step Independent variablePartial r2
Total r2 cumulative
Pre-VAT (CT scan)1Lean body mass 0.2600.2600.003
2Fat mass 0.0740.334 0.004
Pre-VAT (Bertin index)1 Fat mass 0.375 0.375<0.00
2 Lean body mass0.0840.4590.001
Post-VAT (CT scan)1Fat mass 0.201 0.2010.01
2Lean body mass0.044 0.245 0.03
3 HDL-cholesterol0.0340.279 0.03
Post-VAT (Bertin index)1Fat mass 0.2940.294 <0.00
2Lean body mass 0.0560.3500.006
CT, computer tomography; HDL, high-density lipoprotein; VAT, visceral adipose tissue.
890 Download full-text
VOLUME 20 NUMBER 4 | apRiL 2012 | www.obesityjournal.org
Methods and techniques
3. Després JP, Arsenault BJ, Côté M, Cartier A, Lemieux I. Abdominal obesity:
the cholesterol of the 21st century? Can J Cardiol 2008;24 Suppl D:7D–12D.
4. Després JP, Lemieux I, Bergeron J et al. Abdominal obesity and the
metabolic syndrome: contribution to global cardiometabolic risk. Arterioscler
Thromb Vasc Biol 2008;28:1039–1049.
5. Després JP. Cardiovascular disease under the influence of excess visceral
fat. Crit Pathw Cardiol 2007;6:51–59.
6. Lamarche B, Lemieux S, Dagenais GR, Després JP. Visceral obesity and the
risk of ischaemic heart disease: insights from the Québec Cardiovascular
Study. Growth Horm IGF Res 1998;8 Suppl B:1–8.
7. Mathieu P, Lemieux I, Després JP. Obesity, inflammation, and cardiovascular
risk. Clin Pharmacol Ther 2010;87:407–416.
8. Pascot A, Després JP, Lemieux I et al. Deterioration of the metabolic risk
profile in women. Respective contributions of impaired glucose tolerance
and visceral fat accumulation. Diabetes Care 2001;24:902–908.
9. Baerlocher MO, Detsky AS. Discussing radiation risks associated with CT
scans with patients. JAMA 2010;304:2170–2171.
10. Baskerville JR, Chang JH, Viator M et al. Dose versus diagnosis: iatrogenic
radiation exposure by multidetector computerised tomography in an
academic emergency department with measurement of clinically actionable
results and emergently treatable findings. Emerg Med J 2009;26:15–19.
11. Berrington de Gonzalez A, Mahesh M, Kim KP et al. Projected cancer risks
from computed tomographic scans performed in the United States in 2007.
Arch Intern Med 2009;169:2071–2077.
12. Brenner DJ, Doll R, Goodhead DT et al. Cancer risks attributable to low
doses of ionizing radiation: assessing what we really know. Proc Natl Acad
Sci USA 2003;100:13761–13766.
13. Smith-Bindman R, Lipson J, Marcus R et al. Radiation dose associated with
common computed tomography examinations and the associated lifetime
attributable risk of cancer. Arch Intern Med 2009;169:2078–2086.
14. Plank LD. Dual-energy X-ray absorptiometry and body composition. Curr
Opin Clin Nutr Metab Care 2005;8:305–309.
15. Bertin E, Marcus C, Ruiz JC, Eschard JP, Leutenegger M. Measurement
of visceral adipose tissue by DXA combined with anthropometry in obese
humans. Int J Obes Relat Metab Disord 2000;24:263–270.
16. Brochu M, Malita MF, Messier V et al. Resistance training does not
contribute to improving the metabolic profile after a 6-month weight
loss program in overweight and obese postmenopausal women. J Clin
Endocrinol Metab 2009;94:3226–3233.
17. Messier V, Rabasa-Lhoret R, Doucet E et al. Effects of the addition of a
resistance training programme to a caloric restriction weight loss intervention
on psychosocial factors in overweight and obese post-menopausal women:
a Montreal Ottawa New Emerging Team study. J Sports Sci 2010;28:83–92.
18. Karelis AD, Faraj M, Bastard JP et al. The metabolically healthy but obese
individual presents a favorable inflammation profile. J Clin Endocrinol Metab
19. Karelis AD, Lavoie ME, Messier V et al. Relationship between the metabolic
syndrome and physical activity energy expenditure: a MONET study. Appl
Physiol Nutr Metab 2008;33:309–314.
20. Aubertin-Leheudre M, Lord C, Goulet ED, Khalil A, Dionne IJ. Effect of
sarcopenia on cardiovascular disease risk factors in obese postmenopausal
women. Obesity (Silver Spring) 2006;14:2277–2283.
21. Iacobellis G, Assael F, Ribaudo MC et al. Epicardial fat from
echocardiography: a new method for visceral adipose tissue prediction.
Obes Res 2003;11:304–310.
22. Silaghi AC, Poanta L, Valea A, Pais R, Silaghi H. Is epicardial adipose tissue,
assessed by echocardiography, a reliable method for visceral adipose tissue
prediction? Med Ultrason 2011;13:15–20.
23. Iacobellis G, Singh N, Wharton S, Sharma AM. Substantial changes in
epicardial fat thickness after weight loss in severely obese subjects. Obesity
(Silver Spring) 2008;16:1693–1697.
24. Kim MK, Tanaka K, Kim MJ et al. Comparison of epicardial, abdominal
and regional fat compartments in response to weight loss. Nutr Metab
Cardiovasc Dis 2009;19:760–766.