• adiponectin;
  • visceral fat;
  • macrovascular disease;
  • oxidative stress


  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Objective: Adiponectin is a collagen-like product of visceral fat that offers apparent protection against macrovascular disease. We evaluated the relationships of concentrations of adiponectin with oxidative stress and the major risk factors for and/or the presence of macrovascular disease.

Research Methods and Procedures: Adiponectin was measured by radioimmunoassay in serum from 3045 fasting participants (ages 33 to 45) of the Coronary Artery Risk Development in Young Adults Study. Cross-sectional correlation of the concentrations of adiponectin with F2-isoprostane concentrations (a marker of systemic oxidative damage), coronary artery calcification (CAC; an estimate of early macrovascular disease), and several macrovascular risk factors was analyzed.

Results: F2-isoprostanes and CAC were unrelated to adiponectin after minimal adjustment for gender, race, and center. After additional adjustment for insulin resistance and waist circumference and other macrovascular risk factors, adiponectin correlated positively with high-density lipoprotein-cholesterol (p < 0.0001), F2-isoprostanes (p < 0.0001), and CAC (less strongly, p < 0.01) and negatively with triglycerides (p < 0.0001) and C-reactive protein (marking inflammation, p = 0.01).

Discussion: Although these data are consistent with reduced cardiovascular disease risk imparted by adiponectin, the higher circulating levels of adiponectin present with oxidative stress and CAC (adjusting for waist and insulin resistance) may indicate an enhanced adiponectin secretory response of adipose tissue to the metabolic environment present in the early development of macrovascular disease. Thus, the elevated levels of adiponectin may comprise an attempt to alleviate risk for additional development and progression of macrovascular disease in an at-risk environment.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Low concentrations of adiponectin, a protein with significant homology to collagens VIII and X and complement factor C1q, have been measured in obesity (1, 2) and type 2 diabetes (3, 4) and in individuals with advanced macrovascular disease (5). Adiponectin has been shown to correlate with concentrations of circulating lipids (2), positively with high-density lipoprotein (HDL)1-cholesterol and negatively with triglycerides and low-density lipoprotein (LDL)-cholesterol (2). In addition, concentrations of adiponectin at baseline in an intervention trial have been inversely related to a positive outcome (6). Together, these and other data suggest a beneficial effect of adiponectin on the development and progression of macrovascular disease. The Coronary Artery Risk Development in Young Adults (CARDIA) cohort constitutes a unique group of healthy young subjects (with weight gain comparable to that of U.S. adults in their age range) (7) to assess the relationships of serum adiponectin concentrations with these and other risk factors for macrovascular disease. Previously, we had demonstrated interactions of concentrations of adiponectin and waist circumference in modulating their relationships with measures of insulin resistance (8). In this study, we extend those analyses to the correlations of circulating concentrations of adiponectin with coronary artery calcification (CAC), an early marker of macrovascular disease, and several different factors reflecting mechanisms reported to underlie the risk of macrovascular disease; e.g., aspects of lifestyle, blood pressure (BP), circulating lipid concentrations, inflammation, and oxidative damage. We determine whether and how levels of circulating adiponectin may correlate independently with any or all of these multiple pathophysiologic systems.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References


From 1985 to 1986, CARDIA recruited 5115 black and white men and women, 18 to 30 years old, from four clinical centers in Birmingham, AL, Chicago, IL, Minneapolis, MN, and Oakland, CA, with follow-up examinations completed 2, 5, 7, 10, and 15 years later. At the Year 15 examination, 3672 persons were reexamined, constituting nearly 74% of the original cohort.

Evaluated for this report were 3355 participants in whom serum adiponectin was measured at the Year 15 examination. Of these, 217 were eliminated due to a non-fasting specimen (<8 hours fast). An additional 21 were eliminated due to high creatinine, three due to missing measurements of HDL-cholesterol, triglycerides, or C-reactive protein (CRP), and 69 because they were taking medication for diabetes. Thus, the sample of 3045 people had complete data on age, race, gender, center, education, anthropometrics, adiponectin, lipids, CRP, and lipid-lowering medications. Contrasting the 3045 evaluated in this report with the 2070 (of 5115) who did not participate, we found those participating to be slightly older (25.1 vs. 24.5 years), more likely to be women (56% vs. 53% women), and less likely to be African-American (46% vs. 60% black). In addition, the current participants were less likely to be current smokers at baseline (26% vs. 36%) but more likely to have more education (14.0 vs. 13.4 years). We also analyzed F2-isoprostane concentrations and CAC; measurements of each were available in only 2545 of the 3045 participants. After accounting for occasional other missing data, 2483 participants remained with complete data on all relevant variables. Findings for all variables other than F2-isoprostanes and CAC were similar in the set of 3045 and the set of 2483; for simplicity, we report data from the set of 2483. An additional 22 of the 2483 participants did not have LDL-cholesterol levels calculated due to levels of fasting triglycerides > 4.52 mM. Of the 2483 subjects, 2334 had normal fasting glucose levels (<5.6 mM), 125 had impaired fasting glucose (5.6 to 6.9 mM), and 24 had untreated diabetes (≥7.0 mM).


Information on race, gender, and age was provided by self-report with interviews; measurements of anthropometric values were obtained as previously described (8). Smoking in CARDIA has been captured at all examinations, with categories including: currently smoking (and numbers of cigarettes per day), previously smoking, and never having smoked. The total quantity of alcohol consumed per day (in milliliters) was self-reported at each examination based on reported consumption of wine, beer, and liquor per week, multiplying by the estimated concentration of ethanol and summing the volumes. BP was measured three times after a 5-minute rest, with 1-minute intervals between repeat measures, using a Hawksley random zero sphygmomanometer (W.A. Baum Co., Copaigue, NY) and averaging the second and third measurements. Hypertension was defined according to the guidelines set by the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (9): systolic BP (SBP) ≥ 140 mm Hg or diastolic BP (DBP) ≥ 90 mm Hg or taking antihypertensive medication. Blood samples were collected, stored, and shipped as previously described (8).

CAC Measurements

CAC was measured in Oakland, CA and Chicago, IL by electron beam computerized tomography (GE Healthcare, Chalfont St. Giles, United Kingdom) and in Birmingham, AL and Minneapolis, MN by multidetector computerized tomography (GE Healthcare). The protocol included procedures to sustain consistency of measurements and standardization among and within the four institutions, as previously described (10). Briefly, Agatston scores were summed across all arteries to obtain the total calcium score for each scan, with an overall score as a mean of two scans. Approximately 100 scan sets showed calcium in one scan but not the other; these were reviewed side-by-side, after which somewhat fewer than one-half were judged indicative of no calcification. For scan sets with no calcification detected, the overall score of the scan set was recorded as zero.

Adiponectin, Lipids, CRP, and F2-Isoprostanes

Concentrations of total cholesterol, HDL-cholesterol, and triglycerides in plasma have been measured in CARDIA at the Northwest Lipid Research Center, University of Washington, by enzymatic procedures certified by the National Heart, Lung, and Blood Institute-Centers for Disease Control and Prevention Lipid Standardization Program (11). LDL-cholesterol was estimated by the Friedewald equation (12). CRP was measured at the University of Vermont by a high-sensitivity immunoassay (13).

Adiponectin was measured in serum by radioimmunoassay at LINCO, Inc. (St. Charles, MO), with an interassay coefficient of variation of 7% to 9% (2, 8). Fasting glucose and insulin were also measured at LINCO (8). The latter two measurements permitted us to calculate the homeostasis model assessment (HOMA) (14), an estimate of insulin resistance (8): (insulin glucose)/22.5, with insulin expressed as milliunits per liter and glucose as millimolar. F2-isoprostanes (stereoisomers) were assayed by a highly sensitive and specific gas chromatography-mass spectrometry method with an interassay coefficient of variation of 10%.

Statistical Methods

Both minimally adjusted and multivariate analyses were conducted on these cross-sectional data. Because concentrations of adiponectin were skewed toward higher levels, we analyzed natural logarithm transformed adiponectin [ln(adiponectin)] and report geometric mean concentrations of adiponectin. Triglycerides and CRP also displayed highly right-skewed distributions; therefore, geometric means were obtained by exponentiating log-transformed values of these measures. CRP was expressed as ln(CRP + 1) to produce positive numbers because CRP concentrations were often <1 mg/L. Previously (8), we demonstrated that waist circumference related more closely to concentrations of adiponectin than did BMI; therefore, all calculations to examine relationships with adiposity relate to waist circumference at Year 15. Linear regression was used to determine the associations of several independent variables (e.g., concentrations of HDL-C, CRP, and F2-isoprostanes), with adiponectin as the dependent variable, first adjusting for gender, race, age, and examination center. Slopes indicate approximately the percentage change in adiponectin per the indicated unit (SD for continuous variables or category for discrete variables). Given the previously demonstrated (8) correlations of adiponectin with HOMA and with waist circumference, we also added waist and lnHOMA to the regression models. Thus, we could examine the relationships of adiponectin concentrations with and without adjustment for amounts of visceral fat and/or levels of insulin resistance. All analyses were completed using the SAS statistical package version 8.2 (SAS Inc., Cary, NC).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Descriptive characteristics are given according to race and gender in Table 1. Blacks, particularly women, tended to have higher BMI values than whites. Blacks had higher BP than whites. White men had the most adverse lipid profile. HOMA was higher in blacks than in whites. Blacks were more likely to smoke than whites, although heavier smoking occurred at similar rates across race-gender groups. CRP concentrations were higher in women than in men. The men tended to drink more alcohol. The women had higher concentrations of F2-isoprostanes.

Table 1. . Descriptive characteristics according to race and gender
 Black men (453)Black women (627)White men (682)White women (721)
NMean, median, or %SD or (25th to 75th percentiles)Mean, median, or %SD or (25th to 75th percentiles)Mean, median, or %SD or (25th to 75th percentiles)Mean, median, or %SD or (25th to 75th percentiles)
  1. SD, standard deviation; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; HOMA, homeostasis model assessment; CRP, C-reactive protein; CAC, coronary artery calcification. Most values are given as number, mean, and SD. For skewed variables, values are given as median (25th to 75th percentiles). Other values are stated as percentages.

Age (years)394404413413
BMI (kg/m2)285327284276
Waist (cm)9213901494118114
SBP (mm Hg)11715116161131210712
DBP (mm Hg)7812761275107010
Alcohol (mL/d)17356271527815
Cigarette smoking        
 Never smoked (%)58 66 65 57 
 Former smokers (%)12 14 20 27 
 Current < 15/d (%)22 14 6 9 
 Current ≥15/d (%)8 7 10 7 
Cholesterol (mM)4.810.944.680.814.940.984.730.81
HDL-cholesterol (mM)1.240.351.410.341.120.311.460.38
Triglycerides (mM)0.97(0.66 to 1.46)0.79(0.60 to 1.05)1.19(0.81 to 1.77)0.87(0.63 to 1.27)
LDL-cholesterol (mM)3.030.872.850.763.120.852.790.78
HOMA (mU/mL)× (mM)3.5(2.7 to 4.8)3.8(2.9 to 5.2)3.5(2.7 to 4.5)3.0(2.5 to 3.9)
CRP (mg/L)1.1(0.8 to 1.7)1.7(1.1 to 3.4)1.0(0.8 to 1.4)1.2(0.8 to 2.0)
F2-isoprostanes (pM)137551929714459191118
CAC (%)11 4 17 5 
Adiponectin (mg/L)6(4 to 9)9(6 to 13)8(6 to 12)14(10 to 19)

Associations of Several Factors with Circulating Levels of Adiponectin

We performed a series of linear regressions (one for each row and each of two models presented in Table 2), with ln(adiponectin) as the dependent variable, first (Model 1) adjusted only for gender, race, and center and then (Model 2) for gender, race, center, lnHOMA, and waist circumference. Data are presented only for those associations that were significant within Model 2. Thus, associations between ln(adiponectin) and DBP and SBP, use of antihypertensive medications, use of lipid-lowering medications, and LDL-cholesterol were not significant with Model 2. Total cholesterol had a minimal relationship with adiponectin (Table 2, Model 2, p = 0.05). Relationships of adiponectin were positive with HDL-cholesterol and alcohol consumption and negative with ln(triglycerides), ln(CRP), and smoking (at higher consumption, ≥15 cigarettes/d). With Model 1, circulating levels of adiponectin were not associated with either circulating F2-isoprostanes or CAC. However, with additional adjustment for waist and lnHOMA, adiponectin was positively associated with F2-isoprostanes and CAC (Model 2, Table 2).

Table 2. . Several linear regression analyses of ln(adiponectin) in serum (milligrams per liter) as the dependent variable with cardiovascular risk factors
  Model 1: adjusted for age, race, gender, and centerModel 2: adjusted for age, race, gender, center, waist circumference, and lnHOMA
 Comparison unit or SDβSEpβSEp
  1. SD, standard deviation; HOMA, homeostasis model assessment; SE, standard error; HDL, high-density lipoprotein; CRP, C-reactive protein; CAC, coronary artery calcification. Risk factors listed in the table had significant p values in model 2. β is approximately the percentage change in adiponectin per SD or indicated comparison unit of each independent (row) variable.

Total cholesterol (mM)0.89−0.0160.0110.140.0190.0100.05
HDL-cholesterol (mM)0.370.2510.011<0.00010.1720.011<0.0001
ln(triglycerides) (mM)0.553−0.1900.011<0.0001−0.0950.011<0.0001
ln(CRP + 1) (mg/L)0.473−0.1220.011<0.0001−0.0320.0110.005
Alcohol (mL/d)26.10.0430.011<0.00010.0230.0100.02
Smokers (≥15/d)vs. All others−0.0900.0410.03−0.0830.0360.02
F2-isoprostanes (pM)930.0010.0120.960.0730.011<0.0001

Associations of Combinations of Factors with Circulating Levels of Adiponectin

The next linear regression analysis (Table 3) used ln(adiponectin) as the dependent variable, adjusted for gender, race, center, waist, and lnHOMA, and studied combinations of the factors previously found in Model 2 of Table 2 to have p ≤ 0.05. Cholesterol lost significance; it was confounded with HDL-cholesterol and ln(triglycerides) (data not shown). The positive association with alcohol consumption was eliminated; separate analyses showed that HDL-cholesterol was the variable with which alcohol intake was confounded (data not shown). Some circulating fasting lipids [HDL-cholesterol and ln(triglycerides)] and a measure of inflammation [ln(CRP + 1)] all remained significant independent variables (Table 3). HDL-cholesterol was positively related to adiponectin, whereas ln(CRP + 1) and ln(triglycerides) correlated negatively. Adiponectin remained positively associated with both F2-isoprostanes and CAC. Because these findings were contrary to expectation (given that both are risk factors for macrovascular disease, whereas adiponectin is inversely related to that risk), we explored them further.

Table 3. . Multiple linear regression model with ln(adiponectin) in serum (milligrams per liter) as the dependent variable
  1. SE, standard error; HDL, high-density lipoprotein; CRP, C-reactive protein; CAC, coronary artery calcification. All variables are contained in a single regression model with ln(adiponectin) as the dependent variable. The independent variables are listed with the comparison unit or standard deviation shown in parentheses.

HDL-cholesterol (0.37 mM)0.1430.013<0.0001
Alcohol (26.1 mL/d)0.0020.0100.85
Smokers (≥15/d vs. all others)−0.0640.0350.06
Total cholesterol (0.89 mM)0.0110.0110.31
ln(Triglycerides) [0.553 ln(mM)]−0.0540.013<0.0001
ln(CRP + 1) [0.473 ln(mg/L)]−0.0320.0110.003
F2-isoprostanes (93 pM)0.0570.010<0.0001
CAC (any vs. none)0.0830.0330.01

Adiponectin and F2-Isoprostanes

In a separate analysis incorporating quintiles of F2-isoprostane concentrations as independent variables, the fully adjusted (as in Table 3) estimated geometric mean adiponectin concentrations increased monotonically from 8.5 mg/L in the lowest quintile, through 8.6, 8.7, and 9.4 mg/L, to 9.9 mg/L in the highest quintile. We then performed a series of analyses to estimate the impact of waist and/or HOMA on the relationship between adiponectin and F2-isoprostanes (Table 4). From essentially a β of 0 with adjustment for age, sex, race, and center (row 1), the β increased whenever waist circumference was included in the model, with a lesser effect of adjustment for HOMA (Table 4). Similar effects of waist were seen in a model with multiple covariates (Table 4).

Table 4. . Multiple linear regression models with several levels of adjustment
 F2-isoprostane concentrations (pM)CAC (present, absent)
  • CAC, coronary artery calcification; SE, standard error; HOMA, homeostasis model assessment. ln(Adiponectin) concentration in serum (milligrams per liter) is the dependent variable, with F2-isoprostane concentrations and CAC as the independent variables. In the first four rows, adjustment was for age, race, gender, and center (designated as covariate set A); then waist circumference and/or lnHOMA were added to each model. In the last four rows, the remaining variables in Table 3 were added as covariates (designated as covariate set B). Thus, the last row (multiple and waist and lnHOMA) has identical βs to those listed in Table 3.

  • *

    Minimal variable adjustment: age, race, gender, and center.

  • Multiple variable adjustment: age, race, gender, center, total cholesterol, HDL-cholesterol, alcohol, smoking ≥ 15 cigarettes/d, DBP, ln(triglycerides), and ln(CRP + 1).

Covariate set A      
 Minimal and waist0.0730.011<0.00010.0540.0360.14
 Minimal and lnHOMA0.0460.011<0.00010.0730.0350.04
 Minimal and waist and lnHOMA0.0710.011<0.00010.0710.0350.04
Covariate set B      
 Multiple and waist0.0570.011<0.00010.0730.0340.03
 Multiple and lnHOMA0.0450.010<0.00010.0860.0330.01
 Multiple and waist and lnHOMA0.0570.010<0.00010.0830.0330.01

Adiponectin and CAC

In all subjects adjusted for age, gender, race, and center, concentrations of adiponectin were slightly, but not significantly, higher in those with CAC compared with those without (data not shown). Progressive adjustment for waist and then lnHOMA resulted in a mean ln(adiponectin) that was 0.079 higher among those with CAC than among those without CAC (p = 0.02, Model 2, Table 2). Further adjustment for several other factors (Table 3) confirmed this difference to 0.081 (p = 0.01, Table 4). The fully adjusted estimated geometric mean adiponectin levels were 8.9 mg/L in those without CAC and 9.7 mg/L in those with CAC. Again, through a series of additional analyses, we showed HOMA to be the predominant confounder affecting the positive correlation of adiponectin and CAC (Table 4).

Variance Explained by Models

Although waist- and HOMA-adjusted F2-isoprostane concentrations and CAC had associations with concentrations of adiponectin that were counterintuitive, other factors explained considerably more of the variance in adiponectin than did F2-isoprostane concentrations and CAC. The r2 values for the models in Table 4 increased from r2 = 0.240 in the model with age, sex, race, and center to 0.456 when waist, lnHOMA, and all other covariates except F2-isoprostane concentrations and CAC were added. Further addition of both F2-isoprostane concentrations and CAC to this full model increased the r2 value to 0.464.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

The causal relationships (i.e., the apparent protection against macrovascular disease) of adiponectin to other circulating compounds, to medications, and to other factors remain to be determined by prospective or interventional studies, rather than by cross-sectional studies such as this. Nevertheless, to try to understand these biological/biochemical interactions in a cross-sectional manner, we have examined factors that may correlate with levels of adiponectin (e.g., medications, smoking, and intake of alcohol). Also, we have explored the correlations of adiponectin with concentrations of circulating lipids, CRP, and F2-isoprostanes and with CAC. Previously, we demonstrated strong correlations of adiponectin with waist circumference and HOMA, a measure of insulin resistance (8). Thus, the amount of visceral adipose tissue may affect the secretion of adiponectin and its potential role in increasing insulin sensitivity (2, 4, 15), which may, in turn, lead to some of the associations shown here, as reflected by adjustments for waist circumference and HOMA.

Consistent with the apparent protective effect of adiponectin against macrovascular disease and independent of central adiposity and/or insulin resistance, elevated circulating adiponectin was related to increased levels of HDL-cholesterol and reduced levels of triglycerides and LDL-cholesterol. These findings were similar to the results with lipids by Cnop et al. (2). Elevated circulating adiponectin was also related to lower levels of CRP. Without adjustment for waist and/or HOMA, it was unrelated to F2-isoprostanes or CAC. However, conversely, and not obviously consistent with apparent protection, levels of adiponectin adjusted for waist and HOMA were positively related to circulating levels of F2-isoprostanes, a marker of oxidative damage that we have shown to be positively associated with CAC. Furthermore, waist- and HOMA-adjusted adiponectin levels related positively to the presence of CAC itself, but less strongly than to F2-isoprostanes. Although the increment in variance of adiponectin explained by F2-isoprostanes and CAC was not large, it is important to emphasize that we did not expect either F2-isoprostanes or CAC to have a positive relation with concentrations of adiponectin.

Therefore, the positive associations of fasting levels of adiponectin with CAC and F2-isoprostanes contrast with most other associations of adiponectin noted in this paper. They may be explained by the secretion of adiponectin in response to factors that lead to CAC, one of which may be increasing levels of oxidative damage (i.e., with F2-isoprostanes as a marker of oxidative damage). Thus, increasing circulating levels of adiponectin may reduce the burden of these factors, including oxidative damage, subsequently limiting the progression of CAC to more advanced macrovascular disease. However, it is important to note the lack of a complete protective effect of adiponectin; i.e., after adjustment for waist and HOMA, the subjects had CAC and higher levels of F2-isoprostanes. It is important to note that adiponectin exists in several forms, each possibly with different arrays of activities (16, 17). The immunoassay used herein has not been characterized as to the form(s) of adiponectin measured; thus, inferences cannot be made relating structural variants of adiponectin to the activities and associations found herein.

Alternatively, adiponectin may influence the concentrations of F2-isoprostanes by enhancing insulin action and thereby increasing the uptake of glucose and the flow of its metabolites through mitochondrial pathways, generating more free radicals, leading to higher levels of F2-isoprostanes. This possible mechanism is suggested by several types of studies. First, it is well-known that hyperglycemia results in elevated levels of oxidative stress and an elevated formation of F2-isoprostanes (18), presumably due to an increased flux of substrate through mitochondrial pathways, presumably under the effective influence of insulin. From another perspective, insulin resistance is associated with a decrease in the uptake of glucose and oxidative phosphorylation, which would result in a decrease of oxidative stress. Thus, the modulation of insulin sensitivity and glucose concentrations can influence oxidative stress.

The effect of enhanced levels of F2-isoprostanes may be detrimental or beneficial; longitudinal experiments are needed to interpret their effect. Prior associations of circulating F2-isoprostane concentrations with CAC would suggest a pathway in which increased F2-isoprostanes lead toward disease progression (10). If the concentrations of F2-isoprostanes reflect an increasing risk of oxidative damage disrupting cells (e.g., endothelial cells, macrophages, and other cells) or causing the early manifestations of macrovascular disease (i.e., CAC), then the corresponding rise in adiponectin reflects its capacity to reduce damage that would otherwise lead to an enhanced the risk of macrovascular disease (19). In this case, F2-isoprostanes would stimulate the production of adiponectin, and this model would reflect the observations of adiponectin playing a putative positive role in reducing vascular damage, by leading to an increased clearance of oxidized-LDL and stimulation of reverse cholesterol transport (20, 21, 22). The mechanisms behind this seemingly paradoxical relationship, where high adiponectin is generally associated with favorable risk factor levels, may reflect a specific property of isoprostanes, which may become depleted in response to increased vascular permeability with injury but are elevated when oxidative damage occurs (e.g., when LDL oxidation is present in the lesion).

The mechanisms postulated above are likely to be quite complex, with several different factors seen to influence the levels of circulating adiponectin (visceral fat, insulin sensitivity, lipids, inflammation, measures of oxidative damage, and CAC itself). Thus, many of the originally simple associations of factors or circulating compounds with adiponectin were altered (either eliminated or added to the model) with the addition of waist and lnHOMA. In this complicated model, there remain the possibilities of both stimulatory and inhibitory influences on the secretion (and possibly clearance) of adiponectin, with subsequent change in the levels of the cardiovascular risk factors, including some lipids and a measure of oxidative damage. Overall, these observations support the postulated role of adiponectin in altering the influences of other factors related to increased vascular disease risk. Sequential measurements would provide much more information on the temporal relationships among these various factors, especially with the progression of macrovascular disease in the CARDIA population.

  • 1

    Nonstandard abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; CARDIA, Coronary Artery Risk Development in Young Adults; CAC, coronary artery calcification; BP, blood pressure; CRP, C-reactive protein; SBP, systolic BP; DBP, diastolic BP; HOMA, homeostasis model assessment; ln, natural logarithm.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References
  • 1
    Arita, Y., Kihara, S., Ouchi, N., et al (1999) Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochem Biophys Res Commun. 257: 7983.
  • 2
    Cnop, M., Havel, P. J., Utzschneider, K. M., et al (2003) Relationship of adiponectin to body fat distribution, insulin sensitivity and plasma lipoproteins: evidence for independent roles of age and sex. Diabetologia. 46: 459469.
  • 3
    Hotta, K., Funahashi, T., Arita, Y., et al (2000) Plasma concentrations of a novel, adipose-specific protein, adiponectin, in type 2 diabetic patients. Arterioscler Thromb Vasc Biol. 20: 15951599.
  • 4
    Lindsay, R. S., Funahashi, T., Hanson, R. L., et al (2002) Adiponectin and development of type 2 diabetes in the Pima Indian population. Lancet. 360: 5758.
  • 5
    Zoccali, C., Mallamaci, F., Tripepi, G., et al (2002) Adiponectin, metabolic risk factors, and cardiovascular events among patients with end-stage renal disease. J Am Soc Nephrol. 13: 134141.
  • 6
    Pischon, T., Girman, C. J., Hotamisligil, G. S., Rifai, N., Hu, F. B., Rimm, EB. (2004) Plasma adiponectin levels and risk of myocardial infarction in men. JAMA. 291: 17301737.
  • 7
    Lewis, C. E., Jacobs, D. R. Jr, McCreath, H., et al (2000) Weight gain continues in the 1990s: 10-year trends in weight and overweight from the CARDIA study: Coronary Artery Risk Development in Young Adults. Am J Epidemiol. 151: 11721181.
  • 8
    Steffes, M. W., Gross, M. D., Schreiner, P. J., et al (2004) Serum adiponectin in young adults: interactions with central adiposity, circulating levels of glucose and insulin resistance, the CARDIA Study. Ann Epidemiol. 14: 492498.
  • 9
    The sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med. 157: 24132446.; (1997).
  • 10
    Gross, M., Steffes, M., Jacobs, D. R. Jr, et al (2005) Plasma F2-isoprostanes and coronary artery calcification: the CARDIA Study. Clin Chem. 51: 125131.
  • 11
    Marcovina, S. M., Gaur, V. P., Albers, JJ. (1994) Biological variability of cholesterol, triglyceride, low- and high-density lipoprotein cholesterol, lipoprotein(a), and apolipoproteins A-I and B. Clin Chem. 40: 574578.
  • 12
    Friedewald, W. T., Levy, R. I., Fredrickson, DS. (1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 18: 499502.
  • 13
    Tracy, R. P., Lemaitre, R. N., Psaty, B. M., et al (1997) Relationship of C-reactive protein to risk of cardiovascular disease in the elderly: results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol. 17: 11211127.
  • 14
    Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., Turner, RC. (1985) Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 28: 412419.
  • 15
    Weyer, C., Funahashi, T., Tanaka, S., et al (2001) Hypoadiponectinemia in obesity and type 2 diabetes: close association with insulin resistance and hyperinsulinemia. J Clin Endocrinol Metab. 86: 19301935.
  • 16
    Tsao, T. S., Murrey, H. E., Hug, C., Lee, D. H., Lodish, HF. (2002) Oligomerization state-dependent activation of NF-kappa B signaling pathway by adipocyte complement-related protein of 30 kDa (Acrp30). J Biol Chem. 277: 2935929362.
  • 17
    Tsao, T. S., Tomas, E., Murrey, H. E., et al (2003) Role of disulfide bonds in Acrp30/adiponectin structure and signaling specificity: different oligomers activate different signal transduction pathways. J Biol Chem. 278: 5081050817.
  • 18
    Gopaul, N. K., Anggard, E. E., Mallet, A. I., Betteridge, D. J., Wolff, S. P., Nourooz-Zadeh, J. (1995) Plasma 8-epi-PGF2 alpha levels are elevated in individuals with non-insulin dependent diabetes mellitus. FEBS Lett. 368: 225229.
  • 19
    Matsuda, M., Shimomura, I., Sata, M., et al (2002) Role of adiponectin in preventing vascular stenosis: the missing link of adipo-vascular axis. J Biol Chem. 277: 3748737491.
  • 20
    Rader, DJ. (2003) Regulation of reverse cholesterol transport and clinical implications. Am J Cardiol. 92: 429J.
  • 21
    Schmitz, G., Kaminski, WE. (2001) ABC transporters and cholesterol metabolism. Front Biosci. 6: D505D514.
  • 22
    Miller, M., Rhyne, J., Hamlette, S., Birnbaum, J., Rodriguez, A. (2003) Genetics of HDL regulation in humans. Curr Opin Lipidol. 14: 273279.