Obesity and regional body-fat distribution in men: separate

Downloaded from https://academic.oup.com/ajcn/article-abstract/60/5/682/4732014 by East Carolina University user on 11 January 2019

and joint relationships to glucose tolerance and plasma
lipoproteins1 -3

Mauro Zamboni, Fabia Armellini, Luciano Cominacini, Emanuela Turcato, Tiziana Todesco,
Luisa Bissoli, Rocco Micciolo, Ivo Andrea Bergamo-Andreis, and Ottavio Bosello

ABSTRACT Relationships between obesity and fat distri- visceral fat accounted for a greater proportion of the variance of
bution as evaluated by computed tomography and metabolic van- cardiovascular disease risk factors than did total body fat mass,
ables were studied in 35 men. Significant correlations emerged and with the results of Larsson et al (12) and Lapidus et al (13),
between body mass index and sum of glucose during oral glucose who, in a prospective study of risk factors for ischemic heart
load and HDL triglycerides and also between visceral abdominal disease in a group of 54 elderly men with a mean body mass
fat and triglycerides. apolipoprotein B, sum of insulin during oral index of 25 and in a group of women aged 38-60 y with a mean
glucose load, very-low-density-lipoprotein (VLDL) cholesterol, body mass index of 24. 1 , found the greatest risk of cardiovascular
and VLDL and low-density-lipoprotein (LDL) triglycerides. Vis- disease in lean men with the highest waist-to-hip ratio. Although
ceral abdominal fat correlated negatively with the ratio of HDL the relationships between visceral fat and metabolic disturbances
to LDL cholesterol. When the subjects were subdivided into four have been confirmed by several studies in women, few studies
groups according to body mass index ( 26.7, > 26.7) and me- are available in men, particularly on the relationships between
dian visceral abdominal fat, no significant differences were found visceral fat and lipoprotein concentrations.
in body mass index, whereas significant differences were found Very little has been reported on the joint and separate effects
for triglycerides, cholesterol, apolipoprotein B, VLDL choles- of obesity and visceral fat on metabolic disturbances, particularly
terol, HDL:LDL cholesterol, and VLDL triglycerides. Our study lipid disorders. The aim of this study was to analyze the joint and
shows that the amount of visceral abdominal fat is the most rel- separate relationships of obesity and visceral adipose tissue to
evant factor for metabolic abnormalities. Our data also suggest metabolic variables in a group of male subjects, discriminating
that the effect of visceral fat is independent of body mass in- between their respective interrelationships with cardiovascular
dcx. Am J C/in Nutr l994;60:682-7. risk factors.

KEY WORDS Visceral fat, obesity, glucose, insulin, lipo-

Subjects and methods
proteins

Experimental subjects
Introduction
The study was conducted in 35 male subjects ranging in age
from 34 to 68 y with a body mass index (BMI) ranging from
Recent studies have demonstrated that regional body-fat dis-
19.5 to 32.5. All subjects were hospitalized in our institute,
tnibution rather than obesity itself is a risk factor for cardiovas-
whose Ethics Committee approved the study. Subjects presenting
cular disease ( 1 , 2). Central or visceral body-fat distribution is
> 10% weight reduction over the 6 mo before the study, as eval-
associated with carbohydrate and lipid metabolism (3-6). Re-
uated by self-recorded body weight, were excluded. All subjects
cently, several studies have evaluated the separate contributions
had stable weights during the month preceding the study. Sub-
of obesity and regional body-fat distribution as cardiovascular
jects with a history and/or clinical signs of congestive heart fail-
risk factors, although there is still a great deal of uncertainty in
ure and alimentary problems were also excluded, as were subjects
this connection and not all studies have yielded unambiguous
with hypertension. None of the subjects in either group was di-
results (7-13).
Pouliot et al (7) found no significant association between vis-
ceral fat and
plasma lipoprotein in lean premenopausal women. I From the Institutes of Internal Medicine and Radiology, University
of Verona, Policlinico di Borgo Roma, Verona, Italy and the Institute of
These results
are in line with those of Landin et al (9), who, in
Statistics. University of Trento, Italy.
a comparison of lean and obese middle aged women without
2 Supported by grants from MURST and the CNR FATMA C I 79037N
diabetes, demonstrated the importance of obesity in expressing
project.
metabolic risk factors associated with central body-fat distribu- ‘ Address reprint requests to M Zamboni, Clinica Medica, Policlinico
tion. They seem to be at variance, however, with those reported di Borgo Roma, 37134 Verona, Italy.
by Peiris et al (I 1), who, in a sample of 33 healthy premenopausal Received September 23. 1993.
women without diabetes with a wide range of obesity, found that Accepted for publication May 9, 1994.

682 Am J C/in Nuir l994;60:682-7. Printed in USA. © 1994 American Society for Clinical Nutrition
 VISCERAL FAT, OBESITY, AND METABOLIC VARIABLES 683

abetic or was taking medication known to affect carbohydrate TABLE I

and lipid metabolism. All patients were hospitalized throughout Anthropometric characteristics of the study sample’

the study. An isoenergetic diet containing 55% carbohydrates,

Value
30% lipids, and 15% protein was given to all the men before the
study for 3 d. Patient characteristics are listed in Table 1. Age (y) 52.7 ± 8.5 (34-68)
Ten subjects were smokers at the beginning of the study, I1 Weight (kg) 73.5 ± 10.3 (54-97.9)
had stopped smoking 1 y before, and 14 had never smoked. BMI2 25.7 ± 3 (19.5-32.5)
None of the subjects drank more than two small glasses of wine Total AT (cm2) 250 ± 92.5 (94-491.3)

a day (< 20 g alcohol). All had sedentaryjobs and did not engage Subcutaneous AT (cm2) 143.1 ± 66.9 (42.3-335)

Downloaded from https://academic.oup.com/ajcn/article-abstract/60/5/682/4732014 by East Carolina University user on 11 January 2019

Visceral AT (cm2) l07.6 ± 34.3 (5 1.7-202)
in regular exercise. Subjects were divided into four groups ac-
Visceral:subcutaneous AT 0.85 ± 0.32 (0.36- 1.9)
cording to their body mass index ( 26.7, > 26.7) on the basis
of the Italian suggested cutoff point for overweight (14) and the
‘ I ± SD; range in parentheses. AT, adipose tissue.
median value of visceral abdominal adipose tissue (AT) ( 107 2 In kglm2.
vs > 107 cm2). The cutoff point for visceral AT was chosen to
obtain two groups with the same numbers of subjects.
erides contained in VLDL, LDL, HDL1, and HDL3 was always
Anthropometric indexes within 100 ± 6% of total plasma cholesterol and triglycerides.
Apolipoprotein (apo) A-I in plasma was measured by the ra-
The following anthropometric indexes were evaluated in all
dioimmunodiffusion method (20) with Immuno apo A-I plates
subjects: weight, height, BMI, and body-fat distribution, as cab-
(Immuno Diagnostica GES, MBH, Vienna). Differences between
culated by measuring abdominal fat areas by computed tomog-
mntraassay and interassay replicates were < 3.5% and 5.0%, re-
raphy. BMI was evaluated as body weight/height2. Body-fat dis-
spectively. After extensive dialysis total protein in VLDL and
tribution was determined by computed tomography according to
LDL was measured by a modification of the Lowry procedure.
Sj#{246}str#{246}m
(15). Total abdominal AT, visceral abdominal AT, sub-
Apo B concentrations in the lipoproteins were calculated as the
cutaneous abdominal AT, and the ratio of visceral to subcuta-
difference between total protein in the fractions and soluble pro-
neous abdominal AT were evaluated by a single scan at the L4
teins after precipitation of apo B by isopropyl alcohol.
level. Subject centering was obtained by a longitudinal topogram
at the L4 level. The preselected attenuation interval was - 150 to Statistical analysis
-50 Hounsfield units. A cursor was used to define the total cross-
sectional area and the area of visceral fat (area inside the rectus Results are presented as mean ± SD. Although all the variables
abdominis muscles). Data were elaborated by using a histogram- considered were not significantly different from the Gaussian dis-
based statistical program. Our margin of error was 1 . I % for dou- tribution (Kolmogorov-Smirnov test), some data transformation
ble determination. was used to make the empirical distribution more symmetrical.
The square root transformation was used for total AT, subcuta-
Metabolic variables neous AT, and visceral AT, whereas for the other variables (cx-
cept age) the natural logarithm was used. The correlation coef-
Patients were given an oral-glucose-tolerance test (OGTT, 75 ficient (r) was used for correlation analysis. To analyze differ-
g glucose). Blood samples for determining glucose and insulin ences in the mean values of the variables considered among the
concentrations were obtained at 0, 30, 60, 90, 120, and 180 mm. four groups identified on the basis of BMI and visceral abdominal
Plasma glucose was measured by using a glucose-oxidase- AT values and to quantify the distinct roles of visceral AT and
method analyzer (Beckman Instruments Inc, Palo Alto, CA). The BMI, a two-way analysis of variance (ANOVA) for a factorial
intraassay CV was 1 .5%. Plasma immunoreactive insulin under- design was used (21). Analysis of covariance was used to per-
went duplicate measurement by double-antibody radioimmuno- form age-adjusted analyses (2 1 ). For all analyses a significance
assay with a commercial kit (Diagnostic Products Corporation, level of 0.05 was used.
Los Angeles). The detection limit of the insulin assay was 6
pmollL (1 tU/mL) and the intraassay CV was 4.9%.
A blood sample was obtained from an antecubital vein in the Results
morning after a 12-h fast. Cholesterol and triglycerides in lipo-
protein fractions and very-low-density-lipoprotein (VLDL), low- The results of correlations between BMI and visceral AT with
density-lipoprotein (LDL), and HDL subfractions (HDL and metabolic variables are shown in Table 2. Significant correlations
HDL3) were evaluated. Blood was collected into EDTA-contain- emerged between BMI and sum of glucose during OGTT (r
ing tubes (2.68 mmol/L blood). The plasma was separated within = 0.42, P < 0.01) and HDL3 triglycerides (r = 32, P < 0.05),
2 h. Lipoproteins were isolated by preparatory ultracentrifugation but not triglycerides, cholesterol, apo B, or sum of insulin during
(model L5-65; Beckman Instruments Inc) (16). Sodium bromide the OGTT.
was used for density gradient separations: VLDL, d < 1.006; A significant positive correlation emerged between visceral ab-
LDL, d < 1 .006- 1 .063; HDL2, d = 1 .063- 1 . 125; and HDL3, d dominal AT and triglycerides (r = 0.47, P < 0.01), apo B (r
= 1. 125- 1 .210. Total HDL cholesterol was determined after pre- = 0.41, P < 0.05), sum of insulin during the OGTT (r = 0.43,
cipitation with manganese heparin (17). P < 0.01), VLDL cholesterol (r = 0.42, P < 0.01), VLDL tri-
Cholesterol and triglycerides in plasma and lipoprotein frac- glycerides(r = 0.43, P < 0.01), and LDL triglycerides(r 0.41,
=

tions were measured by Technicon autoanalyzer II methodology P < 0.05); a significant negative correlation emerged between
(Technicon Inc. Co. Tarrytown, NY) (18, 19). The recovery in visceral abdominal AT and HDL:LDL cholesterol (r = -0.33, P
the lipoprotein fractions, ie, the sum of cholesterol and triglyc- < 0.05) (Fig 1).
684 ZAMBONI ET AL

TABLE 2 role in determining the other metabolic disorders (22). Fujioka

Correlations between body mass index. visceral adipose tissue (AT), et al (23) demonstrated an association between visceral AT and
and metabolic variables’
impairment of glucose tolerance in both men and women, but
their study sample showed a wide variation in age and included
Body mass index Visceral AT
diabetic patients.
Triglycerides (mmol/L) 0.17 0.472 Seidell et al (24) showed that male abdominal obesity, char-
Cholesterol (mmol/L) 0. I 3 0.27 acterized by an increase in visceral fat, is associated with elevated
Apolipoprotein A-I (mgfL) -0.05 -0.13 concentrations of glucose, insulin, and C-peptide in both fasting
Apolipoprotein B (mglL) 0.10 0.4l

Downloaded from https://academic.oup.com/ajcn/article-abstract/60/5/682/4732014 by East Carolina University user on 11 January 2019

conditions and during the OGTT, although in this study the group
2-h Glucose during OGTI’ (mmollL) 0.502 0.20
of subjects with high visceral AT also had higher BMI and higher
: Glucose during OGTI’ (mmol/L) 0.422 0.30
subcutaneous AT values compared with men with low visceral
: Insulin during OGTF (pmol/L) 0.05 0.432

VLDL cholesterol (mmollL)

AT (BMI 26.9 vs 25.3 and subcutaneous AT 246. 1 vs I 5 1 cm2,
0.23 0.422
LDL cholesterol (mmollL) 0.07 0.20 respectively) and therefore it is difficult to distinguish the effect
HDL2 cholesterol (mmollL) -0.19 -0.29 of visceral AT on metabolic variables from that of BMI and sub-
HDL, cholesterol (mmol/L) -0.02 -0.21 cutaneous AT.
HDL:LDL cholesterol -0.23 On comparing both fertile and menopausal obese women, we
VLDL triglycerides (mmollL) 0.26 0432 already demonstrated that visceral fat is the variable with the
LDL triglycerides (mmol/L) 0.30 0.4I greatest predictive power with regard to hemodynamic and met-
HDL2 triglycerides (mmolIL) 0.26 0.23 abolic indexes and that, after adjustment for amount of visceral
HDL, triglycerides (mmollL) 0.32 0.14
AT, but not for BMI, the metabolic differences observed between
fertile and menopausal women disappear (25).
I OGIT. oral-glucose-tolerance test; , sum of values; VLDL, very-
Pouliot et al (8) recently studied the relationships between re-
low-density lipoprotein: LDL. low-density lipoprotein; HDL, high-den-
gional body-fat distribution and insulin glucose homeostasis in
sity lipoprotein.
2p < 0.01.
3 P < 0.05.

a
Table 3 shows the mean values for the untransformed variables
In Tg 6.5
as well as the results of the ANOVA analysis for the transformed 0
(mmollL) 6
values. When the anthropometric and metabolic variables were a
analyzed, no significant interaction was found; thus we can say
5.5
a
that the effect of BMI is the same irrespective of the level of
r= 0.51
visceral abdominal AT and that the effect of visceral abdominal
5 p<0.01
AT is the same irrespective of the BMI. As far as the relationship a
a
of BMI is concerned, no significant results were found for any 4.5
ID

of the metabolic variables.

When the relationship of visceral abdominal AT was analyzed, 4
significant results were found for the following variables: tri-
3.5
glycerides (P < 0.05), cholesterol (P < 0.05), apo B (P < 0.05),
VLDL cholesterol (P < 0.05), HDL:LDL cholesterol (P < 0.05), 0 100 200 300 2

and VLDL triglycerides (P < 0.05). Independently of values of Visceral AT (cm

BMI, higher mean values in subjects with visceral abdominal AT
> 107 were found for triglycerides, cholesterol, apo B, VLDL
b
cholesterol, and VLDL triglycerides; mean values of HDL:LDL
cholesterol were higher in males with visceral abdominal AT
In IRI 9.3 7.5 In : IRI
107, whether obese or nonobese. (pmollL) a , (.tUlml)
Table 4 shows the age-adjusted mean values of the variables 8.8
considered, as well as the results of the analysis of covariance. 8.3 6.5
After adjustment for the effects of age, the results effectively
proved unaltered. Smoking and age-adjusted mean values of the 7.8 6

variables considered as well as the results of the analysis of co-

7.3 5.5
variance remained unaltered (data not shown).
6.8 5

Discussion 6.3 4.5

5.8 4
The results of the study show that in men metabolic abnor-
malities regarded as risk factors for cardiovascular diseases are 0 100 200 300

due more to the location of body fat than to the actual degree of VisceralAT (cm2)
obesity. It is very interesting to observe that visceral AT in our FIG 1 . Correlations between visceral adipose tissue (AT) area and
subjects is associated with insulin concentrations during the natural logarithm of triglycerides (In Tg) and of sum of insulin (In IRI)
OGTTs which is generally acknowledged as playing a crucial during oral-glucose-tolerance test in 35 male subjects.
 VISCERAL FAT, OBESITY, AND METABOLIC VARIABLES 685

TABLE 3
Anthropometric and metabolic variables, and results of ANOVA for transformed values’

BMI 26.7 BMI > 26.7

Visceral AT Visceral AT Visceral AT Visceral AT F

107 (cm2) > 107 (cm2) 107 (cm2) > 107 (cm2)
Variable (n = 13) (,z = 7) (n = 5) (n = 10) Visceral AT BMI Interaction

Age (y) 51.9 ± 9.9 55.1 ± 9.0 53.4 ± 8.9 51.9 ± 7.0 0.59 0.10 0.55

Downloaded from https://academic.oup.com/ajcn/article-abstract/60/5/682/4732014 by East Carolina University user on 11 January 2019

Total AT (cm2) 167.77 ± 48.94 288.73 ± 61.47 240.40 ± 51.13 336.04 ± 79.15 20.792 6.652 0.86
Subcutaneous AT (cm2) 91.10 ± 36.06 151.87 ± 44.07 145.08 ± 53.60 203.84 ± 67.43 8.292 4.882 0.17
Visceral:subcutaneous AT 0.92 ± 0.27 0.96 ± 0.28 0.84 ± 0.60 0.70 ± 0.23 0.07 I .06 0.27
Triglycerides (mmollL) 1.60 ± 0.57 3.31 ± 2.42 2.55 ± 1.56 2.31 ± 0.72 7452 2.82 3.28
Cholesterol (mmoLIL) 5.67 ± I .20 6.99 ± I .44 6.48 ± I .07 6.57 ± I .42 5.022 1 .79 1.99
Apolipoprotein A-I (mgIL) I 177.5 ± 166.8 1097.1 ± 99.3 1 155.0 ± I 16.2 1221.4 ± 204.0 1.02 0.03 1.12
Apolipoprotein B (mglL) 1178.3 ± 275.9 1517.1 ± 203.4 1305.0 ± 201.0 1430.0 ± 168.0 8.882 1.19 1.39
2-h Glucose (mmollL)’ 7.10 ± 2.22 6.93 ± 1.02 7.94 ± 0.55 8.34 ± 2.94 0.01 0.81 0.01
: Glucose during 0011’
(mmol/L) 41.92 ± 9.75 51.08 ± 18.00 51.81 ± 7.59 53.95 ± 9.84 2.92 3.10 0.72
: Insulin during 0011’
(pmol/L) 1700.8 ± 1094.3 2606.2 ± 1852.7 2155.5 ± 982.4 1493.3 ± 389.3 3.17 1.17 3.42
(jiUlmL) 283.46 ± 182.39 434.37 ± 308.79 359.25 ± 163.74 248.89 ± 64.88 3.17 1.17 3.42
VLDL cholesterol (mmollL) 0.5 1 ± 0.26 1 .26 ± I .20 0.86 ± 0.7 1 0.72 ± 0.29 6.452 1 .85 2.89
LDL cholesterol (mmol/L) 4.01 ± 1.17 4.78 ± 0.46 4.01 ± 0.71 4.72 ± 1.17 3.84 0.56 0.56
HDL2 cholesterol (mmollL) 0.40 ± 0.13 0.30 ± 0.66 0.37 ± 0.12 0.35 ± 0.14 2.43 0.13 0.50
HDL cholesterol (mmollL) 0.77 ± 0.21 0.71 ± 0.20 0.73 ± 0.14 0.76 ± 0.29 0.26 0.03 0.05
HDL:LDL cholesterol 0.1 1 ± 0.05 0.06 ± 0.01 0.09 ± 0.04 0.08 ± 0.04 5.622 0.50 0.98
VLDL triglycerides (mmollL) 0.87 ± 0.46 1 .84 ± I .45 1 .35 ± I .04 I .29 ± 0.67 4.7 1 2 1 .32 1.85
LDL triglycerides (mmoLfL) 0.66 ± 0.19 0.98 ± 0.52 0.80 ± 0.33 0.92 ± 0.38 3.35 0.62 0.46
HDL2 triglycerides (mmol/L) 0.12 ± 0.04 0.12 ± 0.07 0.14 ± 0.07 0.18 ± 0.13 0.007 0.29 0.29
HDL triglycerides (mmollL) 0.18 ± 0.06 0.17 ± 0.04 0.20 ± 0.07 0.21 ± 0.07 0.04 0.45 0.1 1

‘ 1 ± SD. AT. adipose tissue; 0011’, oral-glucose-tolerance test; L, sum of values; VLDL, very-low-density lipoproteins; LDL, low-density
lipoproteins; HDL, high-density lipoproteins.
2P < 0.05.
3 During 0011’.

58 obese and 29 lean control men, and demonstrated that visceral elderly women, noted a significant association between central
AT, as evaluated by computed tomography, is associated with body-fat distribution and serum cholesterol, serum triglycerides,
insulin concentrations only in the group of obese men. They also and LDL cholesterol. In this same study, moreover, a close as-
demonstrated that obese men with a large amount of visceral AT sociation emerged between central AT distribution and indexes
had higher insulin concentrations in both fasting conditions and of obesity, with the result that it proves difficult to distinguish
during the OGTF than did obese men with only a small amount between the effects of one and the other.
of visceral AT, despite the fact that they had comparable amounts Recently, Leenen et al (29) showed that there are significant
of subcutaneous AT and percentages of body fat. According to differences between the sexes as regards the association between
the results of these authors (8), obesity and/or a certain amount amount of visceral AT and lipoprotein profiles; in women the
of visceral AT is required to observe the adverse metabolic ef- amount of visceral AT is associated with elevated concentrations
fects of visceral obesity or of an increase in visceral AT. of serum triglycerides and with reduced concentrations of HDL
In our study, BMI is associated with glucose concentrations cholesterol, even after adjustment for age and BMI, whereas this
during an OGTT, thus confirming that obesity negatively influ- association disappears in men after such adjustment.
ences glucose tolerance. This result is in line with that of Ohlson Our study showed a positive association between the amount of
et al (26), who showed an interaction between obesity as evalu- visceral AT and triglycerides, apo B, VLDL cholesterol, VLDL
ated by BMI and body-fat distribution as evaluated by waist-to- triglycerides, and LDL triglycerides, and a negative association
hip ratio, affecting the risk of diabetes. It must be stressed that, between visceral AT and HDL:LDL cholesterol. When we per-
in the study by Ohlson et al, the fact that the subjects with the formed multiple comparisons by subdividing the study sample on
lowest tertile of waist-to-hip ratio showed no rise in incidence of the basis of amount of visceral fat and BMI, the effect of BMI on
diabetes with increasing body fatness seems to suggest that re- metabolic variables appears to be very little. On the contrary, the
gional body-fat distribution is the critical factor for observing relationship between visceral AT and lipoprotein concentrations
metabolic disorders (27). seems to be very important, also after adjustment for BMI.
Regional body-fat distribution has been shown to be related Our results are at variance with those of other researchers (7-
not only to insulin-glucose homeostasis alterations but also to 9) and seem to suggest that the relationship between visceral fat
lipoprotein disorders (5, 10, 23, 28, 29). Haarbo et al (28), using and metabolic variables is independent of BMI. These inconsis-
a densitometric procedure to evaluate body-fat distribution in 96 tencies could be explained by differences in methods of evalua-
686 ZAMBONI ET AL

TABLE 4
Age-adjusted mean anthropometric and metabolic variables, and results of ANCOVA for transformed values’

BMI 26.7 BMI > 26.7

Visceral AT Visceral AT Visceral AT Visceral AT F

107 (cm2) > 107 (cm2) 107 (cm2) > 107 (cm2)
Variable (Pt = 13) (n = 7) (n = 5) (n = 10) Visceral AT BMI Interaction

TotalAT(cm2) 166.17 293.20 241.59 334.40 22.662 7.232 1.23

Downloaded from https://academic.oup.com/ajcn/article-abstract/60/5/682/4732014 by East Carolina University user on 11 January 2019

5572
Subcutaneous AT (cm2) 89.54 156.22 146.23 202.24 9.922 0.32
Visceral:subcutaneous AT 0.93 0.94 0.84 0.7 1 0.005 1.28 0. 1 1
Triglycerides (mmol/L) 1.61 3.29 2.55 2.32 6.712 2.66 2.89
Cholesterol (mmollL) 5.67 6.98 6.48 6.57 4.842 1.74 1.93
Apolipoprotein A-I (mg/L) I 177.5 1096.5 1 154.4 1222.4 0.98 0.03 1.04
ApolipoproteinB(mg/L) 1178.2 1511.2 1299.5 1439.4 8.352 1.10 1.12
2-h Glucose (mmol/L)3 7.1 I 6.94 7.94 8.34 0.01 0.77 0.01
E Glucose during 0011’
(mmol/L) 41.99 50.54 51.83 54.38 2.46 3.03 0.53
z Insulin during 0011’
(pmol/L) 1713.0 2558.1 2161.6 1510.2 2.82 1.12 3.13
(UlmL) 285.50 426.35 360.27 251.73 2.82 1.12 3.13
VLDL cholesterol (mmol/L) 0.50 1 .27 0.86 0.72 6.352 1.82 2.89
LDL cholesterol (mmolIL) 3.99 4.81 4.28 4.71 4.042 0.60 0.67
HDL2 cholesterol (mmollL) 0.40 0.29 0.37 0.36 2.92 0.18 0.74
HDL1 cholesterol (mmol/L) 0.78 0.70 0.73 0.77 0.45 0.05 0.15
HDL:LDL cholesterol 0. 1 1 0.06 0.09 0.08 6.50 0.62 1.37
VLDL triglycerides (mmol/L) 0.87 1 .85 1 .36 1 .29 4.622 1 .32 1.85
LDL triglycerides (mmol/L) 0.66 0.97 0.79 0.93 2.92 0.56 0.34
HDL2 triglycerides (mmol/L) 0.12 0.12 0.14 0.18 0.07 0.22 0.50
HDL triglycerides (mmol/L) 0.18 0.17 0.20 0.21 0.1 1 0.37 0.23

‘ AT, adipose tissue; 0011’, oral-glucose-tolerance test; I, sum of values; VLDL, very-low-density lipoproteins; LDL, low-density lipoproteins;
HDL, high density lipoprotcins.
2P < 0.05.
3 During 0011’.

tion of regional body-fat distribution, subject characteristics, and studies (8, 29), in which no subjects (8) or only a few subjects
criteria for subdividing the study sample with regard to the BMI (29) were smokers. Because the relationship between visceral AT
cutoff. and metabolic variables remained unchanged after adjustment for
In some studies, body-fat distribution was measured by waist- smoking habits, our data seem to suggest that the relationship
to-hip ratio (9, 12, 13), but recent hypotheses concerning the between visceral AT and metabolic variables is independent of
associations between body-fat distribution and glucose tolerance smoking habits and thus smoking cannot be the reason for the
and lipid alterations emphasize the importance of visceral fat and lack of consistency between our study and the others.
thus the need to use a direct method, such as computed tomog- Our data also seem to show that the relationship between vis-
raphy, to evaluate regional fat distribution (30, 31). Only a few ceral AT and metabolic variables is independent of age, because
studies have used direct methods to evaluate the amount of vis- the relationship of visceral AT to metabolic variables persisted
ceral fat (7, 8, 10, 1 1). Some were conducted in obese women after adjustment for age.
(7, 10, 29), and Leenen et al (29) demonstrated that sex must be In our study the subjects with higher BMI showed higher total
considered when evaluating the relationships between visceral AT and subcutaneous AT values; the association between BMI
fat and metabolic variables. On this basis our study can be com- and total and subcutaneous AT has been demonstrated ( 15), and
pared only with those of a few researchers who have studied the the fact that the amount of subcutaneous and total AT is similar
relationships of visceral fat (evaluated by direct methods) and in the nonobese subjects with higher visceral AT values and in
obesity to metabolic variables in men (8, 29). obese subjects with lower amounts of visceral AT despite the
In our study the BMI cutoff used for subdividing obese and significant differences in metabolic variables seems to be further
nonobese subjects is a little lower than that adopted by Pouliot evidence of the independent role of visceral fat.
et al (8). The mean BMI values of the obese men in the other Our study appears to demonstrate that the subjects with the
studies (8, 29) are higher than in our study, and therefore we may highest amounts of visceral AT, whether obese or normal-weight,
suppose that differences in the BMI cutoff or differences in the show the worst patterns of metabolic variable. The lack of any
degree of obesity may explain our different results. Our subjects significant interaction between visceral AT and BMI on the met-
are overweight or almost obese men and this may account for abolic variables analyzed means that the average differences we
the lack of consistency with the previous study. found between subjects with visceral AT 107 and subjects with
Differences in smoking habits may be an important factor in visceral AT > 107 are the same in both obese and nonobese
explaining the discrepancies between our present and previous males. Obviously, it is possible that this nonsignificant result may
 VISCERAL FAT, OBESITY, AND METABOLIC VARIABLES 687

be due to low power of the statistical test for the interaction and tissue volume. In: Berry EM. Blondheim SH. Elihau HE. Shafrir E.
the finding needs to be confirmed in a larger number of subjects. eds. Recent advances in obesity research: V London: John Libbey

In any case, our data do not confirm that obesity is a prerequisite Publishing 1987:66-76.
16. Havel Ri, Eder MA, Bragdon JM. The distribution and chemical
for observing a significant association between visceral fat and
composition of ultracentrifugally separated lipoproteins in human
metabolic disorders (7-9). In conclusion, our study seems to
serum. I Clin Invest 1955;34:1345-53.
confirm that visceral AT is the main factor responsible for met-
17. Russel 0. Warnik JJ. A comprehensive evaluation of the heparmn-
abolic abnormalities. The associations with visceral adipose tis-
manganese precipitation procedure for estimating high density Ii-
sue seem to be independent of BMI. U poprotein cholesterol. J Lipid Res 1978;19:65-76.

Downloaded from https://academic.oup.com/ajcn/article-abstract/60/5/682/4732014 by East Carolina University user on 11 January 2019

18. Block WD, Janet Mi Jr. Levine JR. Use of a single color reagent to
References improve the automated determination of serum total cholesterol. In:
Skeggs LT Jr. ed. Automation in analytical chemistry. New York:
I . Bjorntorp P. Regional obesity. In: Bjorntorp P. Brodoff BN. eds.
Mediad. 1974:345-53.
Obesity. New York: J B Lippincott Company, 1992:579-86.
19. Kessler 0. Lederer H. Fluorimetric measurement of triglycerides.
2. Bouchard C, Bray GA, Hubbard VS. Basic and clinical aspects of
In: Skeggs LT Jr. ed. Automation in analytical chemistry. New York:
regional fat distribution. Am J Clin Nutr 1990;52:946-50.
Mediad, 1965:341-50.
3. Larsson B. Obesity. fat distribution and cardiovascular disease. Int
20. Cheung MC, Albers JJ. The measurement of apolipoprotein A-I and
I Obes 1991;l5:53-7.
A-Il levels in men and women by immunoassay. J Clin Invest
4. Bjomtorp P. Abdominal obesity and the development of noninsulin-
1977;60:43-50.
dependent diabetes mellitus. Diabetes Metab Rev l988;4:6l5-22.
21. Sweden OW, Cochran WG. Statistical methods. Ames, IA: Iowa
5. Despr#{233}sJP. Obesity and lipid metabolism: relevance of body fat
State University Press, 1967.
distribution. Curr Opin Lipidol 1991;2:5-15.
22. Kissebah AH, Peiris AN. Biology of regional body fat distribution:
6. Zamboni M, Armellini F, Sheiban I, et al. Relation of body fat dis-
relationship to non-insulin-dependent diabetes mellitus. Diabetes
tribution in men and degree of coronary narrowings in coronary
Metab Rev 1989:5:83-109.
artery disease. Am I Cardiol l992;70:l 135-8.
23. Fujioka S. Matsuzawa Y, Togunaga K. Tarui S. Contribution of
7. Pouliot MC, Despr#{233}sJP. Moorjani S. et al. Computed tomography
intra-abdominal fat accumulation to the impairment of glucose and
measured trunk fat and plasma lipoprotein levels in nonobese
lipid metabolism in human obesity. Metabolism 1987:36:54-9.
women. Metabolism l989;38: 1244-50.
24. Seidell JC, Bjorntorp P. Sj#{246}strOmL, Kvist H, Sannerstedt R. Visceral
8. Pouliot MC, Despr#{233}s
JP, Nadeau A, et al. Visceral obesity in men.
fat accumulation in men is positively associated with insulin. glu-
Association with glucose tolerance, plasma insulin, and lipoprotein
cose, and C-peptide levels, but negatively with testosterone levels.
levels. Diabetes 1992;41 :826-34.
Metabolism 1990:39:897-901.
9. Landin K. Krotkiewski M. Smith U. Importance of obesity for the
25. Zamboni M, Armellini F, Milani MP, et al. Body fat distribution in
metabolic abnormalities associated with an abdominal fat distribu-
pre- and post-menopausal women: metabolic and anthropometnc
tion. Metabolism 1989;38:572-6.
variables and their inter-relationships. Int J Obes 1992:16:495-504.
10. Despr#{233}s
JP, Moorjani S. Ferland M, et al. Adipose tissue distribution
26. Ohlson LO, Larsson B, Svardsudd K. et al. The influence of body
and plasma lipoprotein levels in obese women: importance of intra-
abdominal fat. Atherosclerosis 1989;9:203- 10.
fat distribution on the incidence of diabetes mellitus. I 3.5 years of
I I. Peiris AN, Sothmann MS. Hoffmann RG. Adiposity, fat distribution,
follow up of the participants in the study of men born in 1913.

and cardiovascular risk. Ann Intern Med l989;l 10:867-72. Diabetes 1985:34:1055-8.

12. Larsson B, Svardsudd K, Welin L, Wilhelmsen L, Bjorntorp P. 27. Bray GA. Obesity increases risk for diabetes. Int J Obes
Tibblin 0. Abdominal adipose tissue distribution, obesity. and risk 1992;l6(suppl 4):Sl3-7.
of cardiovascular disease and death: I 3 year follow-up of partici- 28. Haarbo J, Hassager C, Riis BJ, Christiansen C. Relation of body fat

pants in the study of men born in 1913. Br Med i l984;288:l4Ol- distribution to serum lipids and lipoproteins in elderly women. Ath-
4. erosclerosis 1989:80:57-62.
13. Lapidus L, Bengtsson C. Larsson B, Pennert K, Rybo E, Sj#{246}strOm 29. Leenen R, van der Kooy K, Seidell JC, Deurenberg P. Visceral fat
L. Distribution of adipose tissue and risk of cardiovascular disease accumulation measured by magnetic resonance imaging in relation
and death: a 12 year follow-up of participants in the population study to serum lipids in obese men and women. Atherosclerosis
of women in Gothenburg, Sweden. Br Med J l984;289:1257-6l. 1992;94:l7l -81.
14. Crepaldi 0, Belfiore F, Bosello 0. et al. Special report: Italian Con- 30. Van der Kooy K, Seidell JC. Techniques for the measurement of
sensus Conference-overweight. obesity and health. Int J Obes visceral fat: a practical guide. Int J Obes l993;17:187-96.
1991;l5:781 -90. 31 . Zamboni M, Armellini F, Milani MP, et al. Evaluation of regional
15. Sj#{246}strOm L. New aspects of weight-for-height indices and adipose body fat distribution: comparison between WIH ratio and computed
tissue distribution in relation to cardiovascular risk and total adipose tomography in obese women. J Intern Med 1992:232:341-7.