Serum Levels of the Adipokine Chemerin in Relation to Renal Function

The design of the study has recently been described in detail (9,,–12). Briefly, 120 Caucasian men (n = 62) and women (n = 58) were recruited with 60 patients having a glomerular filtration rate (GFR) >50 ml/min (control patients), as assessed by the original Modification of Diet in Renal Disease formula (13), and 60 patients being on hemodialysis. Thirty control patients and 32 chronic hemodialysis (CD) patients had type 2 diabetes. Patients with active inflammatory diseases including pneumonia, urinary tract infection, endocarditis, sinusitis, and cholangitis were excluded from the study. Furthermore, patients with end-stage malignant diseases of any origin were excluded. Inactive systemic lupus erythematodes, stable coronary heart disease, and previous stroke were not exclusion criteria. The study was approved by the local ethics committee, and all subjects gave written informed consent before taking part in the study.

Blood samples were taken after an overnight fast. In CD patients, blood was drawn just before hemodialysis started. Chemerin (BioVendor, Modrice, Czech Republic) (intraassay coefficient of variation [CV] 5.1–7.0%, interassay CV 6.9–8.3%), adiponectin (Mediagnost, Reutlingen, Germany) (intraassay CV <4.7%, interassay CV <6.7%), and leptin (Mediagnost, Reutlingen, Germany) (intraassay and interassay CV <10%) were determined with ELISAs according to the manufacturers' instructions. Free fatty acids, cholesterol, triglycerides, C-reactive protein (CRP), insulin, and other routine laboratory parameters were measured in a certified laboratory.

SPSS software version 15.0 (SPSS, Chicago, IL) was used for all statistical analyses as further specified in the results section and in the legend for Table 1. Distribution was tested for normality using Shapiro-Wilk W test, and non-normally distributed parameters were logarithmically transformed before multivariate analyses.

Table 1 summarizes clinical characteristics of the subgroups studied (control and CD). In Table 1 and throughout the text, all continuous variables are given as median ± interquartile range. Median circulating chemerin was more than twofold higher in CD patients (542.2 ± 98.1 μg/l) compared with control patients (254.3 ± 88.7 μg/l, P < 0.001) (Table 1). In contrast, a significant difference in chemerin concentrations could not be demonstrated depending on sex (female subjects 324.4 ± 284.6 μg/l and male subjects 443.6 ± 315.2 μg/l) and type 2 diabetes (type 2 diabetes 388.1 ± 303.7 μg/l and non-type 2 diabetes 331.0 ± 274.6 μg/l). CD patients had a significantly lower BMI compared with that in control patients (P < 0.05) (Table 1).

Using the Spearman rank correlation method, serum chemerin concentrations positively correlated with BMI (r = 0.398, P = 0.002), fasting insulin (FI) (r = 0.408, P = 0.001), leptin (r = 0.516, P < 0.001), and CRP (r = 0.256, P = 0.049) in control patients. In addition, chemerin negatively correlated with GFR (r = −0.372, P = 0.003) in control patients. In CD patients, circulating chemerin levels were negatively associated with GFR (r = −0.413, P = 0.001).

Multiple linear regression analysis revealed that GFR (logarithmically transformed [log], standardized β-coefficient = −0.337, P = 0.013) but not FI (log, standardized β-coefficient = 0.186, P = 0.128), leptin (log, standardized β-coefficient = 0.091, P = 0.588), and CRP (log, standardized β-coefficient = 0.138, P = 0.236) remained independently associated with circulating chemerin (log) levels in control patients after adjustment for age (standardized β-coefficient = −0.033, P = 0.793) and sex (standardized β-coefficient = 0.250, P = 0.097). A similar result was obtained when BMI instead of leptin was included in the model (data not shown). In addition, GFR (log, standardized β-coefficient = −0.351, P = 0.007) predicted circulating chemerin (log) independent of age (standardized β-coefficient = −0.223, P = 0.072) and sex (standardized β-coefficient = −0.076, P = 0.546) in CD patients.

In the current study, we show for the first time that circulating chemerin levels are more than twofold higher in CD patients compared with control patients. Furthermore, CD is a strong independent predictor of chemerin concentrations in multivariate analysis (data not shown). Moreover, GFR remains independently associated with circulating chemerin in multivariate analysis in both control patients and CD patients. In these cases, functional studies including urine analyses should be performed to define whether renal elimination influences serum levels of chemerin. Furthermore, renal production of chemerin has been shown (1,,–4), and it should be determined to what extent this kidney-derived chemerin contributes to circulating levels of the adipokine in control patients and CD patients. Moreover, because chemerin modulates inflammation (14,15), its contribution to renal disease–associated metabolic and vascular complications should be elucidated in future studies.

Recently, an association of chemerin serum levels with metabolic syndrome–related parameters including BMI (5,–7), FI (7), triglycerides (6,–8), HDL cholesterol (5,,–8), leptin (5,6), and CRP (5,6) has been shown. In agreement with these findings, chemerin is positively correlated with BMI, FI, leptin, and CRP in univariate analyses in the control patients in our study. However, these associations in control patients are all lost in multivariate analyses after controlling for renal function, whereas GFR remains independently associated with circulating chemerin. Interestingly, GFR also independently predicts chemerin serum levels in the CD patients in our study. These results indicate that renal function is a significant predictor of circulating chemerin not only in subjects with (near) normal glomerular filtration but also in patients with end-stage renal disease.

Some limitations of the study have to be pointed out: First, the study has a cross-sectional design, and, therefore, causality cannot be established. Second, the sample size is relatively small, and it is possible that various nonsignificant associations in multivariate analyses would have become statistically significant if larger samples were studied. Third, differential misclassification of covariates such as type 2 diabetes is possible because type 2 diabetes was only excluded in the control patients but not in the CD patients by 75-g oral glucose tolerance tests due to the necessary fluid restriction in the latter group.

Taken together, our results suggest that renal filtration independently predicts circulating chemerin.

This study was supported by a grant to M.F. from the Deutsche Forschungsgemeinschaft (DFG), KFO 152, Atherobesity, project FA476/4–1 (TP 4).

No potential conflicts of interest relevant to this article were reported.