Effects of a moderate low-carbohydrate diet on preferential abdominal fat loss and cardiovascular risk factors in patients with type 2 diabetes.

Abstract

Reports have shown that visceral adipose tissue (VAT) is more closely linked to cardiovascular risk factors (CRFs) than subcutaneous adipose tissue (SAT). We aimed to elucidate preferential abdominal fat loss and the correlations between abdominal fat reductions and changes in CRFs achieved with a moderate low-carbohydrate diet (LCD) in patients with type 2 diabetes (T2DM).
Fifty-two outpatients (28 men and 24 women, mean age ± SD: 60.0 ± 10.5 years) with hemoglobin A(1c) (HbA(lc)) levels ≥ 6.5% were on an LCD for 6 months. Over a 6-month period, we measured their abdominal fat distribution (using CT) and assessed CRFs, including body mass index (BMI), HbA(1c), fasting blood glucose (FBG), serum insulin, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglyceride levels.
The patients showed good compliance with the LCD (1812 ± 375 kcal/day, % carbohydrate:fat:protein = 35:40:19 for men; 1706 ± 323 kcal/day, % carbohydrate:fat:protein = 41:36:21 for women). Significant decreases (P = 0.05) in BMI and HbA(1c) levels were observed, along with an increase in HDL-C (P = 0.021) in men and a decrease in LDL-C (P = 0.001) in women. VAT (-21.6 cm(2), P < 0.001 in men; -19.6 cm(2), P < 0.001 in women) and SAT (-13.5 cm(2), P = 0.004 in men; -19.1 cm(2), P = 0.003 in women) significantly decreased. The loss of VAT (%ΔVAT) was greater than that of SAT (%ΔSAT) in women (P = 0.022). A similar but not significant predominance of VAT loss was detected in men (P = 0.111). In women, the %ΔSAT significantly correlated with changes in FBG (ΔFBG) (r = 0.417) and HDL-C (ΔHDL) (r = -0.720), as was %ΔVAT with changes in triglyceride (ΔTG) (r = 0.591).
Six months of a moderate LCD resulted in preferential VAT loss only in women, with significant correlations between %ΔSAT and both ΔHDL and ΔFBG, as well as between %ΔVAT and ΔTG. Our results suggest that an LCD has the potential to reduce abdominal fat in patients with T2DM and deterioration of serum lipid profiles.

Full text

Page 1

© 2011 Sasakabe et al, publisher and licensee Dove Medical Press Ltd. This is an Open Access article
which permits unrestricted noncommercial use, provided the original work is properly cited.
Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 2011:4 167–174
Diabetes, Metabolic Syndrome and Obesity: Targets and TherapyDovepress
submit your manuscript | www.dovepress.com
Dovepress
167
O r i g i n A L r e S e A r c h
open access to scientific and medical research
Open Access Full Text Article
effects of a moderate low-carbohydrate diet on
preferential abdominal fat loss and cardiovascular
risk factors in patients with type 2 diabetes
DOI: 10.2147/DMSO.S19635
Tae Sasakabe1
hajime haimoto2
hiroyuki Umegaki3
Kenji Wakai4
1Department of clinical nutrition,
haimoto clinic, Yayoi, Kasugai, Aichi,
Japan; 2Department of internal
Medicine, haimoto clinic, Yayoi,
Kasugai, Aichi, Japan; 3Department
of geriatrics, nagoya University
graduate School of Medicine,
Tsurumai, Showa, nagoya, Aichi, Japan;
4Department of Preventive Medicine,
nagoya University graduate School
of Medicine, Tsurumai, Showa, nagoya,
Aichi, Japan
correspondence: Tae Sasakabe
haimoto clinic, 1-80 Yayoi,
Kasugai, Aichi 486-0838, Japan
Tel +81 568 85 8226
Fax +81 568 85 8315
email haiclipar-mail78@gol.com
Background: Reports have shown that visceral adipose tissue (VAT) is more closely linked
to cardiovascular risk factors (CRFs) than subcutaneous adipose tissue (SAT). We aimed to
elucidate preferential abdominal fat loss and the correlations between abdominal fat reductions
and changes in CRFs achieved with a moderate low-carbohydrate diet (LCD) in patients with
type 2 diabetes (T2DM).
Patients and methods: Fifty-two outpatients (28 men and 24 women, mean age ± SD:
60.0 ± 10.5 years) with hemoglobin A1c (HbAlc) levels $ 6.5% were on an LCD for 6 months.
Over a 6-month period, we measured their abdominal fat distribution (using CT) and assessed
CRFs, including body mass index (BMI), HbA1c, fasting blood glucose (FBG), serum insulin,
high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C),
and triglyceride levels.
Results: The patients showed good compliance with the LCD (1812 ± 375 kcal/day, % carbo-
hydrate:fat:protein = 35:40:19 for men; 1706 ± 323 kcal/day, % carbohydrate:fat:protein =
41:36:21 for women). Significant decreases (P = 0.05) in BMI and HbA1c levels were observed,
along with an increase in HDL-C (P = 0.021) in men and a decrease in LDL-C (P = 0.001)
in women. VAT (–21.6 cm², P , 0.001 in men; –19.6 cm², P , 0.001 in women) and SAT
(–13.5 cm², P = 0.004 in men; –19.1 cm², P = 0.003 in women) significantly decreased. The loss
of VAT (%∆VAT) was greater than that of SAT (%∆SAT) in women (P = 0.022). A similar
but not significant predominance of VAT loss was detected in men (P = 0.111). In women, the
%∆SAT significantly correlated with changes in FBG (∆FBG) (r = 0.417) and HDL-C (∆HDL)
(r = –0.720), as was %∆VAT with changes in triglyceride (∆TG) (r = 0.591).
Conclusion: Six months of a moderate LCD resulted in preferential VAT loss only in women,
with significant correlations between %∆SAT and both ∆HDL and ∆FBG, as well as between
%∆VAT and ∆TG. Our results suggest that an LCD has the potential to reduce abdominal fat
in patients with T2DM and deterioration of serum lipid profiles.
Keywords: low-carbohydrate diet, visceral adipose tissue, subcutaneous adipose tissue,
cardiovascular risk factors
Introduction
Low-carbohydrate diets (LCDs) have attracted interest for weight loss, glycemic con-
trol, and the management of cardiovascular risk factors in overweight patients with
type 2 diabetes (T2DM).1,2 Among their reported advantages over high-carbohydrate
diets (HCDs) are improvements in both serum high-density lipoprotein cholesterol
(HDL-cholesterol) levels and triglyceride levels in obese patients.3–5
The proportion of energy obtained from carbohydrates in East Asian populations
is generally higher (about 60%) and the fat percentage generally lower (about 25%)
Number of times this article has been viewed
This article was published in the following Dove Press journal:
Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy
26 April 2011

Page 2

Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 2011:4
submit your manuscript | www.dovepress.com
Dovepress
Dovepress
168
Sasakabe et al
compared with Western populations.6–9 Therefore, com-
pared with HCDs, moderate (non-ketogenic) LCDs, but
not strict (ketogenic) LCDs, may be sufficiently effective
for glycemic control in East Asian patients with T2DM.
Moreover, a moderate LCD is more acceptable to patients
with T2DM because they are not required to strictly limit
carbohydrates.10,11 Indeed, we demonstrated the efficacy of
a moderate LCD (30%–45% carbohydrate diet) in glycemic
control, weight loss, and improvement of serum lipid profiles
in T2DM patients with a low attrition rate.11,12 Even in the
Western population, strict and moderate LCDs are reported
to be equally effective in controlling body weight and serum
lipid profiles.13
It has been proposed that abdominal fat accumulation
raises cardiovascular risk. As compared with subcutane-
ous adipose tissue (SAT), visceral adipose tissue (VAT)
in particular is more closely linked to glucose intolerance,
hypertension, dyslipidemia, and insulin resistance.14–16
Therefore, VAT is widely assumed to be a major predic-
tor of metabolic syndrome and insulin resistance. Indeed,
experimental studies have demonstrated that VAT and
SAT play different physiological and biochemical roles.16
Men have significantly more VAT than women, which
may explain why women have fewer cardiovascular
events than men.17 Thus, if moderate LCDs result in a
greater decrease in VAT than SAT, such LCDs may be
useful for controlling not only blood glucose levels, but
also abdominal fat.
Many intervention studies involving HCDs and either
exercise or drugs have been conducted to assess changes
in abdominal fat distribution in obese subjects. However,
although HCDs can result in absolute VAT and SAT loss,
these diets do not result in preferential abdominal fat loss.18
To date, only a few studies have examined the effects of
LCDs on preferential abdominal fat loss and the associa-
tions between the loss of adipose tissue and improvement in
cardiovascular risk factors in nonobese and obese patients
with and without T2DM.19
Therefore, in the present study, we investigated which
abdominal fat tissues (VAT or SAT) would be preferentially
reduced by a 6-month moderate LCD. The subjects were
non-obese patients with T2DM who were typical of those
commonly treated in Japan.20,21 We also sought to determine
whether decreases in VAT or SAT were associated with
changes in cardiovascular risk factors (body mass index
[BMI], blood pressure [BP], hemoglobin A1c [HbAlc] levels,
fasting blood glucose [FBG], serum insulin [IRI] and serum
lipid profiles).
Patients and methods
We invited all new outpatients with T2DM and HbA1c levels
of 6.5% or above to participate in a 6-month intervention
study from April 2007 to November 2008 at the Haimoto
Clinic. Patients with serum creatinine levels . 1.5 mg/dL,
severe complications from diabetes (proliferative retinopathy,
symptomatic neuropathy, or a diabetic foot), ketoacidosis,
soft drink ketosis, cancer, severe heart failure, and liver
cirrhosis were excluded. Five patients on hormone and/or
insulin treatment were also excluded due to the considerable
effects of these drugs on body weight (BW) and serum insulin
levels. The remaining 63 patients participated in the study.
We measured the participants’ BW, BP, and HbA1c levels
every month. Levels of FBG and serum insulin, triglyceride
(TG), low-density lipoprotein cholesterol (LDL-C) and
HDL-C were analyzed in the fasting state at baseline and after
6 months. Changes in VAT, SAT, total adipose tissue (TAT)
and waist circumference (WC) were also assessed at baseline
and after 6 months by computed tomography (CT).
All participants gave informed consent, and the study
protocol was approved by the Ethics Committee of the
Nagoya Tokushukai General Hospital.
Laboratory methods
Plasma glucose concentrations were determined by enzy-
matic methods (Shino-Test Co. Kanagawa, Japan). Serum
insulin levels were measured using the standard double
antibody radioimmunoassay method (Fujirebio Inc. Tokyo,
Japan). Enzymatic methods were used to measure TG levels
(Daiichi Pure Chemicals Co., Tokyo, Japan). Direct methods
were used to assay serum LDL-C and HDL-C levels (Daiichi
Pure Chemicals Co., Tokyo, Japan). HbA1c levels were
measured using high-performance liquid chromatography
(Arkley Co., Kyoto, Japan).
Assessment of body composition
BW was determined using an electronic scale while patients
were wearing only underwear. BMI was calculated by divid-
ing weight in kilograms by the square of height in meters. To
eliminate the effect of sex and age, we used the standard devia-
tion score of BMI as well as crude BMI values. The score was
computed using the following formula: (BMI – mean)/SD.
We used the average and standard deviation of BMI by sex
and 10-year age group in the National Health and Nutrition
Survey, Japan, 200922 because of the small number of partici-
pants in some age groups in the present study.
The surface areas of VAT and SAT were measured by CT
(Hitachi Medical Corporation X-ray CT system Pronto Si), and

Page 3

Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 2011:4
submit your manuscript | www.dovepress.com
Dovepress
Dovepress
169
Abdominal fat loss with a low-carbohydrate diet
patients were scanned at the navel. These areas and WC were
measured using image analysis software (Hitachi Medical Cor-
poration Fat Pointer, version 1.10) as described previously.23
The carbohydrate-reduced diet
The main principle of the LCD used in this study was
to eliminate carbohydrate-rich foods depending on each
patient’s HbA1c levels, as described previously.11,12 In brief,
patients with HbA1c levels , 9.0% were asked to eliminate
carohydrate-rich foods from their dinner, whereas patients
with an HbA1c level $ 9.0% were asked to eliminate car-
bohydrates twice a day (from breakfast and dinner). It was
easy for patients to stay on the LCD because no strict
protein:fat:carbohydrate (PFC) ratio was prescribed, and they
were permitted to eat as much protein and fat as they wanted,
including saturated fat. There were no other restrictions.
A dietician (TS) gave instructions to all participants
3 times during the first month and once a month thereafter.
The participants were instructed to maintain their baseline
physical activity level throughout the study period. Their
dietary intake was assessed after 5 months from 3-day
food records and an interview with a dietitian and the data
were analyzed using the HealthyMaker432 (mushroomsoft,
Okayama, Japan). The participants had not previously been
instructed to follow a LCD before this intervention. They
were asked about changes in their physical activity levels on
a questionnaire that they filled out at the end of the study.
Statistical analysis
Data are presented as mean ± SD. Either the paired t-test or
its nonparametric counterpart (Wilcoxon signed rank test)
was used to assess changes in all the parameters during the
6-month period. The unpaired t-test was applied to assess the
differences in VAT and SAT at baseline and in the percentage
of dietary macronutrient intake between men and women.
We compared the ratio of VAT area to SAT area (V/S
ratio) at baseline and after the 6-month period. The changes in
some variables during the 6-month study period (ie, 6-month
value minus baseline value) are presented as ∆variable
(eg, ∆BMI, ∆SBP, ∆DBP, ∆FBG, ∆IRI, ∆HbA1c, ∆TG,
∆LDL-C, and ∆HDL-C). We used %∆VAT, %∆SAT and
∆V/S as indicators of differential changes in VAT or SAT,
although various parameters have been used in previous inter-
vention studies (eg, ∆VAT, ∆SAT, ∆VAT/∆SAT, %∆VAT,
%∆SAT, %∆VAT/%∆SAT, V/S, and ∆V/S).18,19,24–29 The
parameters %∆VAT and %∆SAT were computed using the
following formula: 100 – (100 × [FAT area at 6 months/FAT
area at baseline]).
The correlations between abdominal fat distribution and
either cardiovascular risk factors or body composition at
baseline and the correlations between changes in these vari-
ables over the 6-month period were assessed with Spearman
correlation coefficients. Multiple regression analyses adjusted
for age were performed on the significant and strong correla-
tions between changes in abdominal fat and cardiovascular
risk factors. All statistical analyses were performed using the
SPSS statistical package (version 11.0; SPSS, Inc., Chicago,
IL). P values less than 0.05 were considered to be statisti-
cally significant.
Results
Baseline characteristics of patients
Of the 63 patients enrolled in the study, 2 patients voluntarily
dropped out, 2 refused the CT assessment at 6 months, 3 began
to require insulin treatment, and 4 were hospitalized because
of diseases other than T2DM. After excluding these patients,
52 were left for the analyses (28 men, mean age ± SD:
57.3 ± 10.6 years, range: 35–73 years and 24 women, mean
age ± SD: 63.6 ± 9.5 years, range: 34–82 years). All but one
of the female participants were postmenopausal.
The patients’ characteristics at baseline are shown in
Table 1. The mean BMI was 25.4 ± 4.5 kg/m² for men and
23.9 ± 3.4 kg/m² for women, and only 5 patients exceeded
30.0 kg/m². The mean standard deviation score of BMI was
0.49 ± 1.34 for men and 0.30 ± 0.93 for women.
HbA1c levels were 8.4% ± 1.5% in men and 8.6% ± 2.0%
in women. VAT was much larger in men than in women
(men: 150.9 ± 57.5 cm², women: 99.6 ± 51.7 cm²), whereas
SAT was larger in women (men: 123.4 ± 70.7 cm², women:
187.3 ± 70.0 cm²). No changes in physical activity levels
during the study period were reported in the questionnaire.
As for the correlations of VAT or SAT distribution and
WC with cardiovascular risk factors in men, we found a
strong positive correlation between VAT and BMI (Spearman
correlation coefficient [r] = 0.669, P , 0.001). There were
also strong, positive correlations between SAT, TAT, or WC
and either BMI (r = 0.590, 0.819 or 0.806, respectively and
P # 0.001 for all) or serum insulin levels (r = 0.655, 0.565,
0.673 and P , 0.001, P = 0.002, P , 0.001, respectively).
Moderate inverse correlations between either SAT or WC
and serum HDL-C levels (r = –0.473, –0.472 and P = 0.020
or P = 0.020, respectively) were also observed. In women,
we found strong positive correlations between VAT, SAT,
TAT, or WC and either BMI (r = 0.652, 0.730, 0.796, or 0.865
and P # 0.001 for all) or serum insulin levels (r = 0.658,
0.663, 0.762, or 0.759 and P , 0.001 for all). A significant

Page 4

Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 2011:4
submit your manuscript | www.dovepress.com
Dovepress
Dovepress
170
Sasakabe et al
inverse correlation between VAT and serum HDL-C levels
(r = 0.567 and P = 0.011) was also observed. Applying the
standard deviation score of BMI did not materially change
the correlations of VAT, SAT, TAT, or WC with BMI in
both sexes (P # 0.001 for all).
Assessment of dietary intake
Total energy intake, assessed by 3-day food records, was
1812 ± 375 kcal/day for men and 1706 ± 323 kcal/day for
women. The average daily percentages of intake for carbo-
hydrates, fat, and protein (% energy) were 34.5% ± 8.0%,
39.6% ± 9.0%, and 19.1% ± 3.4% for men and 40.6% ± 8.1%,
36.4% ± 7.4%, and 20.5% ± 2.4% for women, respectively
(Table 2). The carbohydrate intake (% energy) was signifi-
cantly higher in women than in men (P = 0.010).
changes in anthropometric variables,
cardiovascular risk factors, and antidiabetic
drug use over 6 months
In men, the mean BMI (-0.7 ± 1.2 kg/m²), the standard devia-
tion score of BMI (-0.23 ± 0.37), HbA1c (-1.8% ± 1.4%) and
FBG (-20 ± 42 mg/dL) were significantly decreased, and serum
HDL-C (4.5 ± 8.8 mg/dL) levels were significantly increased
after 6 months. However, no significant changes were observed
in the serum LDL-C and TG levels (Table 1). In women, how-
ever, the mean BMI (-0.8 ± 1.0 kg/m²), the standard deviation
score of BMI (-0.22 ± 0.30), HbA1c (-1.7% ± 1.6%) and
LDL-C levels (-24 ± 27 mg/dL) were significantly decreased,
whereas the mean serum TG and HDL-C levels were not sig-
nificantly altered after 6 months (Table 1).
Sixteen of the patients had already been prescribed antidi-
abetic drugs by other physicians before initiation of the study.
The number of patients (n) who took sulfonylureas decreased
and some of the remaining patients’ doses were lowered over
the 6 months (glibenclamide: n; from 5 to 1, mean dose; from
5 to 2.5 mg/day; glimepiride: n; from 6 to 7, mean dose; from
2.8 to 1.5 mg/day; gliclazide: n; from 1 to 1, mean dose;
from 80 to 20 mg/day). As for other antidiabetic drugs, the
number of patients (n) who received nateglinide, mitiglinide,
voglibose, and pioglitazone also decreased (nateglinide: n;
from 1 to 0, mitiglinide: n; from 1 to 0, pioglitazone: n; from
4 to 1, voglibose: n; from 8 to 3). The number of patients who
took metformin slightly increased (from 5 to 7), although the
mean dose decreased from 600 to 500 mg/day.
Table 1 changes in abdominal fat distribution and cardiovascular risk factors during the 6-month moderate low-carbohydrate diet
Men (n = 28)
Baseline
Women (n = 24)
BaselineAfter 6 monthsP After 6 monthsP
Body weight (kg)
Body mass index (kg/m²)
Standard deviation score of body mass index
Systolic blood pressure (mmhg)
Diastolic blood pressure (mmhg)
hemoglobin A1c (%)
Fasting blood glucose (mg/dL)
Fasting serum insulin (μiU/mL)
Serum triglyceride (mg/dL)
Serum LDL cholesterol (mg/dL)
Serum hDL cholesterol (mg/dL)
VAT (cm²)
∆VAT (cm²)
%∆VAT
SAT (cm²)
∆SAT (cm²)
%∆SAT
TAT (cm²)
∆TAT (cm²)
%∆TAT
V/S ratio
Waist circumference (cm)
70.7 ± 13.6
25.4 ± 4.5
0.49 ± 1.34
132 ± 23
81 ± 15
8.4 ± 1.5
157 ± 47
9.11 ± 9.11
126 ± 120
134 ± 39
47 ± 13
150.9 ± 57.5
68.7 ± 12.7
24.6 ± 4.1
0.26 ± 1.24
133 ± 16
78 ± 9
6.5 ± 0.6
136 ± 29
8.18 ± 5.15
110 ± 77
125 ± 29
52 ± 14
129.4 ± 57.8
-21.6 ± 26.2
-14.2 ± 21.6
109.9 ± 63.8
-13.5 ± 22.8
-9.0 ± 23.1
239.3 ± 103.6
-35.1 ± 42.1
-12.8 ± 19.3
1.30 ± 0.58
86.3 ± 10.2
0.004*
0.003*
0.004*
0.864
0.813
0.000
0.024
0.467*
0.361*
0.076*
0.021*
0.000
55.8 ± 7.2
23.9 ± 3.4
0.30 ± 0.93
143 ± 22
80 ± 13
8.6 ± 2.0
164 ± 50
6.63 ± 3.61
112 ± 63
157 ± 35
59 ± 14
99.6 ± 51.7
54.1 ± 7.4
23.2 ± 3.4
0.08 ± 0.93
134 ± 19
78 ± 9.6
7.0 ± 1.0
144 ± 33
6.87 ± 3.81
105 ± 59
133 ± 28
60 ± 16
80.0 ± 40.7
-19.6 ± 19.5
-18.9 ± 19.0
168.1 ± 67.3
-19.1 ± 29.9
-8.8 ± 17.9
248.1 ± 94.6
-38.8 ± 36.0
-13.4 ± 15.2
0.49 ± 0.21
84.6 ± 9.0
0.002
0.002
0.002
0.142
0.460
0.000*
0.059
0.634*
0.468*
0.001*
0.319
0.000*
123.4 ± 70.7
0.004
187.3 ± 70.0
0.003*
274.4 ± 106.4
0.000
286.9 ± 103.5
0.000
1.40 ± 0.67
88.9 ± 10.1
0.069*
0.000
0.57 ± 0.30
87.0 ± 8.8
0.013*
0.002
Notes: Values are means ± SD. Patients receiving lipid-lowering drugs (4 men and 5 women) were excluded from the analysis of lipid profiles, as were patients
receiving antihypertensive drugs (11 men and 10 women) from the analysis of blood pressure; *nonparametric counterpart (Wilcoxon signed rank test).
Abbreviations: VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue; TAT, total adipose tissue; V/S ratio, ratio of visceral adipose tissue area to subcutaneous
adipose tissue area.

Page 5

Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 2011:4
submit your manuscript | www.dovepress.com
Dovepress
Dovepress
171
Abdominal fat loss with a low-carbohydrate diet
Table 2 Daily dietary intake at 5 months after baseline
Men (n = 28)
1812 ± 375
153.8 ± 37.7
2.7 ± 1.0
10.1 ± 3.7
34.5 ± 8.0
79.4 ± 24.3
21.7 ± 7.2
32.7 ± 11.1
15.5 ± 5.6
39.6 ± 9.0
85.4 ± 19.7
19.1 ± 3.4
Women (n = 24)
1706 ± 323
171.8 ± 40.5
3.1 ± 1.0
10.6 ± 4.0
40.6 ± 8.1
69.6 ± 21.7
17.4 ± 6.8
27.7 ± 10.0
15.8 ± 5.7
36.4 ± 7.4
86.8 ± 15.4
20.5 ± 2.4
Total energy intake (kcal)
carbohydrate (g)
Soluble fiber (g)
Insoluble fiber (g)
carbohydrate (% energy)
Fat (g)
Saturated fat (g)
Monounsaturated fat (g)
Polyunsaturated fat (g)
Fat (% energy)
Protein (g)
Protein (% energy)
Foods rich in carbohydrate
rice (g)
noodles (g)
Bread (g)
Potatoes (g)
Fruit (g)
confectioneries (g)
Sugar (g)
166 ± 85
41 ± 46
24 ± 35
18 ± 34
34 ± 73
17 ± 24
7 ± 7
160 ± 57
57 ± 61
19 ± 29
22 ± 46
64 ± 78
19 ± 16
8 ± 7
Note: values are means ± SD.
changes in abdominal fat distribution
The parameters ∆VAT, ∆SAT, and WC were significantly
decreased in both men and women during the 6-month period
(P , 0.05 for all, Table 1).
In men, despite an absolute decrease in VAT
(–21.6 ± 26.2 cm²), %∆VAT (–14.2% ± 21.6%) was not
significantly different from %∆SAT (–9.0% ± 23.1%)
(P = 0.111). However, a significant decrease in V/S
(P = 0.069) was not observed (Table 1). In women, along with
an absolute decrease in VAT (–19.6 ± 19.5 cm²), %∆VAT
(–18.9% ± 19.0%) was significantly higher than %∆SAT
(–8.8% ± 17.9%) (P = 0.022), and V/S was also significantly
decreased (P = 0.013).
Associations between changes
in abdominal fat distribution
and cardiovascular risk factors
In men, no significant correlations were observed between
changes in abdominal adipose tissue distribution (∆VAT,
∆SAT, ∆TAT, %∆VAT, %∆SAT, and %∆TAT) and changes
in cardiovascular risk factors (Table 3). Meanwhile, ∆TAT,
%∆SAT, and %∆TAT were strongly correlated with ∆FBG
in women (r = 0.533 for ∆TAT, r = 0.417 for %∆SAT, and
r = 0.461 for %∆TAT, Table 3). Similarly, ∆SAT, ∆TAT,
%∆SAT, and %∆TAT were all strongly and inversely cor-
related with ∆HDL-C in women (r = -0.699 for ∆SAT,
r = –0.567 for ∆TAT, r = –0.720 for %∆SAT, and r = –0.554
for %∆TAT, Table 3). In addition, a significant inverse corre-
lation between ∆WC and ∆HDL-C was detected (r = –0.622),
as was a similar correlation between either ∆SAT or ∆WC
and ∆SBP (r = –0.543 for ∆SAT and r = –0.550 for ∆WC,
Table 3). In contrast, only %∆VAT significantly correlated
with ∆TG (r = 0.591, Table 3).
In multiple regression analyses that were adjusted for age,
the following associations remained significant in women:
between %∆VAT and ∆TG (P = 0.005), between ∆SAT
and ∆HDL-C (P , 0.001), between %∆SAT and ∆HDL-C
(P , 0.001), and between ∆WC and ∆HDL-C (P , 0.001).
When five patients who had received pioglitazone, which
influences abdominal fat distribution and serum lipid profiles,
were excluded, the strong correlations observed in women
were not materially altered (data not shown).
Discussion
This study is the first to show that, in outpatients with T2DM,
a significant decrease in both VAT and SAT in men and
women with good glycemic and BMI control, as well as
a tapering off of sulfonylurea use and an improvement in
serum lipid profiles, can be achieved with a moderate LCD
(about a 38% carbohydrate diet). Significant preferential
VAT loss was achieved in women. VAT loss was greater
than SAT loss in men as well, but the difference did not
reach statistical significance. At the same time, significant
correlations of %∆SAT with ∆FBG and ∆HDL-C and sig-
nificant correlations of %∆VAT with ∆TG were observed
only in women.
Limitations
First, the use of antidiabetic drugs (especially pioglita-
zone) that influence SAT and HDL-C may have affected
the results.30,31 However, in the analysis that excluded
the 4 patients who took pioglitazone, a strong correlation
remained between the change in abdominal fat distribu-
tion and ∆HDL-C in women. Second, all the participants
in this study were non-obese, with an average BMI (±SD)
of 24.7 ± 4.0. Caucasians tend to be more obese and insulin
resistant than East Asian populations.32,33 Therefore, studies
of non-obese patients with T2DM may be more relevant to
East Asian populations. Cross-sectional studies of apparently
lean Japanese individuals demonstrated that a significant
accumulation of VAT was associated with increased risk
factors for cardiovascular disease and diabetes. These results
were similar to what has been observed in obese patients
with T2DM.23,34,35 Third, MRI abdomen is more accurate in