Our study was designed to examine the relationship of adiponectin with total coronary atherosclerotic plaque burden and atherosclerotic plaque morphology in humans. We demonstrate that 1) adiponectin levels are predictive of total coronary plaque burden, and 2) adiponectin levels are inversely correlated with the number of mixed and non-calcified plaques. No significant correlation was seen with calcified plaques. Our data suggest that the number of mixed and non-calcified plaques decrease by 1 for each 11.5 µg/ml and 13.2 µg/ml increase in serum adiponectin levels, respectively.
Pundziute et al. 
reported that the number of coronary segments with mixed plaques as determined by MSCT was an independent predictor of acute cardiac events including cardiac death, nonfatal myocardial infarction and unstable angina requiring hospitalization. Hoffmann et al. 
demonstrated that in contrast to calcified plaques, non-calcified plaques as determined by MSCT were consistently present in culprit lesions of patients with acute coronary syndrome. We have recently shown that rapid progression of angiographic stenosis severity occurred most frequently in coronary segments with non-calcified or predominantly non-calcified plaques as determined by MSCT 
. Wolk et al. 
reported that higher plasma adiponectin levels were associated with a lower risk of acute coronary syndrome, suggesting that the pathophysiological role of adiponectin may be related to the stability of atherosclerotic plaque rather than atherosclerotic burden. Our data indicate that adiponectin may reduce the number of non-calcified and mixed plaques without affecting stable calcified plaques. Consistent with these observations, Marso et al. 
showed in a very recent study that low adiponectin levels were associated with lipid-rich yet not calcified coronary plaques in non-diabetic patients in bivariate analysis as determined by IVUS. Furthermore, Otake et al. 
demonstrated that the necrotic core component ratio of culprit lesions as determined by IVUS in patients with acute coronary syndrome was negatively correlated with adiponectin levels in bivariate analysis. Therefore, the finding of an independent association between adiponectin and non-calcified and mixed coronary plaques may have important therapeutic and preventive implications for decreasing the risk of acute coronary events.
Although the exact mechanisms of the negative association between adiponectin and mixed and non-calcified plaques remain to be determined, several possible explanations should be considered. In addition to its beneficial effects on insulin sensitivity and lipid metabolism 
, adiponectin exerts its vasculoprotective effects through its direct actions on endothelial cells, monocytes, macrophages and other inflammatory cells, platelets and smooth muscle cells, thus modulating initiation and progression of atherosclerosis. Adiponectin augments endothelial NO production 
, inhibits ox-LDL–induced endothelial ROS generation 
, suppresses the expression of endothelial adhesion molecules 
, attenuates leucocyte-endothelium interactions 
and protects endothelial cells from apoptosis 
. Furthermore, adiponectin inhibits macrophage activation 
and foam cell formation 
, promotes the clearance of early apoptotic cells by macrophages 
, inhibits smooth muscle cell proliferation 
and antagonizes the stimulatory effect of TNF-α on vascular smooth muscle cell calcification 
The present study has several strengths and limitations. Our data support the concept that adiponectin is an important marker in the pathogenesis of atherosclerosis since adiponectin remains significantly associated with plaque morphology in a fully adjusted multivariate model containing age, sex, BMI, hypertension, diabetes mellitus, smoking, family history of CAD, LDL-C, HDL-C, triglycerides, hsCRP, medication and PAT volume. However, since this was an association study, our study does not establish a causal relationship between adiponectin and coronary plaque morphology. Furthermore, the clinical implications of our results still need to be determined.
There are conflicting epidemiological data regarding the role of adiponectin in atherosclerosis with some studies showing strong inverse associations between adiponectin levels and CAD 
and others failing to detect any association 
. Differences in study design, population characteristics and statistical adjustments may in part explain the apparent discrepancies. In addition, different methods of CAD assessment (CT-angiography or IVUS versus conventional angiography or CT-based assessment of coronary artery calcification) may yield distinct results. Adiponectin may modulate early stages of atherogenesis thus mainly affecting the development of non-calcified or mixed plaques.
In the present study adiponectin is not correlated with calcified plaques, the majority of coronary plaques in our study cohort. Therefore, although statistically highly significant, adiponectin levels are associated with a very modest decrease in total coronary plaque burden accounting for only 3% of the variability in total number of coronary plaques. In contrast, adiponectin accounts for approximately 20% of the variability in mixed and non-calcified plaque burden. Determination of the bioactive high molecular weight form of adiponectin may have possibly resulted in an even stronger association of adiponectin with plaque burden and morphology. Overall, our data emphasize the importance of additional risk factors in the pathogenesis of atherosclerosis.
Storage conditions of blood samples (e.g. storage time and temperature) were suggested to alter adiponectin levels 
. However, comparing adiponectin concentrations in serum aliquots stored for 21 and 33 months at −70°C, adiponectin levels were not significantly different (p
0.45) suggesting that storage conditions did not impact on our results due to adequate specimen stability.
Finally, our study results may only be applicable to non-diabetic patients. Diabetics were clearly underrepresented in the present study (7% of patients). The small number of diabetic patients may also explain the lack of association between adiponectin levels and diabetes.
In summary, we demonstrate that adiponectin is inversely associated with coronary plaque burden and mixed and non-calcified plaques, suggesting an important role of adiponectin in the pathogenesis of atherosclerosis and possibly in the pathophysiology of acute coronary syndrome.