|Home | About | Journals | Submit | Contact Us | Français|
We evaluated the hypothesis that plasma levels of adiponectin and leptin are independently but oppositely associated with coronary calcification (CAC), a measure of subclinical atherosclerosis. In addition, we assessed which biomarkers of adiposity and insulin resistance are the strongest predictors of CAC beyond traditional risk factors, the metabolic syndrome and plasma C-reactive protein (CRP).
Adipokines are fat-secreted biomolecules with pleiotropic actions that converge in diabetes and cardiovascular disease.
We examined the association of plasma adipocytokines with CAC in 860 asymptomatic, non-diabetic participants in the Study of Inherited Risk of Coronary Atherosclerosis (SIRCA).
Plasma adiponectin and leptin levels had opposite and distinct associations with adiposity, insulin resistance and inflammation. Plasma leptin was positively (top vs. bottom quartile) associated with higher CAC after adjusting for age, gender, traditional risk factors and Framingham Risk Scores (FRS) [tobit regression ratio 2.42 (95% CI 1.48–3.95, p=0.002)] and further adjusting for metabolic syndrome and CRP [ratio 2.31 (95% CI 1.36–3.94, p=0.002)]. In contrast, adiponectin levels were not associated with CAC. Comparative analyses suggested that levels of leptin, IL-6 and sol-TNFR2 as well as HOMA-IR predicted CAC scores but only leptin and HOMA-IR provided value beyond risk factors, the metabolic syndrome and CRP.
In SIRCA, while both leptin and adiponectin levels were associated with metabolic and inflammatory markers, only leptin was a significant independent predictor of CAC. Of several metabolic markers, leptin and the HOMA-IR index had the most robust, independent associations with CAC.
Adipokines are fat-secreted biomolecules with pleiotropic actions and represent novel markers for cardiovascular risk. We examined the association of plasma adipocytokines with CAC in 860 asymptomatic, non-diabetic Caucasians. Leptin was positively (top vs. bottom quartile) associated with higher CAC even after adjustment for age, gender, traditional risk factors, Framingham Risk Score, metabolic syndrome, and CRP [ratio 2.31 (95% CI 1.36–3.94, p=0.002)]. Adiponectin levels were not associated with CAC. Comparative analyses suggested that levels of leptin, IL-6 and sol-TNFR2 as well as HOMA-IR predicted CAC scores, but only leptin and HOMA-IR provided value beyond risk factors, the metabolic syndrome and CRP.
Adipokines are fat-secreted biomolecules with diverse signaling effects that modulate insulin resistance, hepatic lipoprotein production and vascular inflammation (1). Two in particular, adiponectin and leptin, are almost exclusively fat-derived and have antithetical actions in insulin resistance and in vascular signaling (2). Because of these properties, adiponectin and leptin have been proposed as biomarkers of adipose function that may add value in predicting cardiovascular disease (CVD) risk and provide targets for therapeutic interventions.
Levels of adiponectin, an insulin sensitizing hormone with anti inflammatory properties (3), are reduced in obesity, type 2 diabetes and coronary artery disease (CAD) compared to controls (4–6). Indeed, several (7,8) but not all (9,10), epidemiological studies suggest that reduced plasma adiponectin levels are independent predictors of CVD. Leptin, on the other hand, is a pleiotropic adipokine that modulates innate immune functions and vascular signaling in addition to its central role in regulation of appetite and energy expenditure (11). In contrast to adiponectin, leptin levels directly correlate with insulin resistance, obesity (12,13) and several CVD risk factors (14). Leptin levels have been associated with CVD beyond BMI, in some (15–17), but not all studies (18).
We previously examined the association of plasma levels of CRP, resistin, interleukin-6 (IL-6) and soluble TNF receptor 2 (sol-TNFR2), as well as the metabolic syndrome, with coronary artery calcification (CAC) in the Study of Inherited Risk of Coronary Atherosclerosis (SIRCA) (19–22). In this report, we examined the association of adiponectin and leptin with CVD risk factors and CAC in SIRCA and then compared the relative value of all measured biomarkers of adiposity and insulin resistance in predicting CAC scores beyond traditional risk factors, metabolic syndrome and plasma CRP.
The Study of Inherited Risk of Coronary Atherosclerosis (SIRCA) is a single center community based cross-sectional study of factors associated with Coronary Artery Calcium (CAC) (21,23). Participants were healthy adults aged 30–75 with a family history of premature CVD, but without evidence of clinical CAD (defined as myocardial infarction, coronary revascularization, angiographic evidence of CAD, or ischemia seen on a cardiac stress test), diabetes, elevated serum creatinine >3.0 mg/dl or elevated total cholesterol (>300mg/dL). This report focuses on 860 unrelated, non-diabetic SIRCA participants.
Study subjects were evaluated in a fasting state at the GCRC at the Hospital of the University of Pennsylvania (21,23). Plasma levels of adiponectin, leptin, resistin and insulin (Linco, St Charles MO), as well as IL-6 and sol-TNFR2 (R+D Systems, Minneapolis) were measured by ELISAs. CRP levels were assayed as described (21). The intra- and inter-assay c.v.’s for pooled human plasma were 5.7% and 9.9% for adiponectin; 5.5% and 12.4% for leptin; 4.6% and 4.3% for resistin; 4.1% and 11.6% for insulin; 8.7% and 10.9% for IL-6; 5.3% and 12.1% for sol-TNFR2; and 8.0% and 8.3% for CRP respectively. Framingham Risk Scores (FRS), were calculated as described by Wilson et. al. (24). Participants were classified as having the metabolic syndrome using the National Cholesterol Education Program (NCEP) definition (25). The homeostasis model assessment (HOMA-IR index = fasting glucose (mmol/L) × fasting insulin (µU/mL)/22.5) (26) was used as a measure of insulin resistance. Global Agatston CAC scores (27), measured at electron beam tomography (Imatron, San Francisco, CA) were determined as previously described (21,23).
Data are reported as median with first and third quartiles (Q1 = 25th percentile, Q3 = 75th percentile), or mean ± SD, for continuous variables, and as proportions for categorical variables. Spearman correlations of plasma adiponectin and leptin with other continuous variables are presented. Crude associations of adipokine levels with categorical variables were examined using the Kruskal-Wallis rank test. Tobit regression, using natural log (CAC+1) as the outcome, was used for the analysis of CAC data because of its marked right skewed distribution and the presence of many zero scores (28). The tobit model is designed to assess the relationship between explanatory variables and a censored dependent variable at one end, where many observations are clustered. We chose this modeling since CAC scores are censored at zero and the use of ordinary least-squares regression on such a non normal distribution would produce biased estimates and invalid inference. Tobit modeling has otherwise similar assumptions about error distributions as the linear regression model.
Because of potential gender differences in adipose associations with CVD, models are presented for each gender separately and combined when appropriate. The association between CAC and highest vs. lowest quartile of adiponectin and leptin were assessed in incremental models including the variables age (age and age2), race, gender, family history of CAD, exercise (none versus any), medications (aspirin, statins, angiotensin converting enzyme inhibitors), Framingham Risk Score, metabolic syndrome, and CRP. Gender differences in the association of adipokines with CAC were assessed using the likelihood-ratio test (LRT). A priori, BMI data was not included in the models because adipokines may be intermediate in the causal pathway between adiposity and sub-clinical atherosclerosis.
We used the LRT in nested models to assess the incremental value of each biomarker of adiposity and insulin resistance, adiponectin, leptin, resistin, IL-6, sol-TNFR2, HOMA-IR data (all included as log transformed variables) and metabolic syndrome, in predicting CAC scores beyond established risk factors. Statistical analyses were performed using Stata 9.0 software (Stata Corp, College Station, TX. A tobit regression model was fit in Stata (http://www.ats.ucla.edu/stat/stata/dae/tobit.htm) using the tobit command with the ll(0) option to indicate left-censoring at a CAC score of zero.
As previously described, (19,21) the SIRCA sample is predominantly Caucasian (Table 1). Plasma levels of adiponectin and leptin were significantly higher in women than in men (p< 0.001 for both). Almost 40% had CAC scores above the 70th percentile, consistent with accelerated atherosclerosis most likely related to recruitment strategy based on family history of CVD.
Adiponectin and leptin correlated only modestly (and inversely) with each other while associations with lipid, metabolic and inflammatory variables were greater for both adipokines in women than men (Appendix Table A). Among all factors, adiponectin’s strongest (direct) correlation was with plasma levels of HDL cholesterol, whereas it was only modestly and inversely correlated with HOMA-IR, adiposity and inflammatory markers. In contrast, leptin was strongly associated with BMI, waist circumference and HOMA, moderately correlated with inflammatory markers, blood pressure and apoB lipoproteins but only weakly inversely related to HDL cholesterol.
In simple models, adiponectin levels were not significantly correlated with CAC in either men or women while leptin had strong direct associations with CAC across gender (Table 2). There were suggestive gender differences in the association of adipokines with CAC in cruder models, which were significant for adiponectin in age and race adjusted models (p<0.01). In fully adjusted models even after controlling for FRS, metabolic syndrome and plasma CRP levels, the top quartile of leptin, but not adiponectin, was significantly associated with CAC scores (Table 2).
In SIRCA there was no evidence for major gender differences in the association of metabolic syndrome, HOMA-IR, IL-6, sol-TNFR2, and resistin with CAC in adjusted analyses (e.g., p=0.10, 0.37, 0.68, 0.18 respectively for gender difference in age and race adjusted models). Except for adiponectin therefore, results of these analyses are presented for both genders combined. The HOMA-IR index, plasma levels of leptin, IL-6 and sol-TNFR2 as well as the NCEP defined metabolic syndrome (glucose cut-point >110 mg/dL), provided significant improvements in the association with CAC beyond traditional risk factors and the FRS (Table 3). After further adjustment for the metabolic syndrome and CRP data, only HOMA-IR (LRT χ 2 10.39, p<0.01) and plasma leptin levels (LRT χ2 6.87, p<0.01) significantly improved model prediction of CAC (Table 3).
Adiponectin and leptin are fat secreted hormones with opposing actions on insulin resistance and vascular inflammation. While plasma leptin and adiponectin had opposite correlations with lipid, metabolic and inflammatory risk factors, we found that only plasma leptin levels were independently associated with CAC. Further, in a comparison of several metabolic biomarkers, leptin and the HOMA-IR index of insulin resistance had the most robust associations with CAC scores beyond traditional risk factors, NCEP defined metabolic syndrome and plasma levels of CRP.
Leptin is an important negative regulator of body weight (11). Paradoxically, obesity is associated with increased plasma leptin levels, most likely due to resistance to its actions in the setting of increased production by adipose tissue (29). Leptin activates the endothelium, induces smooth muscle cell proliferation and its receptors are expressed in atherosclerotic plaques (30). Recent studies suggest an association between plasma leptin levels and atherosclerotic CVD in humans including angiographic CAD (31) and CVD events (32). In a case (n=377) control (n=783) study nested within the WOSCOPS clinical trial, plasma leptin levels predicted CVD even after adjusting for traditional risk factors, BMI and plasma CRP levels (15). However, in a nested case-control study from the Quebec Cardiovascular Study Cohort, plasma leptin levels were not related to CVD events (18).
Few data are available on the association between leptin and direct measures of atherosclerosis in humans. Van den Beld et. al. found no association between plasma leptin levels with carotid intima-media thickness (IMT) in 403 healthy elderly men (33), while Ciccone et. al. reported an association of leptin with IMT in 126 healthy Italians (34). We previously reported that leptin levels were associated with CAC in a type-2 diabetic sample even after controlling for establish risk factors including CRP and measures of sub-clinical vascular disease (16). Recently, Iribarren et. al. reported an association of plasma leptin levels with CAC in older women in the ADVANCE study, but this association was not significant after controlling for metabolic risk factors and BMI data (35). In SIRCA, we found an association of plasma leptin with CAC even after controlling for metabolic syndrome and CRP.
Adiponectin has emerged as a unique fat secreted hormone that regulates insulin sensitivity (36). Atheroprotective effects may be directed through inhibition of the NF-κB inflammatory pathway in vascular cells (37) and by attenuation of foam cell formation (38). Plasma levels are depressed in patients with CAD (39) and are associated with clinical CVD in diabetics (7). A nested case control study by Maahs and colleagues suggested low plasma adiponectin predicted short term CAC progression, more so in non diabetics (40). Several recent prospective studies of clinical CVD, however, have been negative. In a nested case-control study from the Strong Heart Study, there was no association with incident CAD events (10). Similarly, in the British Women's Heart and Health Cohort Study, adiponectin levels were not associated with CVD (9). More recently, Sattar et. al looked at 589 men with fatal and non fatal CAD and 1231 controls and found no difference in median adiponectin levels despite adiponectin associations with HDL and CRP. A seven-study meta-analysis by the same authors failed to demonstrate a consistent relationship of adiponectin with CAD events (41).
Despite correlations with lipids, metabolic factors and insulin resistance, we also did not find an inverse association of adiponectin levels with CAC. The reasons for conflicting study findings are uncertain but may reflect differences in study design and populations as well heterogeneous outcomes including sub-clinical atherosclerotic and different CVD outcomes. In fact, Steffes and colleagues unexpectedly found a positive association of adiponectin with CAC in a study of over 3,000 young adults aged 33 to 45 years (42). Finally, several studies suggest that the high molecular weight adiponectin complex, but not the lower molecular weight hexamer, may be the active signaling molecule (43,44). Few epidemiological studies, however, have assayed the different circulating forms of adiponectin.
We also determined which of several metabolic biomarkers predicted CAC scores beyond established clinical CVD risk factors. Leptin, HOMA-IR, and, to a lesser extent, IL-6 and sol-TNFR2, provided incremental value beyond FRS, metabolic syndrome and CRP. Our finding that HOMA-IR levels are associated with CAC beyond all other risk factors is consistent with most (22) but not all (45) studies which found hyperinsulinemia and insulin resistance indices were independently associated with atherosclerosis and CVD. The clinical application of insulin-based measures, however, is challenging given the lack of assay standardization and because of ultradian and circadian variation in circulating insulin.
This study has several limitations. The study sample is cross-sectional and is not capable of determining causal relationships. Moreover, it is a study of a population consisting primarily of Caucasians with a family history of premature CVD who are otherwise deemed to be at low risk, therefore the generalizability of our findings across other populations and ethnic groups is uncertain. In addition, CAC is not a direct measure of coronary atherosclerosis. In autopsy studies, however, CAC has been shown to be a quantitative estimate of coronary atherosclerosis (46). It has also been shown to be an independent predictor of CVD (47).
In summary, we found that plasma levels of leptin but not adiponectin were associated with CAC after controlling for traditional cardiovascular risk factors, metabolic syndrome and CRP levels. Whether leptin signaling promotes human atherosclerotic CVD directly remains to be established. Finally, leptin levels and the HOMA-IR index had stronger associations with CAC scores than other adipocytokines in this asymptomatic sample. A systematic comparison of multiple adipocytokine and insulin resistance biomarkers across diverse clinical settings is warranted in order to establish which provide utility as metabolic biomarkers of clinical CVD.
This work was supported by a Clinical and Translational Science Award (RFA-RM-06-002) from the National Center for Research Resources (NCRR) and by a Diabetes Endocrinology Research Award (P30 DK-019525) from the NIH (Bethesda, MD) to the University of Pennsylvania, by RO1 HL-073278, RO1 DK-021224, P50 HL-083799 (SCCOR) and W.W. Smith Charitable Trust (West Conshohocken, PA; Grant #H0204) awards to MR.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
There are no conflicts of interest.
Atif Qasim, Cardiovascular Division and Center for Experimental Therapeutics, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA.
Nehal N. Mehta, Cardiovascular Division and Center for Experimental Therapeutics, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA.
Mahlet G. Tadesse, Department of Mathematics, Georgetown University, Washington, DC.
Megan L. Wolfe, Cardiovascular Division and Center for Experimental Therapeutics, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA.
Thomas Rhodes, The Department of Epidemiology, Merck Research Laboratories, West Point, PA.
Cynthia Girman, The Department of Epidemiology, Merck Research Laboratories, West Point, PA.
Muredach P Reilly, Cardiovascular Division and Center for Experimental Therapeutics, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA.