Circulating adiponectin levels associate with inflammatory markers ...

International Journal of Obesity (2008) 32, 772–779 & 2008 Nature Publishing Group All rights reserved 0307-0565/08 $30.00 www.nature.com/ijo

ORIGINAL ARTICLE Circulating adiponectin levels associate with inflammatory markers, insulin resistance and metabolic syndrome independent of obesity J Hung1,2, BM McQuillan1,2, PL Thompson1,2 and JP Beilby3,4 1 Sir Charles Gairdner Hospital Campus of the Heart Research Institute of Western Australia, Perth, Western Australia, Australia; 2School of Medicine and Pharmacology, University of Western Australia, Perth, Western Australia, Australia; 3 Clinical Biochemistry, PathWest Laboratory Medicine, Perth, Western Australia, Australia and 4School of Surgery and Pathology, University of Western Australia, Perth, Western Australia, Australia

Background: Adiponectin is an abundantly expressed adipocyte-specific protein, whose level is decreased in obesity, and which appears to be a key participant in developing inflammation, insulin resistance and metabolic syndrome (MetS). We examined whether the relationship between adiponectin and inflammatory markers, insulin resistance and MetS was independent of obesity. Methods and results: The study was performed in 1094 men and women, aged 27–77 years, from a representative community population. We measured serum inflammatory markers, homoeostasis model assessment of insulin resistance (HOMA-IR) and prevalent MetS using National Cholesterol Education Program ATPIII criteria. Sex- and age-adjusted plasma adiponectin concentration was inversely correlated with body mass index (BMI), waist–hip ratio, diastolic blood pressure, triglycerides, glucose and fasting insulin, and positively correlated with HDL cholesterol (all Po0.005). Log plasma adiponectin was a significant negative correlate of the levels of C-reactive protein, interleukin-6, interleukin-18, fibrinogen and white cell count independent of level of obesity. Log plasma adiponectin was also an inverse associate of log HOMA-IR (Po0.001) independent of obesity. Subjects in the top compared to bottom sex-specific plasma adiponectin quartile had a multivariate-adjusted odds ratio (OR) of 0.21 (95% CI, 0.11–0.42; Po0.001) for prevalent MetS, and the association was independent of age, sex, BMI, log insulin and log intereukin-18 levels. Conclusion: Our findings suggest that higher circulating adiponectin levels may mitigate against adipose-related inflammation, insulin resistance and MetS as much in lean as obese persons. At any rate circulating adiponectin level is a strong risk marker for MetS, which is independent of measures of adiposity, insulin resistance and inflammatory markers. International Journal of Obesity (2008) 32, 772–779; doi:10.1038/sj.ijo.0803793; published online 5 February 2008 Keywords: adiponectin; metabolic syndrome; insulin resistance; inflammation

The metabolic syndrome (MetS) is a clustering of obesityrelated disorders in an individual that includes visceral adiposity, insulin resistance, dyslipidemia and hypertension, and is a strong determinant of type 2 diabetes and cardiovascular disease.1 The global obesity pandemic is having a major impact on the incidence of MetS and cardiovascular disease.2 It is also recognized that obesity

Correspondence: Professor J Hung, School of Medicine and Pharmacology, M503 4th Floor, G Block, Sir Charles Gairdner Hospital, Hospital Avenue, Nedlands, WA 6009, Australia. E-mail: [email protected] Received 8 August 2007; revised 8 November 2007; accepted 15 November 2007; published online 5 February 2008

has a strong inflammatory component, which may contribute to developing MetS and vascular disease.3 A fundamental change has been the recognition of adipose tissue as an active endocrine organ, which produces a host of bioactive molecules, known as adipokines, which may mediate the systemic effects of obesity on health. Some examples of these adipokines include tumour necrosis factora (TNF-a), interleukin-6 (IL-6), interleukin-18 (IL-18), leptin and adiponectin.3,4 Adiponectin, also called AdipoQ and Acrp30, is a cytokine that is exclusively and abundantly produced by adipocytes.5,6 Unlike most other adipocytederived hormones, circulating adiponectin levels are decreased in obesity.6,7 Adiponectin increases fatty acid oxidation and directly sensitizes the body to insulin.8,9 Adiponectin also reverses insulin resistance in animal models

Adiponectin, inflammation and metabolic risk J Hung et al

773 of obesity and diabetes.9,10 In humans, hypoadiponectinemia has been associated with insulin resistance,11–13 and longitudinally with an increased risk of type 2 diabetes.14,15 Adiponectin also appears to have novel anti-inflammatory actions. In vitro, adiponectin downregulates inflammatory responses of endothelial cells through nuclear factor-kb signalling pathways,16,17 inhibits macrophage functions18 and also downregulates cytokine secretion from adipocytes.19 These studies support the key participation of adiponectin in the mechanism of obesity-related inflammation, and metabolic and vascular disease, and hence adiponectin may be a novel therapeutic target.4,20 However, since obesity per se is associated with hypoadiponectinemia, it is important to determine whether adiponectin associates with inflammation, insulin resistance and metabolic risk independently of the effects of obesity. Furthermore, populationbased studies that have examined the association of plasma adiponectin with prevalent MetS have often not or only partially adjusted for these confounding associations.21–24 We have previously reported that elevated levels of IL-18, a proinflammatory cytokine, are independently associated with MetS in a healthy general population.25 In the present study, we examined for an independent (additive) association of adiponectin with inflammatory markers, insulin resistance and prevalent MetS in the same population.

Methods Study population The selection criteria and study design of the communitybased Carotid Ultrasound Disease Assessment Study have been detailed previously.26,27 In brief, this was a random electoral roll sample of 1111 predominantly Caucasian subjects aged 27–77, mean 53 years, from Perth, Western Australia, who were assessed for cardiovascular risk factors and had carotid B-mode ultrasound performed. The numbers of subjects were equally distributed between sexes and across each age decile. This present study sample was confined to the 1094 subjects (547 women, 547 men) who had a fasting plasma adiponectin concentration measured. Subjects also had available measurements of fasting serum glucose (n ¼ 1092), insulin (n ¼ 1088), high-sensitive C-reactive protein (hsCRP) (n ¼ 1059), IL-6 (n ¼ 1019) and IL-18 (n ¼ 1025). A self-administered questionnaire was used to record a clinical history of smoking, hypertension or diabetes. Anthropomorphic measurements and the lower of two resting blood pressures were recorded. Body mass index (BMI) was calculated as weight (kg)/height (m)2. The study protocol was approved by the Institutional Ethics Committee of the University of Western Australia. Written informed consent was obtained from all study participants. Definition of the metabolic syndrome and insulin resistance The 2001 National Cholesterol Education Program Adult Treatment Panel III (ATP III) report defined ATPIII–MetS as

three or more of the following characteristics: central obesity as measured by waist circumference 4102 cm in men, 488 cm in women; hypertriglyceridaemia X150 mg per 100 ml (1.7 mmol l1); low high-density lipoproteins (HDL) cholesterol o40 mg per 100 ml (1.03 mmol l1) in men and o50 mg per 100 ml (1.29 mmol l1) in women; high-blood pressure X130 mm Hg systolic or X85 mm Hg diastolic or current use of antihypertensive drugs and high fasting glucose X110 mg per 100 ml (6.1 mmol l1).1 Subjects were also grouped according to the new International Diabetes Federation (IDF) definition of IDF–MetS.28 This was defined as central obesity, measured by ethnicspecific waist circumferences, plus any two of the following: hypertriglyceridaemia, low HDL cholesterol, high-blood pressure and raised fasting glucose, the cutpoints for which were the same as the ATPIII criteria except for fasting glucose which used a lower cutpoint of X100 mg per 100 ml (5.6 mmol l1). We used the Europid waist circumference cutoffs (X94 cm in men, X80 cm in women) to define central obesity, as this was a predominantly Caucasian population. The total number of separate MetS components defined by ATPIII or IDF criteria, from 0 to 5, was also determined for each person. The insulin resistance index was assessed by the homoeostasis model assessment score (HOMA-IR), calculated as the product of the fasting serum glucose in mmol l1 and fasting insulin level in milliunit per litre, divided by 22.5.29

Biochemical analysis A fasting blood sample was obtained from each subject. Plasma adiponectin levels were measured by a commercially available quantitative sandwich enzyme immunoassay technique (R&D Systems Inc., Minneapolis, Minnesota, USA). The intra-assay coefficients of variation (CVs) ranged from 3.6 to 5.8% and the inter-assay CVs ranged from 7.0 to 8.5%. Serum hsCRP was measured by a microparticle turbidity assay with a range of 0.1–21.0 mg l1.27 Serum IL-6 was measured using an enzyme-linked immunosorbent assay (ELISA) (Quantikine HS, R&D Systems) with an assay range of 0.38–10.0 ng l1.27 Serum IL-18 was measured by a commercially available ELISA method (MBL Co. Ltd., Nagoya, Japan).25 Insulin was measured as milliunit per litre on a Tosoh A1A-600 immunoassay analyser using a two-site immunoenzymometric assay.25 Total cholesterol, HDL cholesterol and triglyceride levels were determined enzymatically with a Hitachi 747 autoanalyser. White cell count (WCC) was measured by a Coulter counter.

Statistical analysis The following positively skewed variables, adiponectin, glucose, insulin, HOMA-IR, hsCRP, IL-6 and IL-18 were log-transformed for analysis. Log mean values were transformed back and presented as geometric means. Means with s.e.m. or 95% confidence intervals (CIs) were provided for International Journal of Obesity


774 each continuous variable. Continuous variables were compared by paired t-test for paired samples and analysis of variance for more than two groups. Categorical variables were compared by w2-test. Linear regression analysis was used to determine independent correlates of the individual inflammatory markers and log HOMA-IR score. Stepwise variable selection procedures were used to determine a subset of independent predictors for each primary outcome, with age and sex always included as covariates. Apart from age and sex, waist–hip ratio, log HOMA-IR and smoking were the principal explanatory variables for inflammatory markers; BMI, ATPIII–MetS component score and WCC were the explanatory variables for log HOMA-IR; and BMI, log insulin and log IL-18 or WCC were selected as covariates for MetS classification. The association of adiponectin with the primary outcome was tested after adjustment for the above covariates as stated in Results. Stepwise logistic regression analysis was used to determine independent associates of ATPIII–MetS and IDF–MetS, and calculate odds ratio (OR) and 95% CI for each risk variable. Adiponectin was entered into the logistic model as sex-specific quartile groups. We used a BMI425 kg m2 to stratify the subgroup who were Table 1

overweight or obese. SPSS version 10.1 for windows was used in analysis. Statistical significance was taken as Po0.05.

Results Clinical and metabolic characteristics according to sex-specific quartiles of plasma adiponectin Women had higher plasma adiponectin levels than men (geometric mean 13.7 vs 7.9 mg l1, Po0.001). Hence, subjects were stratified by sex-specific adiponectin quartiles in Table 1. The cutpoints for plasma adiponectin quartiles in females were o9.7, 9.7–13.8, 13.9–20.0 and 420.0 mg l1; in males, cutpoints were o5.4, 5.4–7.9, 8.0–11.4 and 411.4 mg l1. Continuous variables were also age-adjusted since mean age increased across sex-specific adiponectin quartiles (Po0.001). Age-adjusted BMI, waist circumference, waist–hip ratio, diastolic blood pressure, triglycerides, glucose, insulin and HOMA-IR score were inversely correlated with sex-specific adiponectin quartiles (all Po0.005), while HDL cholesterol had a strong positive association (Po0.001) (Table 1). Systolic blood pressure, smoking and

Baseline clinical, metabolic and inflammatory marker characteristics of the population across sex-specific plasma adiponectin quartiles Sex-specific plasma adiponectin quartiles a Quartile 1 (n ¼ 272)

Quartile 2 (n ¼ 274)



P-value

Age-adjusted continuous variables Age (years) BMI (kg m2) Waist circumference (cm) Waist–hip ratio Systolic BP (mm Hg) Diastolic BP (mm Hg) Smoking (pack-years) LDL cholesterol (mmol l1) Triglyceride (mmol l1) HDL cholesterol (mmol l1) Fasting glucose (mmol l1)b Fasting insulin (mU l1)b HOMA-IRb hsCRP (mg l1)b Interleukin-6 (ng l1)b Interleukin-18 (mg l1)b Fibrinogen (g l1) WCC

49.3 27.9 89.4 0.86 130 82 16.9 3.7 1.6 1.17 5.6 7.2 1.8 2.4 4.0 331 2.9 6.8

(47.9, 50.8) (27.4, 28.4) (87.9, 90.8) (0.85, 0.87) (128, 132) (80, 83) (14.4, 20.0) (3.5, 3.8) (1.5, 1.7) (1.12, 1.21) (5.5, 5.8) (6.6, 7.7) (1.6, 1.9) (2.1, 2.7) (3.8, 4.2) (315, 349) (2.8, 2.9) (6.6, 7.0)

52.2 26.4 85.7 0.84 128 81 15.4 3.6 1.3 1.27 5.4 5.6 1.4 1.9 3.7 295 2.8 6.4

(50.8, 53.7) (25.9, 26.9) (84.3, 87.1) (0.83, 0.85) (126, 130) (80, 82) (12.8, 18.7) (3.5, 3.7) (1.3, 1.4) (1.23, 1.31) (5.3, 5.6) (5.2, 6.0) (1.2, 1.5) (2.6, 2.1) (3.5, 3.9) (280, 310) (2.7, 2.8) (6.2, 6.5)

53.9 25.6 83.7 0.82 129 80 15.6 3.6 1.2 1.39 5.3 4.7 1.1 1.7 3.6 290 2.7 6.3

(52.5, 55.4) (25.1, 26.1) (82.3, 85.1) (0.81, 0.83) (127, 131) (79, 82) (12.9, 18.9) (3.5, 3.7) (1.1, 1.3) (1.35, 1.43) (5.2, 5.4) (4.3, 5.0) (1.0, 1.2) (1.5, 2.0) (3.4, 3.8) (276, 305) (2.6, 2.8) (6.1, 6.4)

58.2 24.5 80.5 0.81 127 78 14.3 3.6 1.1 1.52 5.2 3.9 0.9 1.3 3.4 289 2.6 6.1

(56.8, 59.7) (24.0, 24.9) (79.1, 82.0) (0.80, 0.82) (125, 129) (77, 80) (11.8, 17.4) (3.5, 3.7) (1.0, 1.1) (1.48, 1.57) (5.1, 5.3) (3.6, 4.2) (0.8, 1.0) (1.1, 1.5) (3.2, 3.6) (274, 304) (2.5, 2.7) (5.9, 6.3)

o0.001 o0.001 o0.001 o0.001 0.14 0.003 0.64 0.82 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

Categorical variables Male gender (n, %) Hypertension (n, %) Current smoking (n, %) History MI or stroke (n, %) Diabetes (n, %) ATPIII–MetS (n, %) IDF–MetS (n, %)

137 45 55 21 26 90 109

(50.4%) (16.5%) (20.2%) (7.7%) (9.6%) (33.1%) (40.1%)

137 31 36 17 7 55 77

(50.0%) (11.3%) (13.1%) (6.2%) (2.6%) (20.1%) (28.1%)

137 47 36 17 13 41 65

(50.0%) (17.2%) (13.1%) (6.2%) (4.7%) (15.0%) (23.7%)

136 46 38 32 11 22 46

(49.6%) (16.8%) (13.9%) (11.7%) (4.0%) (8.0%) (16.8%)

0.99 0.18 0.057 0.057 0.002 o0.001 o0.001

Abbreviations: ATPIII, Adult Treatment Panel III; BP, blood pressure; BMI, body mass index; HDL, high-density lipoproteins; HOMA-IR, homoeostasis model assessment of insulin resistance; hsCRP, high-sensitive C-reactive protein; IDF, International Diabetes Federation; LDL, low-density lipoprotein; MetS, metabolic syndrome; MI, myocardial infarction; WCC, white cell count. aThe cutpoints for plasma adiponectin quartiles were o9.7, 9.7–13.8, 13.9–20.0 and 420.0 mg l1 in women and o5.4, 5.4–7.9, 8.0–11.4, and 411.4 mg l1 in men. bThe results for continuous variables are age-adjusted means or geometric means with 95% confidence intervals.

International Journal of Obesity


775 Men

Low

Medium

High

15.0

BMI

0.0

st -h ip

Low Medium High

5.0

ra t io

10.0

Low Medium High

ai

0.0

ai st -h ip

Low Medium High

5.0

ra tio

10.0

20.0

W

15.0

Plasma adiponectin mg/L

20.0

W

Plasma adiponectin mg/L

Women

BMI

Figure 1 Age-adjusted geometric mean adiponectin levels are shown according to low, medium or high-tertile groups of body mass index (BMI) and waist–hip ratio for women (left panel) and men (right panel).

Table 2

Association between log plasma adiponectin and inflammatory markers from linear regression analysis

Inflammatory marker

Model 1 b (s.e.)

Log hsCRP Log interleukin-6 Log interleukin-18 Fibrinogen WCC

0.28 0.09 0.05 0.13 0.37

(0.062) (0.026) (0.023) (0.037) (0.090)

Model 2

T

P-value

4.56 3.46 2.05 3.34 4.12

o0.001 0.001 0.04 0.001 o0.001

b (s.e.) 0.21 0.07 0.03 0.09 0.27

(0.064) (0.027) (0.024) (0.039) (0.092)

T

P-value

3.26 2.47 1.04 2.35 2.94

0.001 0.014 0.29 0.019 0.003

Abbreviations: b, standardized b-regression coefficient; hsCRP, high-sensitive C-reactive protein; s.e., standard error; T, the estimated coefficient, divided by its own standard error; WCC, white cell count. The dependent variable in each case is the inflammatory marker with log plasma adiponectin entered as an explanatory covariate in the linear regression analysis after adjustment for age, sex and waist–hip ratio (Model 1), and adjustment for age, sex, waist–hip ratio, log HOMA-IR and smoking pack-years (Model 2). T-values below 2 or above 2 are considered as useful predictors in the model.

LDL cholesterol were not associated with adiponectin levels (Table 1). In age- and sex-adjusted regression analysis, both waist–hip ratio (b ¼ 4.55, Po0.001) and BMI (b ¼ 0.11, P ¼ 0.002) were significant associates of log plasma adiponectin level. Figure 1 shows that increasing tertile levels of waist–hip ratio and BMI had additive negative effects on ageadjusted plasma adiponectin levels (P ¼ 0.017 for waist–hip ratio–BMI interaction). Women in the lowest compared to highest tertile groups of waist–hip ratio and BMI had E80% higher age-adjusted geometric mean adiponectin level (Po0.001), and men had E48% higher adiponectin level (Po0.001) (Figure 1).

Inflammatory markers according to level of plasma adiponectin Table 1 shows that age-adjusted levels of inflammatory markers, hsCRP, IL-6, IL-18, fibrinogen and WCC, declined progressively across sex-specific adiponectin quartiles (all Po0.001). Log plasma adiponectin was a significant inverse associate of the levels of all inflammatory markers after adjustment for age, sex and waist–hip ratio (Table 2). Adiponectin remained an independent associate of hsCRP,

Table 3 Linear regression analysis of variables with significant effect on log HOMA-IR score Independent variable Log plasma adiponectin Female BMI ATPIII–MetS component score WCC

b (s.e.) 0.19 0.16 0.05 0.21 0.02

(0.032) (0.038) (0.005) (0.017) (0.011)

T

P-value

6.18 5.96 10.25 12.56 1.90

o0.001 o0.001 o0.001 o0.001 0.057

Abbreviations: ATPIII, Adult Treatment Panel III; b, standardized b-regression coefficient; BMI, body mass index; MetS, metabolic syndrome; T, the estimated coefficient, divided by its own standard error; WCC, white cell count. The dependent variable is log HOMA-IR and the explanatory variables are gender, BMI, ATPIII–MetS component score, log plasma adiponectin and WCC. Age is not an independent predictor because of high colinearity with the other covariates in the model. Adjusted R2 for the full model is 0.42. Tvalues below 2 or above 2 are considered as useful predictors in the model.

IL-6, fibrinogen and WCC, but not IL-18, after additional adjustment for log HOMA-IR score and smoking (Table 2).

Plasma adiponectin and HOMA-IR Log plasma adiponectin was an independent associate of log HOMA-IR score (b ¼ 0.19, Po0.001) after adjusting for sex, International Journal of Obesity


776 BMI, ATPIII–MetS component score and WCC, as significant covariates (Table 3). The full model accounted for E42% of the total variance of HOMA-IR score. When subjects were

Figure 2 Age- and sex-adjusted geometric mean adiponectin levels in milligram per litre are shown with standard error bars across groups according to number of metabolic syndrome components, ranging from 0 to 5, using National Cholesterol Education Program Adult Treatment Panel III (ATPIII) criteria, in subgroups separated by body mass index (BMI) p25 kg m2 (hatched bars) and body mass index (BMI) 425 kg m2 (solid black bars). There were no patients in the normal BMI subgroup with more than four metabolic components. See Methods for detailed description of criteria for metabolic syndrome components.

stratified by BMI level, adiponectin remained an independent associate of log HOMA-IR in the 442 subjects with normal BMI (b ¼ 0.23, Po0.001) and 577 subjects with increased BMI (b ¼ 0.15, Po0.001).

Plasma adiponectin and metabolic syndrome MetS by ATPIII and IDF criteria was present in 208 (19.0%) and 297 (27.1%), respectively, of the whole population. The lower prevalence of ATPIII–MetS than IDF–MetS was due to the higher cutpoints used by ATPIII to define central obesity and hyperglycaemia. The proportion of subjects with ATPIII– MetS or IDF–MetS declined across sex-specific adiponectin quartiles (both Po0.001) (Table 1). There was also a higher prevalence of diabetes history in the bottom sex-specific adiponectin quartile group compared to the other groups (9.6 vs 3.8%, P ¼ 0.002) (Table 1). Figure 2 shows that age- and sex-adjusted geometric mean adiponectin levels were inversely related to the number of ATPIII–MetS components in subgroups below and above a BMI of 25 kg m2 (both trend Po0.001). Individuals with compared to those without ATPIII–MetS had a geometric mean (age- and sex adjusted) adiponectin level of 7.5 vs 11.2

Figure 3 The top panels (a, b) show the adjusted odds ratios (’) and their 95% confidence interval bars for metabolic syndrome defined by Adult Treatment Panel III criteria (ATPIII–MetS) (a) and International Diabetes Federation criteria (IDF–MetS) (b) in the whole population according to sex-specific plasma adiponectin quartiles, with quartile 1 as reference (odds ratio 1.0). The model was adjusted for age, sex, body mass index (BMI), log insulin and log interleukin-18 or white cell count (WCC). The bottom panels (c, d) show the adjusted odds ratios (’) and their 95% confidence interval bars according to sex-specific plasma adiponectin quartiles for ATPIII–MetS in the population stratified by BMI p25 kg m2 (c) and BMI 425 kg m2 (d) The model was adjusted for the same covariates as for the whole population except for BMI.



777 (Po0.001); for IDF–MetS, mean adiponectin levels were 8.2 vs 11.3 mg l1, respectively (Po0.001). In the model without adiponectin, logistic regression analysis identified age, sex, BMI, log insulin and log IL-18 as independent associates of ATPIII–MetS, and age, sex, BMI, log insulin and WCC as associates of IDF–MetS. After adjusting for these covariates, sex-specific plasma adiponectin quartiles significantly added to the prediction for MetS. Figure 3 illustrates that the adjusted ORs for MetS declined across increasing sex-specific plasma adiponectin quartiles. Individuals in the top compared to bottom quartile had an adjusted OR of 0.21 (95% CI, 0.11–0.42; Po0.001) for ATPIII–MetS, and OR of 0.46 (95% CI, 0.26–0.80; P ¼ 0.007) for IDF–MetS (Figures 3a and b). When stratified by BMI, subjects in the top compared to bottom adiponectin quartile had a reduced (adjusted) likelihood of ATPIII–MetS, with a similar OR in lean (OR 0.10, 95% CI, 0.08–0.46, P ¼ 0.009) as compared with overweight/obese subjects (OR 0.23, 95% CI, 0.12–0.26, Po0.001) (Figures 3c and d).

Discussion This was a large population-based study of men and women, across a wide age span, encompassing a broad range of adiposity and HOMA-IR scores. This was a generally healthy population with a low prevalence of diabetes and clinical vascular disease, hence reducing the likelihood of reverse causation. The major findings of the study were as follows: (1) plasma adiponectin concentrations were inversely related to many circulating inflammatory markers, independent of obesity and insulin resistance; (2) plasma adiponectin levels were inversely associated with insulin resistance scores independent of obesity and (3) higher circulating adiponectin concentrations were associated with a reduced prevalence of MetS, and the relationship was independent of the level of obesity, insulin resistance and inflammatory markers. Our data confirm that gender and visceral obesity are major determinants of plasma adiponectin concentrations.5–7,30 Higher levels of BMI and waist–hip ratio, reflecting total and visceral adiposity, were associated with an additive fall in plasma adiponectin concentrations (Figure 1). In accordance, adiponectin mRNA levels have been found to be lower in adipose tissue from obese as compared to those in lean subjects, and lower in visceral vs subcutaneous adipose tissue.6,31 Men had considerably lower plasma adiponectin levels than women, concordant with previous studies.7,30 This may partly relate to inhibitory effects of androgens on adiponectin levels.32 There is evidence for the participation of adipokines such as TNF-a, IL-6 and IL-18 in the development of a systemic inflammatory state that contributes to MetS.3,25 In contrast to other adipokines, adiponectin appears to have antiinflammatory actions.16–19 In accordance, we observed a significant inverse association of circulating adiponectin

with the inflammatory markers, hsCRP, IL-6, fibrinogen and WCC, after adjustment for significant covariates (Table 2). Previous studies have been inconsistent, with some reporting an association of circulating adiponectin with inflammatory markers,24,33,34 while other studies have not found a significant association.35,36 Some of these discrepancies may be explained by the different characteristics of the study populations regarding age, sex, smoking and levels of obesity and insulin resistance. In the present study, we have adjusted for these confounders, and confirmed an independent association of circulating adiponectin with a range of systemic inflammatory markers. We also found a strong association between circulating adiponectin and MetS components in accordance with the insulin-sensitizing actions of adiponectin, and its participation in glucose and lipid metabolism.8–10 Our results confirm previous findings across diverse ethnic groups.11,12,22,24,30 Plasma adiponectin level was also strongly associated with the homoeostasis model assessment of insulin resistance, again supported by other population-based studies.11–13,24 Abbasi et al.37 provided initial evidence of a relationship between adiponectin concentrations and insulin resistance independent of obesity in a study of 60 nondiabetic subjects. We confirmed this important finding in a large communitybased study showing that plasma adiponectin remained a strong inverse associate of HOMA-IR score after adjusting for obesity and other major covariates (Table 3). Plasma adiponectin remained inversely associated with HOMA-IR score in subjects with a normal BMI, further indicating that the association of circulating adiponectin with insulin sensitivity is regardless of obesity. Although it is widely believed that adiponectin is a key participant in developing MetS, there are few comprehensive general population-based studies.4,20 Choi et al.21 in a prospective study of 372 elderly Koreans found adiponectin levels to be strongly associated with developing diabetes and MetS. Mohan et al.22 in a small case–control study of Asian Indians showed that lower adiponectin levels were associated with prevalent MetS. Matsushita et al.23 performed a cross-sectional study of 624 Japanese middle-aged men and found that adiponectin level was a more significant predictor than TNFa, IL-6 or CRP for prevalent MetS, although age and smoking were the only covariates adjusted for in their analysis. However, Wannamethee et al.24 in the British Regional Heart Study of 3640 nondiabetic men aged 60–79 years did adjust for BMI among other covariates and found that the likelihood of MetS decreased significantly with increasing adiponectin. Nevertheless, measures of insulin resistance and inflammation were not included as covariates in the model.24 By comparison, our community-based study included both men and women across a wide age span and range of adiposity, and included measures of inflammatory markers and insulin resistance. We have previously reported that elevated levels of IL-18, in addition to obesity and hyperinsulinaemia, were independent associates of prevalent International Journal of Obesity


778 ATPIII–MetS in this population.25 We have now extended our findings to plasma adiponectin, the circulating level of which was additionally associated with prevalent MetS, such that persons with adiponectin concentrations in the top compared to bottom sex-specific quartile had a E80% (adjusted) odds reduction of ATPIII–MetS and E50% odds reduction of IDF–MetS. Adiponectin was as strongly related to prevalent MetS in persons who were lean as compared to those who were obese (Figure 3c), indicating again that the association is independent of adiposity. Irrespective of causality, we have demonstrated plasma adiponectin to be a very strong risk marker for MetS in the general population, which is independent of measures of adiposity, insulin resistance and inflammatory markers. We acknowledge some study limitations. In a crosssectional study, causality cannot be proven, and prospective studies would be helpful to confirm our findings. Diabetic subjects were not excluded because of small numbers and their exclusion would not have materially affected the results. In humans, adiponectin circulates as different molecular weight complexes, and it appears that it is the high molecular weight (HMW) isoforms that are more relevant in the prediction of insulin resistance.4,20 Although a high correlation between HMW and total adiponectin has been demonstrated, future studies which can assess the HMW-to-total adiponectin ratio may be more informative.4,20 In summary, our findings suggest that higher adiponectin levels may mitigate against adipose-related inflammation, insulin resistance and metabolic disorders as much in lean as in obese individuals. Interventions that enhance adiponectin secretion or action may therefore have potential for treatment of metabolic and vascular disease, although the clinical evidence for adiponectin being protective against cardiovascular disease is conflicting.38 At any rate circulating adiponectin level is a strong risk marker for MetS independent of measures of adiposity.

Acknowledgements This study was supported in part by a grant-in-aid from the National Health and Medical Research Council (211980) and HeartSearch, Perth, Western Australia (JB). Disclosure/Conflict of interest None.

3 4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

References 19 1 National Cholesterol Education Program (NCEP). Third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III) final report. Circulation 2002; 106: 3143–3421. 2 Caterson ID, Hubbard V, Bray GA, Grunstein R, Hansen BC, Hong Y et al. Prevention conference VII: obesity, a worldwide epidemic


20

related to heart disease and stroke: group III: worldwide comorbidities of obesity. Circulation 2004; 110: e476–e483. Berg AH, Scherer PE. Adipose tissue, inflammation, and cardiovascular disease. Circ Res 2005; 96: 939–949. Matsuzawa Y. Adiponectin: identification, physiology and clinical relevance in metabolic and vascular disease. Atherosclerosis Supplements 2005; 6: 7–14. Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novel serum protein similar to c1q, produced exclusively in adipocytes. J Biol Chem 1995; 270: 26746–26749. Hu E, Liang P, Spiegelman BM. AdipoQ is a novel adiposespecific gene dysregulated in obesity. J Biol Chem 1996; 271: 10697–10703. Arita Y, Kihara S, Ouchi N, Takahashi M, Maeda K, Miyagawa J et al. Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochem Biophys Res Commun 1999; 257: 79–83. Yamauchi T, Kamon J, Minokoshi Y, Ito Y, Waki H, Uchida S et al. Adiponectin stimulates glucose utilization and fatty-acid oxidation by activating amp-activated protein kinase. Nat Med 2002; 8: 1288–1295. Berg AH, Combs TP, Du X, Brownlee M, Scherer PE. The adipocyte-secreted protein acrp30 enhances hepatic insulin action. Nat Med 2001; 7: 947–953. Yamauchi T, Kamon J, Waki H, Terauchi Y, Kubota N, Hara K et al. The fat-derived hormone adiponectin reverses insulin resistance associated with both lipoatrophy and obesity. Nat Med 2001; 7: 941–946. Yamamoto Y, Hirose H, Saito I, Tomita M, Taniyama M, Matsubara K et al. Correlation of the adipocyte-derived protein adiponectin with insulin resistance index and serum highdensity lipoprotein-cholesterol, independent of body mass index, in the Japanese population. Clin Sci 2002; 103: 137–142. Tschritter O, Fritsche A, Thamer C, Haap M, Shirkavand F, Rahe S et al. Plasma adiponectin concentrations predict insulin sensitivity of both glucose and lipid metabolism. Diabetes 2003; 52: 239–243. Nilsson PM, Engstrom G, Hedblad B, Frystyk J, Persson MM, Berglund G et al. Plasma adiponectin levels in relation to carotid intima media thickness and markers of insulin resistance. Arterioscler Thromb Vasc Biol 2006; 26: 2758–2762. Spranger J, Kroke A, Mohlig M, Bergmann MM, Ristow M, Boeing H et al. Adiponectin and protection against type 2 diabetes mellitus. Lancet 2003; 361: 226–228. Duncan BB, Schmidt MI, Pankow JS, Bang H, Couper D, Ballantyne CM et al. Adiponectin and the development of type 2 diabetes: the atherosclerosis risk in communities study. Diabetes 2004; 53: 2473–2478. Ouchi N, Kihara S, Arita Y, Maeda K, Kuriyama H, Okamoto Y et al. Novel modulator for endothelial adhesion molecules: adipocyte-derived plasma protein adiponectin. Circulation 1999; 100: 2473–2476. Ouchi N, Kihara S, Arita Y, Okamoto Y, Maeda K, Kuriyama H et al. Adiponectin, an adipocyte-derived plasma protein, inhibits endothelial NF-{kappa}B signaling through a cAMP-dependent pathway. Circulation 2000; 102: 1296–1301. Yokota T, Oritani K, Takahashi I, Ishikawa J, Matsuyama A, Ouchi N et al. Adiponectin, a new member of the family of soluble defense collagens, negatively regulates the growth of myelomonocytic progenitors and the functions of macrophages. Blood 2000; 96: 1723–1732. Dietze-Schroeder D, Sell H, Uhlig M, Koenen M, Eckel J. Autocrine action of adiponectin on human fat cells prevents the release of insulin resistance-inducing factors. Diabetes 2005; 54: 2003–2011. Kadowaki T, Yamauchi T, Kubota N, Hara K, Ueki K, Tobe K. Adiponectin and adiponectin receptors in insulin resistance, diabetes, and the metabolic syndrome. J Clin Invest 2006; 116: 1784–1792.


779 21 Choi KM, Lee J, Lee KW, Seo JA, Oh JH, Kim SG et al. Serum adiponectin concentrations predict the developments of type 2 diabetes and the metabolic syndrome in elderly Koreans. Clin Endocrinol 2004; 61: 75–80. 22 Mohan V, Deepa R, Pradeepa R, Vimaleswaran KS, Mohan A, Velmurugan K et al. Association of low adiponectin levels with the metabolic syndromeFthe Chennai Urban Rural Epidemiology Study (CURES-4). Metabolism 2005; 54: 476–481. 23 Matsushita K, Yatsuya H, Tamakoshi K, Wada K, Otsuka R, Takefuji S et al. Comparison of circulating adiponectin and proinflammatory markers regarding their association with metabolic syndrome in Japanese men. Arterioscler Thromb Vasc Biol 2006; 26: 871–876. 24 Wannamethee SG, Tchernova J, Whincup P, Lowe GD, Rumley A, Brown K et al. Associations of adiponectin with metabolic and vascular risk parameters in the British regional heart study reveal stronger links to insulin resistance-related than to coronory heart disease risk-related parameters. Int J Obes 2007; 31: 1089–1097. 25 Hung J, McQuillan BM, Chapman CML, Thompson PL, Beilby JP. Elevated interleukin-18 levels are associated with the metabolic syndrome independent of obesity and insulin resistance. Arterioscler Thromb Vasc Biol 2005; 25: 1268–1273. 26 McQuillan BM, Beilby JP, Nidorf M, Thompson PL, Hung J. Hyperhomocysteinemia but not the C677T mutation of methylenetetrahydrofolate reductase is an independent risk determinant of carotid wall thickening. The Perth Carotid Ultrasound Disease Assessment Study (CUDAS). Circulation 1999; 99: 2383–2388. 27 Chapman CML, Beilby JP, McQuillan BM, Thompson PL, Hung J. Monocyte count, but not C-reactive protein or interleukin-6, is an independent risk marker for subclinical carotid atherosclerosis. Stroke 2004; 35: 1619–1624. 28 Alberti KGM, Zimmet P, Shaw J. The metabolic syndromeFa new worldwide definition. Lancet 2005; 366: 1059–1062. 29 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and

30

31

32

33

34

35

36

37

38

beta-cell function from fasting insulin and glucose concentrations in man. Diabetologia 1985; 28: 412–419. Lihn AS, Bruun JM, He G, Pedersen SB, Jensen PF, Richelsen B. Lower expression of adiponectin mRNA in visceral adipose tissue in lean and obese subjects. Mol Cell Endocrinol 2004; 219: 9–15. Cnop M, Havel PJ, Utzschneider KM, Carr DB, Sinha MK, Boyko EJ et al. Relationship of adiponectin to body fat distribution, insulin sensitivity and plasma lipoproteins: Evidence for independent roles of age and sex. Diabetologia 2003; 46: 459–469. Nishizawa H, Shimomura I, Kishida K, Maeda N, Kuriyama H, Nagaretani H et al. Androgens decrease plasma adiponectin, an insulin-sensitizing adipocyte-derived protein. Diabetes 2002; 51: 2734–2741. Ouchi N, Kihara S, Funahashi T, Nakamura T, Nishida M, Kumada M et al. Reciprocal association of c-reactive protein with adiponectin in blood stream and adipose tissue. Circulation 2003; 107: 671–674. Schulze MB, Rimm EB, Shai I, Rifai N, Hu FB. Relationship between adiponectin and glycemic control, blood lipids, and inflammatory markers in men with type 2 diabetes. Diabetes Care 2004; 27: 1680–1687. Rothenbacher D, Brenner H, Marz W, Koenig W. Adiponectin, risk of coronary heart disease and correlations with cardiovascular risk markers. Eur Heart J 2005; 26: 1640–1646. Herder C, Hauner H, Haastert B, Rohrig K, Koenig W, Kolb H et al. Hypoadiponectinemia and proinflammatory state: two sides of the same coin?: results from the cooperative health research in the region of Augsburg Survey 4 (KORA S4). Diabetes Care 2006; 29: 1626–1631. Abbasi F, Chu JW, Lamendola C, McLaughlin T, Hayden J, Reaven GM et al. Discrimination between obesity and insulin resistance in the relationship with adiponectin. Diabetes 2004; 53: 585–590. Wannamethee SG, Whincup PH, Lennon L, Sattar N. Circulating adiponectin levels and mortality in elderly men with and without cardiovascular disease and heart failure. Arch Intern Med 2007; 167: 1510–1517.
