The Journal of International Medical Research

2010; 38: 792 – 802 [first published online as 38(3) 30]

Adiponectin Levels in Non-obese First-

degree Relatives of Type 2 Diabetes
Patients and Non-diabetic Subjects:
a 5-Year Follow-up Study
J LIU, F WANG, Y CHA, ZP CHEN AND HY DING
Department of Endocrinology, Shanghai Fifth People’s Hospital, Fudan University,
Shanghai, China

This study was designed to investigate with age, high-density lipoprotein-

adiponectin levels and their relationship to cholesterol and IMT; these factors explained
various parameters at baseline and after 5 45% of the variation in adiponectin in the
years in non-obese first-degree relatives of FDR group. In the normal group, multiple
type 2 diabetes patients (FDR group) versus regression analyses revealed that low-
subjects without a family history of diabetes density lipoprotein-cholesterol and IMT
(normal group). Adiponectin levels at explained 25% of the variability in the
baseline were lower in the FDR group versus adiponectin concentration. In both groups,
the normal group. After 5 years, however, the correlation between
adiponectin levels had fallen significantly adiponectin and IMT just failed to reach
in both the FDR (24.3% reduction) and the statistical significance in this population
normal (35.7% reduction) groups. group. We conclude that adiponectin levels
Adiponectin levels were negatively were reduced in non-obese first-degree
correlated with waist/hip ratio, fasting relatives of patients with type 2 diabetes
plasma glucose, carotid intima-media and normal individuals over a 5-year
thickness (IMT) and insulin resistance in period. This study supports previous
the FDR group. When adjusted for relevant findings that hypoadiponectinaemia is a
risk factors, adiponectin was associated risk factor for atherosclerosis.

KEY WORDS: TYPE 2 DIABETES PATIENTS;ADIPONECTIN; ATHEROSCLEROSIS; RISK FACTORS;

FIRST-DEGREE RELATIVES

Introduction association with obesity,2,3 insulin

Adiponectin, a 30 kDa protein, is synthesized resistance, metabolic syndrome,2 type 2
4

exclusively by adipose tissue1 and has been diabetes,2,5 essential hypertension,6 coronary
reported to have anti-inflammatory, artery disease,7 dyslipidaemia8 and in first-
antidiabetic and antiatherogenic properties. degree relatives of type 2 diabetes patients.9
Decreased serum concentrations of Low baseline adiponectin concentrations
adiponectin have been described in have been shown to predict the future

792
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

development of impaired glucose regulation phenotypic characteristics associated with a

and type 2 diabetes mellitus, whereas hereditary predisposition for type 2 diabetes.
elevated baseline levels have been shown to
be protective against the subsequent Subjects and methods
development of type 2 diabetes.5,10 SUBJECTS
First-degree relatives of individuals with Between January 2003 and May 2008, non-
type 2 diabetes are at increased risk of obese first-degree relatives of type 2 diabetes
developing glucose intolerance and diabetes. patients (FDR group) were recruited from the
Several studies have been published Diabetes Treatment Centre of Shanghai Fifth
concerning circulating adiponectin People’s Hospital, Fudan University,
concentrations in first-degree relatives,9,11 – 13 Shanghai, China, and subjects without a
but the results have been inconsistent. known family history of diabetes (normal
Among these studies, Osei et al.11 reported group) were recruited from out-patient
that low adiponectin levels were an physical health clinics. Subjects with
important determinant of glucose tolerance abnormal renal, hepatic, heart, lung or
in first-degree relatives who subsequently thyroid function were excluded from the
developed type 2 diabetes. study. Members of the normal group were
Insulin resistance and β-cell dysfunction individually matched for age, BMI, WHR,
are prerequisites for the development of type blood glucose and fasting lipids with
2 diabetes. The relative role of each remains members of the FDR group.
controversial, since studies of at-risk first- The Human Rights Committee of the
degree relatives of patients with type 2 Shanghai Fifth People’s Hospital of Fudan
diabetes have yielded conflicting results. University approved the study protocol and all
Some studies have demonstrated insulin participants gave written informed consent.
resistance in first-degree relatives,9,12
whereas others have reported that impaired ASSESSMENT
insulin secretion is the first identifiable Biochemical and anthropometric
phenotype.14 – 17 To date, no consensus has measurements, together with assessments of
been reached. endothelial function and atherosclerosis,
To gain further insight into the were made at baseline and after a 5-year
pathogenesis of type 2 diabetes, the present follow-up period.
study explored the relationship between a
pre-diabetic state and adiponectin. Biochemical measurements
Adiponectin, carotid intima-media thickness Biochemical measurements included fasting
(IMT), lipid profiles, insulin resistance and plasma glucose (FPG), 2-h post-load plasma
insulin secretion were measured in non- glucose, triglycerides, high-density
obese first-degree relatives of type 2 diabetes lipoprotein (HDL)-cholesterol, low-density
patients at baseline and after 5 years, and lipoprotein (LDL)-cholesterol, adiponectin, β-
compared with results seen in normal cell function and insulin resistance.
subjects without a known family history of Blood samples were drawn from the
diabetes. The subjects were individually antecubital vein after overnight fasting for at
matched for age, body mass index (BMI), least 8 h. A 75-g oral glucose tolerance test
waist/hip ratio (WHR), blood glucose and (OGTT) was then administered, and a second
fasting lipids in order to differentiate early blood sample for glucose was drawn after 120

793
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

min. Blood samples were stored at –80 °C until Carotid IMT

analysed. Individuals were excluded from the Carotid IMT was measured using
study if they had physician-diagnosed ultrasonography in the supine position.
diabetes and were currently receiving Well-trained ultrasonographers, blinded to
medication for diabetes or if they had a FPG the subjects’ characteristics, produced high-
level ≥ 6.1 mmol/l or a 2-h post-load glucose resolution B-mode ultrasound images with a
level ≥ 7.8 mmol/l at baseline. 10 MHz linear array transducer (LOGIQ9;
Plasma glucose was measured by the General Electric, Milwaukee, WI, USA).
glucose oxidase method using a Hitachi Longitudinal scanning was performed from
7170S analyser (Hitachi High-Technologies the common carotid artery (CCA) to the
Corp., Tokyo, Japan). The fasting insulin bulbus. The IMT was measured on the B-
concentration was determined by mode scan using electronic calipers to 10
immunochemiluminescent assay (Roche mm proximal to the CCA–bulbus junction.
Diagnostics, Mannheim, Germany). The mean IMT over six segments of both
Adiponectin was measured using a carotid arteries was calculated. The
commercially available enzyme-linked coefficient of variation for measurement of
immunosorbent assay kit (B-Bridge IMT was 3.3 ± 1.60%.
International, Mountain View, CA, USA).
Both the intra- and interassay coefficients of Brachial artery vascular reactivity
variation were < 5.0%. Endothelial-dependent vasodilation (EDVD)
Lipid profiles (triglycerides, total of the right brachial artery was measured
cholesterol, LDL-cholesterol and HDL- after carotid ultrasound according to the
cholesterol) were determined using standard guidelines of the International Brachial
laboratory methods. Homeostasis model Artery Reactivity Task Force.19 The lumen
assessment (HOMA) was utilized to quantify diameter of the artery was defined as the
β-cell function (HOMA-β) and insulin distance between the leading edge of the
resistance (HOMA-IR) using the equations near wall–lumen interface echo to the
HOMA-β = (20 × insulin)/(glucose – 3.5) and leading edge of the far wall–lumen interface
HOMA-IR = (glucose × insulin)/22.5. echo. All scans were electrocardiogram
triggered and made coincident with the R
Anthropometric measurements wave (end of diastole). EDVD was elicited by
Height and weight were measured to the induced hyperaemia following inflation of a
nearest 0.5 cm and 0.1 kg, respectively, and pneumatic tourniquet placed around the
BMI was calculated as weight (kg)/height forearm, distal to the scanned part of the
(m)2. Waist and hip circumferences were artery, up to a pressure of 250 mmHg for 5
determined using the mean of two min followed by sudden deflation. This
measurements; waist circumference was manoeuvre is recognized to cause shear
measured at the smallest circumference of stress on the endothelial cells which, in turn,
the waist and hip circumference was releases nitric oxide, producing vasodilation,
measured at the widest level of the buttocks. and is, therefore, a test for endothelial
The WHR was then calculated. Body fat function. Images of the brachial artery were
percentage was calculated from the BMI, age recorded at 10, 30, 60, 90 and 120 s after
and sex using a Caucasian prediction deflation. The diameter of the brachial
formula.18 artery was determined as the mean of three

794
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

measurements. EDVD was expressed as the EDVD at baseline (Table 1). Comparisons
percentage increase in the maximal between the groups showed that HOMA-IR
diameter after cuff deflation within 120 s and HOMA-β levels were significantly higher
compared to the baseline diameter. in the FDR group, and the FDR group also had
significantly lower adiponectin levels (P < 0.05
STATISTICAL ANALYSIS for all comparisons). The lower adiponectin
Data were reported as the mean ± SD or the levels in the FDR group remained when
median and 25th and 75th percentiles. All adjusting for the small differences in HDL-
analyses were performed using the SPSS® cholesterol (data not shown).
statistical package, version 10.0 (SPSS Inc., As shown in Table 2, the WHR increased
Chicago, IL, USA) for Windows®. Differences significantly over the 5-year follow-up period
between the FDR group and the normal in both the FDR (P < 0.001) and the normal
group were compared using Student’s t-test group (P < 0.05), without a change in BMI in
or the χ2-test, as appropriate. Within-group either group. In parallel with this increase in
comparisons of baseline and follow-up data central adiposity, the body fat percentage,
were made using a paired t-test. Values of total cholesterol and triglycerides also
HOMA-IR, HOMA-β and triglycerides were significantly increased in both groups (P < 0.05
logarithmically transformed due to skewed for all comparisons except total cholesterol in
distributions. Spearman correlation was used the FDR group which was P < 0.001; Table 2).
to analyse associations between adiponectin Systolic and diastolic blood pressure levels and
and other variables. Multiple linear EDVD did not change significantly in either
regression was used to identify independent group during follow-up (Table 2). HOMA-IR
determinants for adiponectin. A P-value of did not change significantly over time in the
≤ 0.05 was considered to be statistically FDR group, whereas β-cell function, as
significant. determined by the HOMA-β value, decreased
by 39.5% over the same period (P < 0.05). The
Results adiponectin level in the FDR group was 24.3%
A total of 53 FDR subjects and 37 normal lower at the 5-year assessment compared with
subjects were included in the study. Of these, baseline (P < 0.001). The FPG increased
29 FDR subjects and 20 normal subjects were significantly in the FDR group at 5 years (P <
followed up 5 years later. The baseline and 0.001; Table 2), without a change in post-load
follow-up clinical and biological glucose. Meanwhile, IMT increased by 5.1%
characteristics of the study participants are over the same period in the FDR group (P <
given in Tables 1 and 2. 0.05) (Table 2). One patient in the FDR group
There were no statistically significant had progressed to diabetes during the 5-year
differences between the FDR and normal follow-up period. In the normal group,
groups in terms of sex distribution. All adiponectin decreased by 35.7% over the 5-
participants had normal glucose tolerance, year follow-up period (P < 0.001) (Table 2).
but the 2-h post-load glucose level during the Insulin resistance, as measured as HOMA-IR,
OGTT was significantly higher in the FDR was 72.2% higher at the 5-year assessment in
group than in the normal group at baseline (P the normal group (P < 0.05). In line with the
< 0.05; Table 1). Both groups had similar increase in insulin resistance, HOMA-β also
anthropometric measures of adiposity and increased significantly during the same time
similar levels of FPG, lipids, carotid IMT and in the normal group (P < 0.05).

795
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

TABLE 1:
Baseline demographic, clinical and biochemical characteristics of non-obese first-degree
relatives of type 2 diabetes patients (FDR group) and subjects without a known family
history of diabetes (normal group)
Demographic or characteristic Normal group FDR group
No of subjects 37 53
Age (years) 49.19 ± 8.76 48.55 ± 7.30
Sex
Male 12 23
Female 25 30
Body mass index (kg/m2) 24.49 ± 3.47 24.72 ± 3.15
Waist/hip ratio 0.86 ± 0.05 0.86 ± 0.07
Body fat (%) 32.02 ± 7.31 29.60 ± 6.66
Systolic blood pressure (mmHg) 122.97 ± 13.87 121.60 ± 15.71
Diastolic blood pressure (mmHg) 77.30 ± 7.51 77.51 ± 10.81
Triglycerides (mmol/l) 1.18 (0.86 – 1.80) 1.13 (0.81 – 1.84)
Total cholesterol (mmol/l) 4.47 ± 0.97 4.26 ± 0.78
LDL-cholesterol (mmol/l) 2.64 ± 0.74 2.59 ± 0.69
HDL-cholesterol (mmol/l) 1.39 ± 0.38 1.23 ± 0.29a
Fasting plasma glucose (mmol/l) 4.48 ± 0.48 4.59 ± 0.71
2-h post-load glucose (mmol/l) 4.76 ± 0.32 5.32 ± 1.61a
Adiponectin (mg/l) 14.43 ± 7.91 10.06 ± 5.79a
Carotid IMT (mm) 0.81 ± 0.09 0.77 ± 0.11
EDVD (%) 9.78 ± 7.23 10.37 ± 6.04
HOMA-IR 0.76 (0.56 – 1.29) 1.24 (0.67 – 1.81)a
HOMA-β 75.29 (57.55 – 127.25) 124.55 (62.62 – 201.19)a
Data shown are the number of subjects, the mean ± SD or the median (25th – 75th percentile).
a
P < 0.05 compared with the normal group.
LDL, low-density lipoprotein; HDL, high-density lipoprotein; IMT, intima-media thickness; EDVD, endothelial-
dependent vasodilation; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β,
homeostasis model assessment of β-cell function.

No correlation was found between the correlations between adiponectin and FPG (P
change in adiponectin over time and the < 0.05) and HOMA-IR (P < 0.01) (Table 3).
evolution of insulin resistance, β-cell Stepwise multiple regression analyses
function and IMT (data not shown). The were performed on the results from the FDR
change in adiponectin was, however, group, with adiponectin as a dependent
inversely correlated with the change in HDL- variable and age, WHR, BMI, LDL-
cholesterol in the FDR group. cholesterol, HDL-cholesterol, FPG, 2 h post-
At follow-up, there were significant load glucose, IMT, EDVD, HOMA-IR and
negative correlations between adiponectin HOMA-β as independent variables. Age,
and WHR, FPG, IMT (all P < 0.05) and HDL-cholesterol and IMT were shown to be
HOMA-IR (P < 0.01) in the FDR group (Table explanatory variables of adiponectin (Table
3). In the normal group, there were 4), with a total explanatory power of 45%;
significant positive correlations between although the correlation between
adiponectin and HDL- and LDL-cholesterol adiponectin and IMT was not statistically
(P < 0.01), and significant negative significant. In the normal group, similar

796
 TABLE 2:
Demographic, clinical and biochemical characteristics of non-obese first-degree relatives of type 2 diabetes patients (FDR
group) and subjects without a known family history of diabetes (normal group) for whom both baseline and 5-year follow-up
data were available
Normal group (n = 20) FDR group (n = 29)
Demographic or characteristic Baseline Follow up Baseline Follow up
Sex
Male 7 7 13 13
Female 13 13 16 16
Age (years) 49.05 ± 8.56 53.55 ± 9.02 48.41 ± 6.57 53.07 ± 6.19
Body mass index (kg/m2) 23.50 ± 2.68 23.68 ± 2.62 24.27 ± 2.87 24.02 ± 2.87
Waist/hip ratio 0.86 ± 0.05 0.91 ± 0.07a 0.86 ± 0.07 0.90 ± 0.06b
a
Body fat (%) 30.30 ± 5.81 31.55 ± 5.61 28.21 ± 6.76 28.99 ± 6.11a
Systolic blood pressure (mmHg) 121.50 ± 14.70 126.30 ± 14.56 123.79 ± 17.35 125.93 ± 15.94
Diastolic blood pressure (mmHg) 77.75 ± 9.66 80.75 ± 9.37 77.93 ± 9.78 79.83 ± 10.82

797
Triglycerides (mmol/l) 1.22 (0.90 – 1.74) 1.40 (0.98 – 1.97)a 1.15 (0.89 – 2.10) 1.27 (0.91 – 2.90)a
Total cholesterol (mmol/l) 4.19 ± 0.89 4.75 ± 0.73a 4.30 ± 0.73 4.89 ± 0.67b
LDL-cholesterol (mmol/l) 2.38 ± 0.60 2.41 ± 0.64 2.65 ± 0.60 2.75 ± 0.55c
HDL-cholesterol (mmol/l) 1.32 ± 0.32 1.25 ± 0.35 1.19 ± 0.31 1.09 ± 0.31a
Fasting plasma glucose (mmol/l) 4.50 ± 0.60 4.68 ± 2.85 4.57 ± 0.60 5.20 ± 0.77b,c
J Liu, F Wang, Y Cha et al.

2-h post-load glucose (mmol/l) 4.73 ± 0.31 5.94 ± 0.87a 5.16 ± 1.43c 5.69 ± 0.93
Adiponectin (mg/l) 13.26 ± 7.30 8.53 ± 4.41 b 9.92 ± 5.07c 7.51 ± 3.72b
Adiponectin levels in type 2 diabetes

Carotid IMT (mm) 0.80 ± 0.10 0.85 ± 0.15 0.79 ± 0.10 0.83 ± 0.11a
EDVD (%) 9.96 ± 9.37 8.88 ± 5.48 9.05 ± 4.93 10.37 ± 8.11
HOMA-IR 0.79 (0.43 – 1.34) 1.36 (0.90 – 2.14)a 1.01 (0.68 – 1.63) 1.12 (0.74 – 1.87)
HOMA-β 75.29 (59.82 – 143.42) 157.33 (97.33 – 190.00)a 126.67 (59.55 – 168.28)c 76.59 (49.68 – 96.19)a,c
Data shown are the number of subjects, the mean ± SD or the median (25th – 75th percentile).
aP < 0.05, bP < 0.001 compared with baseline; cP < 0.05 compared with the normal group.

LDL, low-density lipoprotein; HDL, high-density lipoprotein; IMT, intima-media thickness; EDVD, endothelial-dependent vasodilation; HOMA-IR,
homeostasis model assessment of insulin resistance; HOMA-β, homeostasis model assessment of β-cell function.
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

TABLE 3:
Spearman correlation coefficients (r) for plasma adiponectin compared with other
variables in non-obese first-degree relatives of type 2 diabetes patients (FDR group) and
in subjects without a known family history of diabetes (normal group) after 5 years of
follow-up
FDR group Normal group
(n = 20) (n = 29)
Statistical Statistical
Variable r significance r significance
Age 0.286 NS –0.048 NS
Systolic blood pressure –0.118 NS –0.134 NS
Diastolic blood pressure –0.018 NS –0.255 NS
Body mass index –0.322 NS –0.352 NS
Waist/hip ratio –0.397 P = 0.032 –0.332 NS
Body fat 0.253 NS 0.125 NS
Triglycerides –0.156 NS –0.414 NS
Total cholesterol –0.051 NS 0.385 NS
LDL-cholesterol –0.208 NS 0.648 P = 0.002
HDL-cholesterol 0.353 NS 0.581 P = 0.007
Fasting plasma glucose –0.373 P = 0.046 –0.536 P = 0.015
2-h post-load glucose –0.037 NS 0.061 NS
Carotid IMT –0.372 P = 0.037 –0.256 NS
EDVD 0.278 NS –0.069 NS
HOMA-IR –0.538 P = 0.003 –0.602 P = 0.005
HOMA-β –0.292 NS –0.014 NS
LDL, low-density lipoprotein; HDL, high-density lipoprotein; IMT, intima-media thickness; EDVD, endothelial-
dependent vasodilation; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β,
homeostasis model assessment of β-cell function; NS, not statistically significant (P > 0.05).

analysis revealed that LDL-cholesterol and adiponectin levels in relatives have been
IMT explained 25% of the variability in the reported by some,9,11 but not all,12,13
adiponectin level; again, however, the investigators. The FDR group in the present
correlation between adioponectin and IMT study demonstrated a lower adiponectin level
was not statistically significant. at baseline compared with the normal group;
although blood pressure, WHR and BMI were
Discussion all similar between the two groups. These
A decrease in adiponectin levels was seen data were consistent with previous
both in the FDR and normal groups over 5 studies.9,11,12 The reduced level of adipose
years, and adiponectin levels were shown to tissue adiponectin mRNA found in first-
correlate negatively with carotid IMT at degree relatives12 suggests that altered
follow up, although this did not reach adiponectin gene expression plays a
statistical significance in this population. pathogenic role in the development of
A current concept is that decreased diabetes. Another explanation of the role of
adiponectin levels are a key predictor of adiponectin production and secretion in the
impaired glucose regulation and type 2 FDR group is given in the genetic studies
diabetes mellitus.10 Declining fasting performed by Stumvoll et al.,20 who showed

798
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

TABLE 4:
Multiple linear regression analyses of plasma adiponectin with various parameters as
independent variablesa in non-obese first-degree relatives of type 2 diabetes patients (FDR
group) and in subjects without a known family history of diabetes (normal group) after 5
years of follow-up
FDR group Normal group
Regression Statistical Regression Statistical
Variable coefficient significance coefficient significance
Age 0.429 P = 0.015 –0.188 NS
Waist/hip ratio 0.000 NS –0.207 NS
Body mass index 0.024 NS 0.001 NS
LDL-cholesterol –0.201 NS 0.502 P = 0.047
HDL-cholesterol 0.476 P = 0.008 0.146 NS
Fasting plasma glucose –0.170 NS –0.140 NS
2-h post-load glucose 0.056 NS 0.326 NS
Carotid IMT –0.319 NS (P = 0.059) –0.449 NS (P = 0.052)
EDVD 0.337 NS –0.122 NS
HOMA-IR –0.119 NS –0.246 NS
HOMA-β 0.066 NS –0.146 NS
aStepwise multiple regression analyses were performed on the results from the FDR and normal groups, with
adiponectin as a dependent variable and age, waist/hip ratio, body mass index, LDL-cholesterol, HDL-
cholesterol, fasting plasma glucose, 2 h post-load glucose, IMT, EDVD, HOMA-IR and HOMA-β as
independent variables.
r 2 = 0.452 and C = –12.589 for the FDR group, and r 2 = 0.252 and C = –0.464 for the normal group, where
C is the covariance explained by the variable(s).
LDL, low-density lipoprotein; HDL, high-density lipoprotein; IMT, intima-media thickness; EDVD, endothelial-
dependent vasodilation; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β,
homeostasis model assessment of β-cell function; NS, not statistically significant (P > 0.05).

that a silent, highly prevalent T/G statistically significant (P = 0.052). Taken

polymorphism in exon 2 of the adiponectin together, these findings could implicate
gene probably contributes to the adiponectin in the risk of progression to
development of insulin resistance. atherosclerosis in populations with pre-
In the present 5-year prospective study, diabetes and non-diabetes. Accordingly, it
adiponectin levels fell in both groups during has been found that the increased IMT in a
follow-up (by 24.3% in the FDR group and by cohort of middle-aged diabetes patients
35.7% in the normal group). The reason for could be explained by the decreasing
this is unclear, but it may be a consequence adiponectin level.21 It is not surprising that,
of the increased age of the subjects, reflecting in the present study, hypoadiponectinaemia
a global epidemic. There was a correlation seemed to be a significant risk factor for later
between adiponectin and age, HDL- development of atherosclerosis, in
cholesterol and carotid IMT in the FDR agreement with previous studies.21,22
group, although for IMT this was not There were certain advantages to the
statistically significant (P = 0.059). LDL- present study, including the long 5-year
cholesterol and IMT were also associated follow-up period, the use of a matched normal
with adiponectin levels in normal subjects, control group and the fact that carotid IMT
although again for IMT this was not and EDVD can be measured non-invasively

799
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

and are widely used diagnostic tools both in decreased insulin secretion over 5 years,
clinical practice and in epidemiological whereas insulin resistance did not change.
studies. The results correlated well with the This finding is similar to that of the
presence of coronary heart disease and future prospective study by Cnop et al.,16 which
development of myocardial infarction, and showed that the development of reduced
are reproducible.23,24 The present study was, glucose tolerance is possibly associated with
however, somewhat limited by the small β-cell dysfunction in non-obese first-degree
sample size. relatives. The mechanisms underlying the
First-degree relatives of patients with type progressive decline in β-cell function in these
2 diabetes are known to be at increased risk individuals are not fully understood. It may
of developing diabetes.25 It is a much be related to a genetic predisposition
debated question as to whether impaired β- compounded by environmental exposure.
cell function is an early change or whether it Adiponectin has previously been reported
is secondary to insulin resistance in first- to be associated with insulin resistance in
degree relatives.12,13 Lihn et al.12 and Pellmé both epidemiological2,26 and experimental27
et al.9 reported that first-degree relatives with studies. The major findings in these studies
normal glucose tolerance had low insulin were that plasma adiponectin was
sensitivity, whereas there was no difference negatively correlated with BMI and
in insulin secretion. In contrast, Pimenta et WHR,2,4,26 and positively correlated with
al.14 and van Haeften et al.15 reported that β- insulin sensitivity.2,4 However, the only
cell dysfunction was the major determinant finding from the FDR group in the present
of the progression of diabetes in first-degree study was that plasma adiponectin was
relatives. Few longitudinal studies in these positively correlated with HDL-cholesterol
high-risk groups have been performed, after adjusting for relevant risk factors. This
however. Two follow-up studies have may be related to the relatively normal
assessed insulin resistance and β-cell range of BMI (18.1 – 29.9 kg/m2) that the
function in first-degree relatives using participants had at the beginning of the
different designs.16,17 The most recent 7-year study and also the relatively small number
follow-up study showed loss of β-cell of study participants.
function, and a strong relationship between In conclusion, in the present study,
the decline in glucose tolerance and β-cell adiponectin levels were shown to be
dysfunction in the first-degree relative significantly reduced in non-obese first-
group.16 Osei et al.17 showed that first-degree degree relatives of patients with type 2
relatives of African-American patients with diabetes and in normal individuals over a 5-
type 2 diabetes, who progressed to either year period. This study supports previous
impaired glucose tolerance or type 2 findings that hypoadiponectinaemia is a risk
diabetes, had 30% lower β-cell function. In factor for atherosclerosis. Furthermore, a
general, most studies have suggested that a decrease in β-cell function was also seen in
decline in β-cell function is a critical non-obese first-degree relatives at the 5-year
determinant of deteriorating glucose follow-up.
tolerance in first-degree relatives. In the
present study, the first-degree relatives were Conflicts of interest
characterized by increased insulin resistance The authors had no conflicts of interest to
at baseline. These non-obese relatives had declare in relation to this article.

800
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

• Received for publication 28 November 2009 • Accepted subject to revision 10 December 2009
• Revised accepted 22 March 2010
Copyright © 2010 Field House Publishing LLP

References adiposity: a study of first-degree relatives of

1 Scherer PE, Williams S, Fogliano M, et al: A type 2 diabetic subjects. Diabetes Care 2004; 27:
novel serum protein similar to C1q, produced 2033 – 2040.
exclusively in adipocytes. J Biol Chem 1995; 270: 14 Pimenta W, Korytkowski M, Mitrakou A, et al:
26746 – 26749. Pancreatic beta-cell dysfunction as the primary
2 Weyer C, Funahashi T, Tanaka S, et al: genetic lesion in NIDDM. Evidence from studies
Hypoadiponectinemia in obesity and type 2 in normal glucose-tolerant individuals with a
diabetes: close association with insulin first-degree NIDDM relative. JAMA 1995; 273:
resistance and hyperinsulinemia. J Clin 1855 – 1861.
Endocrinol Metab 2001; 86: 1930 – 1935. 15 van Haeften TW, Dubbeldam S, Zonderland
3 Lee S, Bacha F, Gungor N, et al: Racial ML, et al: Insulin secretion in normal glucose-
differences in adiponectin in youth: tolerant relatives of type 2 diabetic subjects.
relationship to visceral fat and insulin Assessments using hyperglycemic glucose
sensitivity. Diabetes Care 2006; 29: 51 – 56. clamps and oral glucose tolerance tests.
4 Arita Y, Kihara S, Ouchi N, et al: Paradoxical Diabetes Care 1998; 21: 278 – 282.
decrease of an adipose-specific protein, 16 Cnop M, Vidal J, Hull RL, et al: Progressive loss
adiponectin, in obesity. Biochem Biophys Res of beta-cell function leads to worsening glucose
Commun 1999; 257: 79 – 83. tolerance in first-degree relatives of subjects
5 Snehalatha C, Mukesh B, Simon M, et al: with type 2 diabetes. Diabetes Care 2007; 30:
Plasma adiponectin is an independent 677 – 682.
predictor of type 2 diabetes in Asian Indians. 17 Osei K, Rhinesmith S, Gaillard T, et al: Impaired
Diabetes Care 2003; 26: 3226 – 3229. insulin sensitivity, insulin secretion, and
6 Adamczak M, Wiecek A, Funahashi T, et al: glucose effectiveness predict future
Decreased plasma adiponectin concentration development of impaired glucose tolerance and
in patients with essential hypertension. Am J type 2 diabetes in pre-diabetic African
Hypertens 2003; 16: 72 – 75. Americans: implications for primary diabetes
7 Zoccali C, Mallamaci F, Tripepi G, et al: prevention. Diabetes Care 2004; 27: 1439 –
Adiponectin, metabolic risk factors, and 1446.
cardiovascular events among patients with 18 Gurrici S, Hartriyanti Y, Hautvast JG, et al:
end-stage renal disease. J Am Soc Nephrol 2002; Differences in the relationship between body fat
13: 134 – 141. and body mass index between two different
8 Matsubara M, Maruoka S, Katayose S: Indonesian ethnic groups: the effect of body
Decreased plasma adiponectin concentrations build. Eur J Clin Nutr 1999; 53: 468 – 472.
in women with dyslipidemia. J Clin Endocrinol 19 Sorensen KE, Celermajer DS, Spiegelhalter DJ, et
Metab 2002; 87: 2764 – 2769. al: Non-invasive measurement of human
9 Pellmé F, Smith U, Funahashi T, et al: endothelium dependent arterial responses:
Circulating adiponectin levels are reduced in accuracy and reproducibility. Br Heart J 1995;
nonobese but insulin-resistant first-degree 74: 247 – 253.
relatives of type 2 diabetic patients. Diabetes 20 Stumvoll M, Tschritter O, Fritsche A, et al:
2003; 52: 1182 – 1186. Association of the T-G polymorphism in
10 Jalovaara K, Santaniemi M, Timonen M, et al: adiponectin (exon 2) with obesity and insulin
Low serum adiponectin level as a predictor of sensitivity: interaction with family history of
impaired glucose regulation and type 2 type 2 diabetes. Diabetes 2002; 51: 37 – 41.
diabetes mellitus in a middle-aged Finnish 21 Dullaart RP, de Vries R, van Tol A, et al: Lower
population. Metabolism 2008; 57: 1130 – 1134. plasma adiponectin is a marker of increased
11 Osei K, Gaillard T, Schuster D: Plasma intima-media thickness associated with type 2
adiponectin levels in high risk African- diabetes mellitus and with male gender. Eur J
Americans with normal glucose tolerance, Endocrinol 2007; 156: 387 – 394.
impaired glucose tolerance, and type 2 22 Pilz S, Horejsi R, Möller R, et al: Early
diabetes. Obes Res 2005; 13: 179 – 185. atherosclerosis in obese juveniles is associated
12 Lihn AS, Østergard T, Nyholm B, et al: with low serum levels of adiponectin. J Clin
Adiponectin expression in adipose tissue is Endocrinol Metab 2005; 90: 4792 – 4796.
reduced in first-degree relatives of type 2 23 O’Leary DH, Polak JF, Kronmal RA, et al for the
diabetic patients. Am J Physiol Endocrinol Metab Cardiovascular Health Study Collaborative
2003; 284: E443 – E448. Research Group: Carotid-artery intima and
13 Kriketos AD, Greenfield JR, Peake PW, et al: media thickness as a risk factor for myocardial
Inflammation, insulin resistance, and infarction and stroke in older adults. N Engl J

801
 J Liu, F Wang, Y Cha et al.
Adiponectin levels in type 2 diabetes

Med 1999; 340: 14 – 22. Diabetes 2000; 49: 2135 – 2141.

24 Campuzano R, Moya JL, García-Lledó A, et al: 26 Lindsay RS, Funahashi T, Hanson RL, et al:
Endothelial dysfunction, intima-media Adiponectin and development of type 2
thickness and coronary reserve in relation to diabetes in the Pima Indian population. Lancet
risk factors and Framingham score in patients 2002; 360: 57 – 58.
without clinical atherosclerosis. J Hypertens 27 Hotta K, Funahashi T, Bodkin NL, et al:
2006; 24: 1581 – 1588. Circulating concentrations of the adipocyte
25 Nielsen MF, Nyholm B, Caumo A, et al: protein adiponectin are decreased in parallel
Prandial glucose effectiveness and fasting with reduced insulin sensitivity during the
gluconeogenesis in insulin-resistant first-degree progression to type 2 diabetes in rhesus
relatives of patients with type 2 diabetes. monkeys. Diabetes 2001; 50: 1126 – 1133.

Author’s address for correspondence

Dr Heyuan Ding
Department of Endocrinology, Shanghai Fifth Hospital, Fudan University, 801 Heqing Rd,
Minhang District, Shanghai 200240, China.
E-mail: [email protected]

802