ORIGINAL ARTICLE

Insulin resistance and metabolic syndrome in normal-weightindividuals

Shanying Chen • Youming Chen • Xinyu Liu •

Mi Li • Bide Wu • Yongqiang Li • Yan Liang •

Xiaofei Shao • Harry Holthofer • Hequn Zou

Received: 31 July 2013 / Accepted: 4 October 2013

� Springer Science+Business Media New York 2013

Abstract We performed this study to investigate the

prevalences of insulin resistance and metabolic syndrome

(MetS) in a Chinese population with normal weight. We

also examined whether fat mass is associated with insulin

resistance and MetS in normal-weight individuals. Data

were drawn from a cross-sectional study in China. Subjects

with diabetes were excluded. The prevalences of insulin

resistance and MetS were calculated. Multivariate logistic

regression analysis was performed repeated separately for

body mass index (BMI) and waist circumference (WC) in

both men and women. We also used the combination of

BMI and WC to predict insulin resistance and MetS.

8.55 % of normal-weight men and 12.62 % of normal-

weight women had insulin resistance. 7.41 % of normal-

weight men and 10.24 % of normal-weight women had

MetS. WC was associated with incident insulin resistance

and MetS independent of BMI in both men and women.

BMI was independently associated with incident MetS in

women. Normal-weight individuals with insulin resistance

and/or MetS are not rare in the Chinese population. Fat

mass is associated with insulin resistance and MetS in

normal-weight subjects. The current findings support using

both BMI and WC in clinical practice.

Keywords Insulin resistance �Metabolic syndrome �Normal weight

Abbreviations

MetS Metabolic syndrome

BMI Body mass index

WC Waist circumference

ATP III Third Report of the National Cholesterol

Education Program’s Adult Treatment Panel

CI Confidence interval

HDL-C High-density lipoprotein-cholesterol

HOMA Homeostasis model assessment

OR Odds ratio

CRP C-reactive protein

eGFR Estimated glomerular filtration rate

Scr Serum creatinine

Introduction

Insulin resistance is a hallmark of metabolic syndrome

(MetS) and largely initiated by obesity [1]. The existence

of a subgroup of individuals with normal weight but with

Shanying Chen and Youming Chen have contributed equally to this

study.

S. Chen � X. Liu � Y. Li � Y. Liang � X. Shao � H. Zou (&)

Department of Nephrology, The Third Affiliated Hospital of

Southern Medical University, 183# Zhongshan Dadao, Tianhe

District, Guangzhou 510000, People’s Republic of China

e-mail: hequnzou@hotmail.com

S. Chen � B. Wu

Department of Nephrology, Zhangzhou Affiliated Hospital of

Fujian Medical University, Zhangzhou, People’s Republic of

China

Y. Chen

Clinical Laboratory, The Third Affiliated Hospital of Southern

Medical University, Guangzhou, People’s Republic of China

M. Li

Blood Purification Center, The Fifth Affiliated Hospital of Sun

Yat-sen University, Zhuhai, People’s Republic of China

H. Holthofer

National Centre for Sensor Research/BioAnalytical Sciences,

Dublin City University, Dublin, Ireland

123

Endocrine

DOI 10.1007/s12020-013-0079-8

metabolic disturbances usually associated with obesity was

first suggested in 1981 by Ruderman et al. [2, 3]. Despite

having normal body mass index (BMI), these individuals

presented with insulin resistance and/or hyperinsulinemia.

They also had hypertriglyceridemia, hypertension, and

were at high risk for cardiovascular disease and type two

diabetes [2–8]. There exists only scarce data on insulin

resistance in normal-weight individuals. However, indi-

viduals with normal weight and metabolic syndrome

(MetS) or insulin resistance are not uncommon [3, 4, 7]. To

our knowledge, there are no data on the prevalence of

insulin resistance or MetS in Chinese subjects with normal

weight.

There is limited evidence that higher fat mass is asso-

ciated with insulin resistance in normal-weight individuals

[3–8]. However, most previous studies were conducted on

small sizes [3, 5, 6, 8]. Large cohort studies were limited to

the United State population [3, 7, 9].

We performed this study to investigate the prevalences

of insulin resistance and MetS in a Chinese population with

normal weight. We also examined whether fat mass is

associated with insulin resistance and MetS in normal-

weight individuals. Both BMI and waist circumference

(WC) were used as anthropometric indexes of fat mass.

Methods

Subjects

We used the data from a population-based, cross-sectional

survey conducted in Wanzhai Town, Zhuhai City. Zhuhai

is located on the southern coastline of China. There are six

communities in Wanzhai Town and three of them were

randomly selected for this survey. All adult residents

(18 years or older) living in the three communities were

invited to participate in this survey. Participants were

recruited by mail and home visit. 2,142 residents volun-

tarily participated in this survey. This study was approved

by the ethics committee of the Third Affiliated Hospital of

Southern Medical University. All subjects gave their

informed consent. We excluded subjects with diabetes. We

described the cross-sectional study in our previous paper

[10].

Data collection

All doctors, medical students, and nurses participating in

this study had received intensive training. Data were col-

lected in local community clinic, health stations, or during

home interview. Data on age, sex, education, current or

past cigarette smoking, alcohol use, physical activity, die-

tary habits, and personal history and family history were

obtained using questionnaires. Blood pressure was deter-

mined with the calibrated mercury sphygmomanometer in

a sitting position after at least 5 min rest. Blood pressure

was determined three times and an average of the three

reading was calculated [10].

Anthropometric measurements

Anthropometric indexes were collected in the community

clinics and measured according to the recommendation by

the World Health Organization. WC was measured at the

level of midway between the lower rib margin and the iliac

crest in the midaxillary line, with the participants standing

with light garments and breathing out gently. Hip cir-

cumference was measured at the widest point around the

buttocks [11]. BMI was calculated as weight (in kg) divi-

ded by the square of the height (in m).

Laboratory variables

All blood specimens were collected after an overnight fast

of at least 10 h in local community clinic. First morning

urine samples were collected. Women who were actively

menstruating were excluded from the urine test. All spec-

imens from collection sites were transported to the central

laboratory in the Third Affiliated Hospital of Southern

Medical University in 3 h and stored at 2–8 �C until

analysis.

Serum creatinine, fasting glucose, serum total cholesterol,

serum triglyceride, and serum high-density lipoprotein-cho-

lesterol (HDL-C) were measured by the colorimetric method.

High sensitivity C-reactive protein (CRP) was measured using

enzymatic immunoassay turbidimetric method. Serum insulin

was measured by electrochemiluminescence immunoassay.

Homeostatic model assessment of insulin resistance (HOMA-

IR) was calculated as fasting plasma glucose (mmol/l) 9

fasting insulin (mU/l)/22.5 [12]. Estimated glomerular filtra-

tion rate (eGFR) was calculated as 186 9 (Scr)-1.154 9

(Age)-0.203 9 (if female, 90.742) [13]. Urinary albumin-to-

creatinine ratio (mg/g) was calculated as the ratio of urinary

albumin to urinary creatinine.

Definitions

Normal weight was defined as BMI \ 24 kg/m2 [14].

Insulin resistance was defined as exceeding the 75 % per-

centile of HOMA-IR in normal glucose tolerance subjects.

According to an epidemiological survey in China, the cut-

off point is 2.69 [15]. Abdominal obesity was defined as a

waist circumference C85 cm in men or C80 cm in women

[14]. Diabetes mellitus was defined as a fasting serum

Endocrine

123

glucose C7.0 mmol/l and/or self-reported diagnosis of

diabetes. A family history of diabetes or hypertension was

defined as at least one diabetic parent or hypertensive

parent. MetS was defined as having at least three of the

following five criteria: abdominal obesity, elevated tri-

glyceride levels (C150 mg/dl), low HDL-C levels

(\40 mg/dl in men, or \50 mg/dl in women), an elevated

blood pressure (C130/85 mmHg), and an elevated fasting

glucose level (C110 mg/dl or 6.1 mmol/l) [16].

Data analyses

Data were analyzed using Stata (version 11). Continuous

variables were shown as mean ± SD if they had normal

distribution. Median and interquartile range were used to

show skewed distributed continuous variables. The cate-

gorical variables were presented as absolute and relative

(%) values. A two-tailed p value \ 0.05 was considered

significant.

Baseline characteristics of men and women were

examined using Student’s t test or Wilcoxon rank-sum test

for continuous variables and the Chi squared test or Fish-

er’s exact test for categorical variables. The prevalences of

insulin resistance, MetS, and abdominal obesity were cal-

culated for men and women separately.

Then based on WC, both men and women were divided

into normal WC and larger WC subgroups (defined as a

WC C 85 cm in men or C80 cm in women). Baseline

characteristics of two subgroups were examined in both

men and women. The prevalences of insulin resistance and

MetS were also calculated. Differences in the prevalences

of metabolic disturbances between the normal WC and the

larger WC subgroups were tested using Chi squared test or

Fisher’s exact test tests as appropriate, and the analyses

were repeated separately for men and women.

Use of body mass index and waist circumference,

respectively, to predict insulin resistance and metabolic

syndrome

In order to examine whether fat mass is associated insulin

resistance and MetS in normal-weight individuals, the

associations of anthropometric indexes with insulin resis-

tance and MetS were examined using logistic regression in

men and women separately. Both BMI and waist circum-

ference (WC) were used as anthropometric indexes of fat

mass.

The first model was unadjusted. Next this model was

adjusted for variables that are likely potential confounders

and unlikely to be in the causal pathway between fat mass

and insulin resistance/MetS. These variables included age,

family history of hypertension, family history of diabetes,

physical inactivity, smoking, alcohol use, education

attainment, and dietary habits.

In sensitivity analysis, CRP was added to the above

model. CRP is potentially in the causal pathway between

fat mass and insulin resistance. Because the distribution of

CRP was highly skewed, the natural logarithm of CRP was

used for analyses.

The above analyses were repeated separately for BMI

and WC in men and women.

Prediction of insulin resistance and MetS

by the combination of BMI and WC

Although BMI and WC were correlated, the colinearity

diagnostics indicated that BMI and WC could be used in

the same regression model. To determine whether BMI and

WC independently contribute to the prediction of insulin

resistance and MetS in normal-weight subjects, both BMI

and WC were included in the adjusted models.

The logistic regression models were adjusted for age, family

history of hypertension, family history of diabetes, physical

inactivity, smoking, alcohol use, education attainment, and

dietary habits. Finally, CRP was added to the models.

The above analyses were repeated separately for men

and women in the entire cohort.

Results

There were 2,142 study subjects (the mean age was

49.55 ± 13.44 years and 796 were men) and 308 subjects

were excluded because of missing data for serum insulin,

serum fasting glucose, serum creatinine, CRP, urinary albu-

min, urinary creatinine, BMI, or waist circumference. We also

excluded 151 subjects with diabetes. We excluded 619 obesity

subjects (BMI C 24 kg/m2). Among 1,064 normal-weight

subjects, the mean age was 50.82 ± 15.56 years and 32.99 %

(351) were men. The employment rate was higher in men and

employers offered free physical examinations. It led to a lower

participation rate for men [10].

Baseline characteristics of male and female subjects

(Table 1)

Male subjects were older and had higher WC. Of the

lifestyle factors, men had a higher percentage of smoking

and alcohol use, and a lower percentage of physical inac-

tivity (p \ 0.001). Men had higher blood pressure, higher

serum creatinine, higher serum uric acid, and higher serum

triglyceride. Women had a higher prevalence of insulin

resistance (p \ 0.001). Women also had higher preva-

lences of abdominal obesity and MetS, but the differences

were not significant.

Endocrine

123

The prevalences of abdominal obesity, insulin resistance

and MetS were 32.82, 12.62, and 10.24 % in women with

normal weight, respectively. In men with normal weight,

the respective prevalences of abdominal obesity, insulin

resistance and MetS were 31.34, 8.55, and 7.41 %. Prev-

alence of metabolic disturbances was shown in Fig. 1. Of

other components of MetS, men had higher prevalences of

elevated blood pressure and elevated triglyceride levels

(p \ 0.01). And women had a higher prevalence of low

HDL-C levels (p \ 0.001).

Baseline characteristics of male and female subjects

according to waist circumference (Table 2)

Based on WC, both men and women were divided into

normal WC and larger WC subgroups. In general, in both

men and women, subjects with larger WC had higher BMI,

higher systolic and diastolic blood pressure, higher serum

Table 1 Baseline characteristics of male and female subjects

Men (n = 351) Women

(n = 713)

p value

Demographics

Age (years) 52.83 ± 16.29 49.82 ± 15.10 0.046

Clinical characteristics

Body mass index

(kg/m2)

21.39 ± 1.96 20.92 ± 21.98 \0.001

Waist circumference

(cm)

80.26 ± 7.36 76.09 ± 7.47 \0.001

History of

hypertension (%)

70 (19.94) 58 (8.13) 0.003

History of coronary

heart disease (%)

5 (1.42) 12 (1.68) 0.75

History of stroke (%) 0 2 (0.28) 1.00

Family history of

diabetes (%)

15 (4.27) 51 (7.15) 0.07

Family history of

hypertension (%)

69 (19.66) 180 (25.25) 0.045

Family history of

coronary heart

disease (%)

12 (3.42) 25 (3.50) 0.75

Family history of

stroke (%)

9 (2.56) 11 (1.54) 0.25

Current smoker (%) 115 (32.76) 9 (1.26) \0.001

Current alcohol

use (%)

46 (13.96) 7 (0.98) \0.001

Education

attainment high

school or above (%)

175 (49.86) 269 (37.73) \0.001

Physical inactivity

(%)

180 (51.85) 420 (58.91) 0.02

Dietary habit

preferring a

vegetarian diet (%)

36 (10.26) 92 (12.90) 0.21

Systolic blood

pressure (mmHg)

127.07 ± 19.86 122.33 ± 19.17 \0.001

Diastolic blood

pressure (mmHg)

77.37 ± 10.85 74.48 ± 10.55 \0.001

Laboratory

Serum creatinine

(lmol/l)

86.84 ± 15.128 63.93 ± 10.02 \0.001

Serum uric acid

(lmol/l)

391.35 ± 85.59 301.61 ± 73.79 \0.001

Estimated

glomerular

filtration rate

(ml/min/1.73 m2)

89.81 ± 18.65 94.21 ± 18.88 \0.001

Urinary albumin-to-

creatinine ratio

(mg/g)

6.10

(4.51–9.81)

8.75

(6.28–13.35)

\0.001

Fasting glucose

(mmo/l)

4.71 ± 0.52 4.66 ± 0.45 0.12

Serum C-reactive

protein

0.69

(0.33–1.70)

0.65

(0.33–1.49)

0.20

Table 1 continued

Men (n = 351) Women

(n = 713)

p value

Serum triglyceride

(mmol/l)

1.14

(0.83–1.64)

1.02

(0.75–1.41)

\0.001

Serum low density

lipoprotein (mmol/l)

3.12 ± 0.92 3.07 ± 0.86 0.37

Serum high-density

lipoprotein (mmol/l)

1.51 ± 0.33 1.64 ± 0.33 \0.001

HOMA-IR

(lU/ml 9

mmol/ml)

1.34

(0.91–1.87)

1.47

(1.08–2.15)

0.001

Metabolic disturbances

Abdominal obesity

(%)

110 (31.34) 234 (32.82) 0.63

Elevated blood

pressure (%)

178 (50.71) 299 (41.94) 0.007

Elevated fasting

glucose level (%)

6 (1.71) 8 (1.12) 0.41

Low high-density

lipoprotein-

cholesterol (%)

5 (1.42) 93 (13.43) \0.001

Elevated triglyceride

level (%)

83 (23.65) 105 (14.73) \0.001

Insulin resistance

(%)

30 (8.55) 90 (12.76) 0.04

Metabolic syndrome

(%)

26 (7.41) 73 (10.24) 0.14

Mean ± SD or median (25th to 75th percentiles) for continuous

variables and absolute and relative (%) values for category variables

are presented. Elevated blood pressure, elevated fasting glucose level,

low high-density lipoprotein-cholesterol, elevated triglyceride level,

and metabolic syndrome diagnostic criteria were defined according to

ATP III

HOMA-IR homeostatic model assessment of insulin resistance

Endocrine

123

uric acid, higher serum fasting glucose, higher serum CRP

and higher serum triglyceride. Subjects with larger WC had

higher prevalences of insulin resistance and MetS in both

men and women. Prevalence of insulin resistance and MetS

in both the normal WC and larger WC subgroups by gender

was shown in Table 2. No differences in a family history of

hypertension and diabetes were observed between the

normal WC and larger WC subgroups in men and women.

In men, subjects with normal WC had a higher proportion

of preferring a vegetarian diet.

Fourteen (5.81 %) male subjects with normal WC and

35 (7.31 %) female subjects with normal WC had insulin

resistance. Eight (1.67 %) female subjects with normal WC

met the metabolic syndrome diagnostic criteria according

to ATP III [15]. No male subject with normal WC had

MetS.

Use of body mass index and waist circumference,

respectively, to predict insulin resistance and metabolic

syndrome (Table 3)

In the unadjusted analyses, both BMI and WC were asso-

ciated with incident insulin resistance in women and men.

After adjusting for age, family history of hypertension,

family history of diabetes, smoking status, alcohol use,

physical inactivity, education attainment, and dietary hab-

its, both BMI and WC were associated with an increased

risk for insulin resistance (Table 3). The odd ratios for BMI

(every 1-kg/m2) were 1.27 (95 % CI 1.00–1.60, p = 0.049)

and 1.36 (95 % CI 1.18–1.55, p \ 0.001) in men and

women, respectively. The respective odd ratios for WC

(every 1-cm increment) in men and women were 1.11

(95 % CI 1.04–1.18, p = 0.001) and 1.10 (95 % CI

1.06–1.14, p \ 0.001). After adjustment for CRP, the

association of BMI and insulin resistance was borderly

significant in men (OR 1.24, 95 % CI 0.98–1.57, p = 0.07)

and still statistically significant in women (OR 1.29, 95 %

CI 1.12–1.49, p \ 0.001). WC was still significantly

associated with insulin resistance in both men and women.

After adjusting for potential confounders and CRP, both

BMI and WC were associated with an increased risk for

MetS. There was a 33 % increased risk of MetS for every

1-kg/m2 increase in BMI (95 % CI 1.01–1.75, p = 0.045)

in men and a 63 % increased risk of MetS (95 % CI

1.36–1.96, p \ 0.001) in women. There was a 28 %

increased risk of MetS for every 1-cm increase in WC

(95 % CI 1.16–1.41, p \ 0.001) in men and 16 %

increased risk of MetS (95 % CI 1.11–1.22, p \ 0.001) in

women.

Prediction of insulin resistance and MetS

by the combination of BMI and WC (Table 4)

Pearson correlation coefficient between BMI and WC were

0.64 in both men and women. When BMI and WC were

included in the logistic regression models, WC was inde-

pendently associated with an increased risk for insulin

resistance in both men and women (OR 1.10, 95 % CI

1.02–1.19, p = 0.009; OR 1.06, 95 % CI 1.02–1.11,

p = 0.005, respectively). The association of WC and

insulin resistance was independent of CRP. However, in

similar models, the associations of BMI and insulin resis-

tance were borderly significant in women and not signifi-

cant in men. When CRP was added to the models, BMI was

p = 0.63

P =0.43

P =0.007

P <0.001

P <0.001

P =0.04

P =0.14

0 .1 .2 .3 .4 .5

Met

aboli

c syn

drom

e

Insu

lin re

sista

nce

A low H

DL lev

el

An ele

vate

d tri

glyce

ride

level

An ele

vate

d blo

od p

ress

ure

level

An ele

vate

d fa

sting

gluc

ose

level

Abdom

inal o

besit

y

women men

Fig. 1 Prevalence of metabolic

disturbances in men and women

Endocrine

123

Table 2 Baseline characteristics of male and female subjects according to waist circumference

Men p value Women p value

WC \ 85 cm

(n = 241)

WC C 85 cm

(n = 110)

WC \ 80 cm

(n = 479)

WC C 80 cm

(n = 234)

Demographics

Age (years) 51.56 ± 16.70 55.59 ± 15.07 0.03 46.49 ± 14.66 56.68 ± 13.62 \0.001

Clinical characteristics

Body mass index (kg/m2) 20.85 ± 2.02 22.57 ± 1.15 \0.001 20.25 ± 1.89 22.29 ± 1.34 \0.001

Waist circumference (cm) 76.65 ± 5.67 88.19 ± 3.32 \0.001 72.04 ± 5.01 84.37 ± 4.10 \0.001

History of hypertension (%) 37 (15.35) 21 (19.09) 0.40 33 (6.89) 37 (15.81) \0.001

History of coronary heart disease (%) 3 (1.24) 2 (18.18) 0.62 10 (2.09) 2 (0.85) 0.35

History of stroke (%) 0 0 – 1 (0.21) 1 (0.43) 0.55

Family history of diabetes (%) 11 (4.56) 4 (3.64) 0.78 38 (7.93) 13 (5.56) 0.28

Family history of hypertension (%) 44 (18.26) 25 (22.73) 0.33 130 (27.14) 50 (21.37) 0.10

Family history of coronary heart disease (%) 7 (2.90) 5 (4.55) 0.53 18 (3.76) 7 (2.99) 0.67

Family history of stroke (%) 8 (3.31) 1 (0.91) 0.28 7 (1.46) 4 (1.71) 0.76

Current smoker (%) 82 (34.02) 33 (30.00) 0.46 5 (1.04) 4 (1.71) 0.49

Current alcohol use (%) 31 (12.86) 15 (13.64) 0.87 2 (0.42) 5 (2.14) 0.04

Education attainment high school or above (%) 117 (48.55) 58 (52.73) 0.49 211 (44.26) 58 (24.79) \0.001

Physical inactivity (%) 129 (53.53) 51 (46.36) 0.21 298 (62.21) 122 (52.14) 0.01

Dietary habit preferring a vegetarian diet (%) 31 (12.86) 5 (4.55) 0.02 57 (11.90) 35 (14.96) 0.25

Systolic blood pressure (mmHg) 125.44 ± 20.038 130.57 ± 18.32 0.03 118.26 ± 17.47 130.27 ± 19.88 \0.001

Diastolic blood pressure (mmHg) 76.35 ± 10.90 79.58 ± 10.46 0.01 73.82 ± 10.75 77.75 ± 10 \0.001

Laboratory

Serum creatinine (lmol/l) 85.95 ± 13.47 86.81 ± 18.69 0.64 63.36 ± 9.86 65.10 ± 10.26 0.03

Serum uric acid (lmol/l) 384.87 ± 86.66 405.54 ± 8.81 0.04 291.91 ± 67.34 321.46 ± 82.14 \0.001

Estimated glomerular filtration rate (ml/min/

1.73 m2)

90.27 ± 18.66 88.81 ± 18.67 0.450 97.10 ± 18.68 90.17 ± 18.48 \0.001

Urinary albumin-to-creatinine ratio (mg/g) 6.01 (4.24–9.72) 6.36

(4.95–10.78)

0.15 8.22

(6.01–12.46)

9.86

(6.98–15.20)

\0.001

Fasting glucose (mmo/l) 4.66 ± 0.51 4.80 ± 0.52 0.02 4.61 ± 0.42 4.75 ± 0.49 \0.001

Serum C-reactive protein (mg/l) 0.60 (0. 29–1.27) 0.96

(0.54–1.78)

\0.001 0.50

(0.28–1.16)

1.102

(0.58–2.05)

\0.001

Serum triglyceride (mmol/l) 1.05

(0.80–1.638)

1.34

(1.03–2.16)

\0.001 0.93

(0.71–1.219)

1.31

(0.94–1.90)

\0.001

Serum low density lipoprotein (mmol/l) 3.06 ± 0.88 3.325 ± 1.00 0.07 2.97 ± 0.84 3.26 ± 0.87 \0.001

Serum high-density lipoprotein (mmol/l) 1.54 ± 0.34 1.44 ± 0.30 0.01 1.67 ± 0.32 1.58 ± 0.33 \0.001

HOMA-IR (lU/ml 9 mmol/ml) 1.21 (0.84–1.68) 1.96

(1.23–2.39)

\0.001 1.34

(0.97–1.79)

1.85

(1.26–2.65)

\0.001

Metabolic disturbances

Elevated blood pressure (%) 110 (45.64) 68 (61.18) 0.005 136 (28.39) 163 (69.66) \0.001

Elevated fasting glucose level (%) 5 (2.07) 1 (0.91) 0.67 4 (0.84) 4 (1.71) 0.45

Low high-density lipoprotein-cholesterol (%) 3 (1.24) 2 (1.82) 0.65 48 (10.02) 45 (19.23) 0.001

Elevated triglyceride level (%) 41 (17.01) 42 (38.18) \0.001 38 (7.93) 67 (28.63) \0.001

Insulin resistance (%) 14 (5.81) 16 (14.55) 0.007 35 (7.31) 56 (23.93) \0.001

Metabolic syndrome (%) 0 26 (23.64) \0.001 8 (1.67) 65 (27.78) \0.001

Mean ± SD or median (25th to 75th percentiles) for continuous variables and absolute and relative (%) values for category variables are

presented. Elevated blood pressure, elevated fasting glucose level, low high-density lipoprotein-cholesterol, elevated triglyceride level and

metabolic syndrome diagnostic criteria were defined according to ATP III

HOMA-IR homeostatic model assessment of insulin resistance

Endocrine

123

not independently associated with an increased risk for

insulin resistance in men and women.

When the above analyses were repeated for MetS, WC

was independently associated with an increased risk for

MetS in both men and women (OR 1.31, 95 % CI 1.18–1.44,

p \ 0.001; OR 1.14, 95 % CI 1.08–1.20, p \ 0.001,

respectively). Further adjustment for CRP had little impact

on the odds ratios. BMI was independently associated with

an increased risk for MetS in women but not men.

In both men and women, family history of hypertension,

family history of diabetes, physical inactivity, and dietary

habits were not associated with insulin resistance and

MetS.

Discussion

The results of this study suggest that 8.55 % of normal-

weight men and 12.76 % of normal-weight women had

insulin resistance. 7.41 % of normal-weight men and

10.24 % of normal-weight women met the metabolic

syndrome diagnostic criteria according to ATP III. WC was

associated with incident insulin resistance and MetS inde-

pendent of BMI in both men and women. BMI was inde-

pendently associated with incident MetS in women. The

association of BMI and incident insulin resistance in

women was borderly significant (p = 0.09). The following

discussion interprets these findings in the context of

existing literature.

Individuals with normal weight but metabolic distur-

bances usually associated with obesity were first suggested

by Ruderman et al. [2–4]. Previous studies suggest that

these individuals present with insulin resistance and pre-

dispose to type 2 diabetes, hypertriglyceridemia, and pre-

mature coronary heart disease [4, 5]. Most of these studies

involved a small number of study subjects [3]. The results

of previous studies suggested that normal-weight individ-

uals with MetS are not rare in the general population [5, 8, 9].

On the large cohort of subjects derived from NHANES III,

Park et al. showed that 4.6 % of normal-weight men and

Table 3 Use of body mass index and waist circumference, respectively, to predict insulin resistance and metabolic syndrome

Insulin resistance Metabolic syndrome

Body mass index (1 kg/m2) Waist circumference (cm) Body mass index (1 kg/m2) Waist circumference (cm)

OR (95 % CI) p value OR (95 % CI) p value OR (95 % CI) p value OR (95 % CI) p value

Model 1a (men) 1.27 (1.01–1. 60) 0.04 1.10 (1.04–1.17) 0.001 1.37 (1.05–1.78) 0.02 1.26 (1.16–1.37) \0.001

Model 1a (women) 1.38 (1.21–1.58) \0.001 1.10 (1.07–1.14) \0.001 1.71 (1.43–2.04) \0.001 1.19 (1.15–1.25) \0.001

Model 2b (men) 1.27 (1.00–1.60) 0.049 1.11 (1.04–1.18) 0.001 1.37 (1.04–1.81) 0.03 1.29 (1.18–1.42) \0.001

Model 2b (women) 1.36 (1.18–1.55) \0.001 1.10 (1.06–1.14) \0.001 1.69 (1.41–2.03) \0.001 1.17 (1.12–1.23) \0.001

Model 3c (men) 1.24 (0.98–1.57) 0.07 1.11 (1.04–1.18) 0.002 1.33 (1. 01–1.75) 0.045 1.28 (1.16–1.41) \0.001

Model 3c (women) 1.29 (1.12–1.49) \0.001 1.08 (1.04–1.12) \0.001 1.63 (1.36–1.96) \0.001 1.16 (1.11–1.22) \0.001

a Unadjustedb Adjusted for age, family history of hypertension, family history of diabetes mellitus, current smoker, current alcohol use, physical inactivity,

education attainment, dietary habit (preferring a vegetarian diet)c Adjusted for above ? C-reactive protein

Table 4 Prediction of insulin resistance and metabolic syndrome by the combination of body mass index and waist circumference

Insulin resistance Metabolic syndrome

Body mass index Waist circumference (cm) Body mass index Waist circumference (cm)

OR (95 % CI) p value OR (95 % CI) p value OR (95 % CI) p value OR (95 % CI) p value

Model 1a (men) 1.02 (0.77–1.36) 0.88 1.11 (1.03–1.19) 0.005 0.87 (0.60–1.25) 0.45 1.31 (1.18–1.44) \0.001

Model 1a (women) 1.15 (0.98–1.36) 0.09 1.08 (1.03–1.12) 0.001 1.31 (1.07–1.561) 0.01 1.14 (1.08–1.20) \0.001

Model 2b (men) 1.02 (0.77–1.35) 0.88 1.10 (1.02–1.19) 0.009 0.88 (0.61–1.27) 0.49 1.19 (1.17–1.43) \0.001

Model 2b (women) 1.13 (0.96–1.34) 0.15 1.06 (1.02–1.11) 0.005 1.29 (1.05–1.59) 0.02 1.13 (1.07–1.19) \0.001

a Adjusted for age, family history of hypertension, family history of diabetes mellitus, current smoker, current alcohol use, physical inactivity,

education attainment, dietary habit (preferring a vegetarian food); BMI and WC were included in the modelsb Adjusted for above ? C-reactive protein

Endocrine

123

6.2 % of normal-weight women met the MetS diagnostic

criteria according to ATP III [3]. Results from a commu-

nity-based, longitudinal study of 2,902 people showed that

7 % normal-weight subjects had MetS [8]. The results of

the current study suggest the prevalences of insulin resis-

tance in normal-weight men and women were 8.55 and

12.62 %, respectively. To our knowledge, this is the first

study based on the large cohort of Asian subjects.

Over the past 20 years, it has become apparent that a

number of interrelated factors are often associated with and

may contribute to the pathogenesis of insulin resistance in

both normal-weight and obese individuals. They include

central obesity, low birth weight, inactivity, and family

history [5].

In a previous study, even there were no differences in

birth weight and BMI between insulin resistance and

insulin sensitive subgroups, insulin resistance subgroup

had higher fat mass measured by dual energy X-ray

absorptiometry [7].Based on one study in non-obese,

normoglycemic subjects, although both insulin resistance

and insulin sensitive subgroups had normal BMI and there

was no difference of BMI between two subgroups, body

adipose stores were significantly increased in insulin

resistance subgroup. Higher total body, abdominal adipose

stores and increased intramyocellular lipids are related to

activation of the intramyocellular lipids pathway which is

associated with insulin resistance. Association of elevated

intramyocellular lipids with decreased insulin action was

observed in non-obese subjects [1, 17]. In the current

study, normal-weight subjects with larger WC had a

significant higher prevalence of insulin resistance and

MetS in both men and women. The results suggest that

relatively higher fat mass might be plausible explanation

for insulin resistance and metabolic disturbances in nor-

mal-weight subjects.

BMI assesses entire body mass without differentiating

different components of subcutaneous and visceral fat,

muscle, and bone [18]. WC is strongly correlated with

abdominal and thought to represent fat stored in visceral

depots [19, 20]. The results of this study suggest that WC

independently contributes to the prediction of insulin

resistance and MetS in normal-weight men and women.

The results suggest that abdominal fat is associated with

insulin resistance in normal-weight subjects and the adi-

posity exists in subjects with normal-weight.

However, the results of this study also suggest that BMI

independently contributes to the prediction of MetS in

normal-weight women. One previous study suggests that

BMI and WC independently contribute to the prediction of

nonabdominal, abdominal subcutaneous, and visceral fat.

Excess nonabdominal, abdominal subcutaneous, or visceral

fat predict the relative risk of disease, however, WC is not a

clinically significant predictor of abdominal subcutaneous

fat [21]. The findings of the current study support the

combined use of BMI and WC to predict in MetS in nor-

mal-weight women. The fact that BMI is associated with

MetS in normal-weight women independent of WC sug-

gests that not only visceral fat but also nonabdominal and

abdominal subcutaneous contribute to the prediction of

MetS.

Some factors other than fat mass, such as physical

inactivity, low birth weight, and family history may con-

tribute to the pathogenesis of insulin resistance. But in the

current study, none of family history of hypertension and

diabetes, dietary habits, and physical inactivity was asso-

ciated with insulin resistance and MetS.

Strengths of this study include the use of a relatively

large population to conduct the analyses and the combined

use of BMI and WC to predict insulin resistance and MetS.

The first limitation of our study is a cross-sectional study

which limits causal inference. Second, homeostatic model

assessment of insulin is a reliable indicator of insulin

resistance, but not the gold standard [12]. HOMA index is

not suitable for all persons [15]. The magnitude of the

correlation coefficients of HOMA index with insulin

resistance determined by euglycemic clamp test is smaller

in the subjects with a lower body mass index, a lower

HOMA-beta cell function and higher fasting glucose levels

[22]. Third, some potential confounders were not included

in our analyses. Birth weight and detailed diet diary were

not included in this study. But one previous study failed to

find the association of birth weight with insulin resistance

in normal-weight subjects [6]. And the results from another

previous study showed that no difference in diary was

found between metabolically obese subjects with normal

weight and control subjects [3]. Finally, only 32.99 % of

the participants were men. The sample is biased, but the

data analyses were repeated separately for men and

women.

Conclusion

In conclusion, this study suggests that normal-weight

individuals with insulin resistance and/or MetS are not rare

in the Chinese population. WC is significantly associated

with insulin resistance and MetS independent of BMI. BMI

is independently associated with incident MetS in women.

The current findings support using both BMI and WC in

clinical practice.

Acknowledgments This study was supported by the following

science foundation: 1. EU FP7 Program, UroSense, 2011; 2. ISN

Research Committee Grant, 2007; 3. ISN Research Committee Grant,

2004; 4. Guangdong Provincial Science and Technique Program (No.

2011B031800386), 2011.

Endocrine

123

Disclosure None.

Conflict of interest None.

References

1. S.P. Bagby, Obesity-initiated metabolic syndrome and the kid-

ney: a recipe for chronic kidney disease? J. Am. Soc. Nephrol.

15(11), 2775–2791 (2004)

2. N.B. Ruderman, S.H. Schneider, P. Berchtold, The ‘‘metaboli-

cally-obese,’’ normal-weight individual. Am. J. Clin. Nutr. 34(8),

1617–1621 (1981)

3. F. Conus, R. Rabasa-Lhoret, F. Peronnet, Characteristics of

metabolically obese normal-weight (MONW) subjects. Appl.

Physiol. Nutr. Metab. 32(1), 4–12 (2007)

4. N.B. Ruderman, P. Berchtold, S. Schneider, Obesity-associated

disorders in normal-weight individuals: some speculations. Int.

J. Obes. 6(Suppl 1), 151–157 (1982)

5. N. Ruderman, D. Chisholm, X. Pi-Sunyer, S. Schneider, The

metabolically obese, normal-weight individual revisited. Diabetes

47(5), 699–713 (1998)

6. R.V. Dvorak, W.F. DeNino, P.A. Ades, E.T. Poehlman, Pheno-

typic characteristics associated with insulin resistance in meta-

bolically obese but normal-weight young women. Diabetes

48(11), 2210–2214 (1999)

7. F. Conus, D.B. Allison, R. Rabasa-Lhoret, M. St-Onge, D.H. St-

Pierre, A. Tremblay-Lebeau, E.T. Poehlman, Metabolic and

behavioral characteristics of metabolically obese but normal-

weight women. J. Clin. Endocrinol. Metab. 89(10), 5013–5020

(2004)

8. J.B. Meigs, P.W. Wilson, C.S. Fox, R.S. Vasan, D.M. Nathan,

L.M. Sullivan, R.B. D’Agostino, Body mass index, metabolic

syndrome, and risk of type 2 diabetes or cardiovascular disease.

J. Clin. Endocrinol. Metab. 91(8), 2906–2912 (2006)

9. M.-P. St-Onge, I. Janssen, S.B. Heymsfield, Metabolic syndrome

in normal-weight Americans. New definition of the metabolically

obese, normal-weight individual. Diabetes Care 27, 2222–2228

(2004)

10. S.Y. Chen, Y.M. Chen, X.Y. Liu et al., Association of insulin

resistance with chronic kidney disease in non-diabetic subjects

with normal weight. PLoS ONE 8(9), e74058 (2013)

11. A. Molarius, J.C. Seidell, S. Sans, J. Tuomilehto, K. Kuulasmaa,

Waist and hip circumferences, and waist-hip ratio in 19 popula-

tions of the WHO MONICA Project. Int. J. Obes. Relat. Metab.

Disord. 23(2), 116–125 (1999)

12. A.J. Hanley, K. Williams, C. Gonzalez, R.B. D’Agostino Jr, L.E.

Wagenknecht, M.P. Stern, S.M. Haffner, San Antonio Heart

Study, Mexico City Diabetes Study, Insulin Resistance Athero-

sclerosis Study, Prediction of type 2 diabetes using simple

measures of insulin resistance: combined results from the San

Antonio Heart Study, the Mexico City Diabetes Study, and the

Insulin Resistance Atherosclerosis Study. Diabetes 52(2),

463–469 (2003)

13. A.S. Levey, J.P. Bosch, J.B. Lewis, T. Greene, N. Rogers, D.

Roth, A more accurate method to estimate glomerular filtration

rate from serum creatinine: a new prediction equation. Modifi-

cation of Diet in Renal Disease Study Group. Ann. Intern. Med.

130(6), 461–470 (1999)

14. Z. Bei-Fan, Cooperative Meta-Analysis Group of Working Group

on Obesity in China, Predictive values of body mass index and

waist circumference for risk factors of certain related diseases in

Chinese adults: study on optimal cut-off points of body mass

index and waist circumference in Chinese adults. Asia Pac.

J. Clin. Nutr. 11(Suppl 8), S685–S693 (2002)

15. X.Y. Xing, W.Y. Yang, Z.J. Yang, The diagnostic significance of

homeostasis model assessment of insulin resistance in Metabolic

Syndrome among subjects with different glucose tolerance

(Chinese). Chin J Diabetes 12(3), 182–186 (2004)

16. National Cholesterol Education Program (NCEP) Expert Panel on

Detection Ea, and Treatment of High Blood Cholesterol in Adults

(Adult Treatment Panel III), Third report of the National Cho-

lesterol Education Program (NCEP) Expert Panel on detection,

evaluation, and treatment of high blood cholesterol in adults

(adult treatment panel III) final report. Circulation 106(25),

3143–3421 (2002)

17. U.B. Masharani, B.A. Maddux, X. Li, G.K. Sakkas, K. Mulligan,

M. Schambelan, I.D. Goldfine, J.F. Youngren, Insulin resistance

in non-obese subjects is associated with activation of the JNK

pathway and impaired insulin signaling in skeletal muscle. PLoS

ONE 6(5), e19878 (2011)

18. J.O. Burton, L.J. Gray, Webb, M.J. Davies, K. Khunti, W. Crasto,

S.J. Carr, N.J. Brunskill, Association of anthropometric obesity

measures with chronic kidney disease risk in a non-diabetic

patient population. Nephrol. Dial. Transpl. 27(5), 1860–1866

(2012)

19. L.M. Browning, S.D. Hsieh, M. Ashwell, A systematic review of

waist-to-height ratio as a screening tool for the prediction of

cardiovascular disease and diabetes: 0.5 could be a suitable global

boundary value. Nutr. Res. Rev. 23(2), 247–269 (2010)

20. L.M. Browning, O. Mugridge, A.K. Dixon, S.W. Aitken, A.M.

Prentice, S.A. Jebb, Measuring abdominal adipose tissue: com-

parison of simpler methods with MRI. Obes. Facts 4(1), 9–15

(2011)

21. I. Janssen, S.B. Heymsfield, D.B. Allison, D.P. Kotler, R. Ross,

Body mass index and waist circumference independently con-

tribute to the prediction of nonabdominal, abdominal subcuta-

neous, and visceral fat. Am. J. Clin. Nutr. 75(4), 683–688 (2002)

22. E.S. Kang, Y.S. Yun, S.W. Park et al., Limitation of the validity

of the homeostasis model assessment as an index of insulin

resistance in Korea. Metabolism 54(2), 206–211 (2005)

Endocrine

123