It is known that plasma levels of AD are able to predict diabetes onset and diabetic patients show lower concentrations of the adipokine compared to the general population [4, 6]. Our study extends these observations to a specific subgroup of DM patients with ischemic chronic heart failure, exploring the controversial relationship that exists between AD and HF.
This study confirms our previous observation that AD increases in patients with CAD and overt HF compared to those with normal or reduced LVEF but clinically asymptomatic . This is in accordance to the literature which reports higher AD levels in patients with advanced NYHA class . We now show that not only clinical staging, but also a more pathophysiological classification of HF as proposed by ACC/AHA guidelines, is able to predict AD levels in this subset of patients. NT-proBNP, HDL cholesterol and age still remain the strongest, independent predictors of AD .
We have shown that AD levels are systematically lower in diabetics compared to non-diabetics, no matter to what HF staging class they belong. These data suggest that diabetic subjects with reduced LVEF orchestrate a slower rise of AD which begins only in face of overt HF (group CD). It is of interest that Chen and coll. in asymptomatic men with uncomplicated type 2 diabetes found a positive association of circulating AD levels with myocardial glucose metabolism but not with cardiac function .
Because of its cross-sectional nature, our study cannot elucidate with clarity the pathophysiological mechanisms that lie under the increase of AD at worsening HF. In any case, some considerations can arise from the data. First of all, cardiac cachexia cannot be the main cause of higher AD levels in HF because none of our study subjects was underweight and the increase of AD observed between A/B and C was not coupled to a decrease of BMI. Therefore, cardiac cachexia is a late process in HF compared to the increase of AD which seems to be a faster phenomenon.
For the same reason, we cannot offer a definite explanation for the hampered rise of AD that we detected in HF diabetic subjects. However, we can speculate that diabetics have lower AD plasma levels to begin with which may depend from their higher BMI. Indeed, in our backward stepwise linear regression models, we found that DM loses its predictivity on AD when BMI was introduced as cofactor. However, if we focus on patients with overt HF (CND vs. CD) we find that despite a significant difference in AD levels, BMI was not different (data not shown). These data suggest that something else, beyond body weight, is able to modulate adiponectin metabolism in diabetic patients with HF. A good candidate could be NT-proBNP which seems to directly induce AD production by adipocytes, in vitro, and in heart failure patients . Even more striking, natriuretic peptides are able to increase total and HMW-adiponectin concentrations in healthy subjects . It could be assumed that diabetic patients produce lower levels of NT-proBNP in response of overt HF. However, this is not the case: in fact the level of NT-proBNP is not significantly different between diabetics and non-diabetics belonging to the A and B clinical stages. Even more, diabetic subjects in the C group have higher, although not significant, levels of NT-proBNP when compared to the non-diabetic C subgroup, in spite of a lower adiponectinemia (data not shown). A question remains as to what lies between the secretion of NT-proBNP and the production of AD by adipocytes. Some sort of “NT-proBNP resistance” may exist in type 2 DM. Further studies will be required to investigate this hypothesis, comparing the different AD response of diabetic and non-diabetic subjects to injection of natriuretic peptides.
Finally, Won and coll. found that plasma concentration of AD was significantly lower in HF patients with metabolic syndrome compared to HF patients without metabolic syndrome . In our series, the prevalence of metabolic syndrome, as defined by ATPIII criteria [33, 34], was 68.2% in diabetic patients and 30.9% in non-diabetic patients. At multivariate analysis, the presence of metabolic syndrome was not an independent predictor of AD plasma levels (data not shown).
Our population comprises only ischemic HF and type 2 DM. Heart failure of different etiology (valvular, hypertensive and idiopathic) was not explored. Similarly, we have no data regarding type 1 DM. Another limitation of the study is the small number of female subjects (only 13%) that were included which depended on the nature of the inclusion criteria (CAD). Finally, as stated above, all limitations of cross sectional studies also apply to our findings.