We showed that increased urinary adiponectin concentration is significantly associated with microalbuminuria and increased mean baPWV. Our results indicate that urinary adiponectin may be a useful biomarker for detecting vascular damage such as microalbuminuria and arterial stiffness.
Several earlier studies focused on the relationship between plasma adiponectin and renal complications in diabetic patients [6, 16]. Few researchers, however, have investigated the association between urinary adiponectin level and vascular damage. One study  assessed whether urinary adiponectin excretion might represent vascular damage in type 2 diabetic patients. One hundred fifty-six diabetic patients with a past history of diabetic nephropathy and 40 healthy subjects were evaluated in that study. The investigators found no significant relationship between urinary adiponectin and 24-hour urinary albumin excretion rate, but the urinary adiponectin level was significantly higher in diabetic patients than in healthy subjects. Furthermore, these researchers reported that adiponectinuria was associated with common carotid artery intima-media-thickness in the diabetic patients. In a study of diabetic patients, Koshimura et al.  showed that the urinary adiponectin excretion level was markedly higher in the group with overt nephropathy than in the group without nephropathy. However, these authors did not find a significant association with urinary adiponectin level in patients with normoalbuminuria and microalbuminuria, only among those with macroalbumi nuria (>300 mg/g). However, their study population comprised only 38 males with type 2 diabetes.
We found that urinary adiponectin concentration was significantly associated with the urinary A/C ratio and that there was a significant difference in urinary adiponectin concentration between subjects with normoalbuminuria and microalbuminuria. The correlation between urinary adiponectin and urinary A/C ratio was significant after adjusting for serum adiponectin level.
The molecular weight of an adiponectin monomer is approximately 28 kDa, while a dimer is 56 kDa and a trimer is 68 kDa . Adiponectin circulates predominantly as a trimer (LMW), hexamer (MMW), or high molecular weight multimer (HMW). Shen et al.  demonstrated that urinary adiponectin is physiologically intact and mainly present as a trimer (LMW). The molecular weight of the trimeric form of adiponectin is similar to the molecular weight of albumin, which is about 67 kDa. Some reports have suggested that vascular complications precede the onset of albuminuria or the detection of diabetes [18, 19]. Therefore, a measure of urinary total adiponectin in the monomeric to trimeric forms could be a useful tool for early detection of vascular changes due to its size, which is smaller than that of albumin. If it were possible to compare the results of renal biopsy specimens according to the grade of adiponectinuria or albuminuria, we could obtain more information on comparison between markers for the detection of diabetic nephropathy. Future studies will be required to clarify the pathologic evidence.
Furthermore, we provided evidence that measuring the urinary adiponectin level is more effective than measuring the serum adiponectin level in order to detect microalbuminuria. Measurement of urinary biomarkers is more useful than that of serum biomarkers, because urine can be sampled simply and non-invasively .
The mechanism through which plasma or urinary adiponectin excretion is related to diabetic nephropathy or other vascular complications is unclear. Adiponectin has been shown in previous studies to protect against vascular damage. Some reports, including an experimental study, revealed that adiponectin-deficient (Ad
) mice treated with adiponectin showed reduced albuminuria and improvement in podocyte function [20, 21]. Sharma et al.  demonstrated that plasma adiponectin levels had a negative correlation with the degree of albuminuria in non-diabetic subjects. They suggested that plasma adiponectin is a regulator of albuminuria and modulates oxidant stress in podocyte foot processes. Shimotomai et al.  reported that increased urinary adiponectin level may be explained by altered glomerular permeability. Consistent with these hypotheses, we found that urinary adiponectin concentration positively correlated with the urinary A/C ratio, and that serum adiponectin concentration negatively correlated with the urinary A/C ratio. In our study, we observed that the mean serum adiponectin concentration increased with increasing urinary adiponectin tertiles. We could not determine the precise mechanism behind this observation. We observed that the mean urinary adiponectin concentration was higher in diabetic patients than in the normal glucose status group. Also, we found that the mean serum adiponectin concentration was lower in diabetic patients than in the normal glucose status group. If the mechanism were due to ‘spill-over’, then the serum adiponectin level would be similar to or higher in the diabetic group than in the normal glucose status group, in order to compensate for the renal loss of adiponectin. We speculate that it may be another disease pattern of the results; that is, serum adiponectin may not have compensated fully as urinary adiponectin level decreased.
We also found that increased urinary adiponectin level was associated with degree of arterial stiffness. Many researchers have assessed the association between plasma adiponectin level and arterial stiffness or blood pressure [7, 23]. Some studies have found that decreased plasma adiponectin concentration was related to stiffness of both the central and peripheral arteries in hypertensive patients [8, 23]. However, another study found no correlation between plasma adiponectin level and peripheral arterial stiffness . In our study, we used brachial-ankle pulse wave velocity (baPWV) as a simple measure of arterial stiffness and as a predictive marker of macrovascular complications . We found that urinary adiponectin concentration, but not serum adiponectin concentration, was an independent predictor of baPWV as a marker of arterial stiffness after adjusting for confounding factors such as age, lipid profile, and systolic blood pressure [26, 27]. Recently, a few studies have appeared reporting that circulating adiponectin did not show a cardioprotective effect in animal models . In humans, serum adiponectin concentrations rise across chronic heart failure stages, although this finding is less evident in type 2 diabetic patients according to the study by Baldasseroni et al. . In our results, we did not find evidence of a protective role for serum adiponectin against vascular damage, in accord with these reports.
Our study has several strengths compared with previous urinary adiponectin studies. We performed our investigation in a heterogeneous population that included healthy normoglycemic subjects, prediabetic subjects, and type 2 diabetic patients. Additionally, we recruited a much larger study population than did previous studies. We excluded those subjects with macroalbuminuria and focused our assessment on the relationship between urinary adiponectin concentration and microalbuminuria.
Several limitations of our study must be considered when interpreting our results. First, our study was cross-sectional. Our findings are therefore limited to documenting causal relationship. Second, we measured total adiponectin concentration, and not that of individual isoforms. Some reports have suggested that high molecular weight adiponectin is a more powerful marker of insulin resistance and arterial stiffness than is low molecular weight adiponectin [8, 30]. This may be the reason why we failed to find a significant relationship between serum adiponectin level and arterial stiffness in our study. Follow-up investigations are needed to examine the potential associations between adiponectin isoforms and vascular complications. Thirdly, we used baPWV as a representative tool for detecting arterial stiffness instead of aortic PWV, which is considered to be a better marker. Further studies are needed to validate the significance of baPWV in a larger prospective design.