Serum irisin levels correlated to peritoneal dialysis adequacy in nondiabetic peritoneal dialysis patients

Background Irisin is a recently discovered myokine thought to be involved in multiple metabolism abnormalities in most dialysis patients. However, the myokine has not been thoroughly studied in peritoneal dialysis. This study aimed to evaluate serum irisin levels and establish their relation to dialysis adequacy, insulin resistance, and bone metabolism status in patients on peritoneal dialysis. Methods A total of 59 nondiabetic prevalent peritoneal dialysis patients and 52 age- and sex-matched healthy controls were enrolled in this cross-sectional study. Serum irisin concentration was assessed by enzyme-linked immunosorbent assay. The correlations between serum irisin and dialysis adequacy, clinical, and metabolic variables were investigated. Results Serum irisin levels were lower in nondiabetic peritoneal dialysis patients (17.02ng/ml) compared with healthy controls (22.17ng/ml, P<0.001). Multivariate regression analysis revealed that fasting glucose levels were correlated inversely with serum irisin levels in peritoneal dialysis patients. Serum irisin levels were associated with neither insulin resistance nor bone metabolism in our patients. Serum irisin levels were positively associated with peritoneal Kt/Vurea (β = 4.933, 95% confidence interval [CI] = 0.536–9.331, P = 0.029) and peritoneal CCr (β = 0.259, 95% CI = 0.053–0.465, P = 0.015) among peritoneal dialysis patients. Conclusions The study demonstrated that non-diabetic peritoneal dialysis patients have lower serum irisin levels, and the levels were correlated with peritoneal dialysis adequacy, indicating adequate dialysis may improve irisin secretion. Additional studies are needed to provide a confirmation.

Introduction syndrome, hemodialysis, or bariatric surgery were excluded. Double bag systems and PD solutions (Dianeal 1.5% or 2.5% dextrose; Baxter Healthcare, Guangzhou, China) were used in all patients. CAPD involved a 2-L fluid exchange 3-5 times per day. After completing detailed clinical questionnaires and routine urine and blood tests to exclude CKD and the conditions described above, 52 age-and sex-matched healthy individuals were also included as controls. A unique identification number was provided for each participant and others had no access to information that could identify individual participants during or after data collection.
When assessing the relationship between irisin and CKD-MBD, the following additional exclusion criteria were applied: biliary obstructive disease, Cushing's syndrome, primary hyperparathyroidism, hypoparathyroidism, hyperthyroidism, rheumatologic diseases, consumption of aluminum-containing drugs, teriparatide and denosumab, paraplegia, genetic diseases including Marfan syndrome, prior musculoskeletal injuries or surgical procedures within 3 months prior to the study, or severe acidosis and acute pancreatitis. Collectively, 59 PD patients were enrolled in the study. The study protocol was approved by the ethics committee at the third Hospital of Sun Yat-Sen University. Written informed consent was provided by all participants.

Analytic procedures and biochemical analyses
Blood samples and data collection took place between March and August 2016. Blood samples were collected in vacutainer tubes without anticoagulant after fasting for at least 8h. Following centrifugation, serum was collected and stored with aprotintin at -80˚C until analysis. Medical history, demographic data, and routine biochemical parameters, including levels of triglycerides [TGs], total cholesterol [TC], HDL cholesterol [HDL-C], LDL cholesterol [LDL-C], blood urea nitrogen [BUN], serum creatinine, and albumin [ALB], were also obtained. Experimental data were obtained using the Hitachi 7180 biochemistry analyzer (Japan), and insulin levels were measured by chemiluminescence (ADVIA Centaur XP: Siemens). The estimated glomerular filtration rate was calculated based on the Chronic Kidney Disease Epidemiology Collaboration creatinine equation [19] and the modified Modification of Diet in Renal Disease formula for Chinese patients [20]. The Geriatric nutritional risk index (GNRI) was calculated using the following formula: GNRI = [1.489 × ALB (g/L)] + [41.7 × (body weight/ideal body weight)]. Ideal body weight was calculated from the Lorentz equations (WLo) [21]. Insulin resistance was estimated using the homeostatic model assessment of insulin resistance (HOMA-IR) equation: HOMA-IR = fasting serum insulin (mU/L) × fasting serum glucose (mmol/L) / 22.5 [12]. Additionally, 24-h urine and dialysate were collected to calculate weekly kidney and peritoneal Kt/V urea and weekly kidney and peritoneal creatinine clearance (kidney C Cr and peritoneal C Cr , respectively). Serum irisin levels were measured by ELISA using commercial kits (EK-067-29, Phoenix Pharmaceutical, Burlingame, CA, USA) and the intra-and inter-assay variations were <10% and <15%, respectively. The experiments were conducted on April 27, June 12 and August 27, 2016, respectively.

Statistical analysis
SPSS Software, version 13.0 was used for all statistical analyses. The results were presented as the means ± standard deviation or medians with interquartile ranges according to the sample distributions. Categorical variables were given as frequencies (percentages). The Student's ttest was used to compare differences between the two groups for normally distributed data, while the Mann-Whitney U test was used for non-normal data. Categorical data were compared using the Chi-square test. Correlations were expressed as Pearson's correlation coefficients for two continuous variables, and Spearman rank correlations were used for non-normally distributed variables. Univariate linear regression analyses were performed to evaluate the determinants of serum irisin levels. Multiple stepwise linear regression models were employed to select variables independently related to serum irisin concentrations. All analyses were two-tailed, and a p<0.05 was considered statistically significant.

Baseline characteristics
The demographic and clinical characteristics of the 59 PD patients and 52 healthy controls are summarized in Table 1. The mean ages of the PD patients and healthy controls were 47.5±11.6 years and 44.3±15.0 years, respectively. Male patients and controls accounted for 57.6% and 57.7% of individuals in each group, respectively. No differences in age, sex, proportion of smokers or drinkers, or TC, TG or HDL-C levels were observed between the two groups. PD patients had a lower BMI and serum levels of fasting glucose, ALB, and LDL-C and higher blood pressure and pulse pressures, levels of BUN, serum creatinine and uric acid, utilization ratios of angiotensinconverting enzyme inhibitors/angiotensin II receptor blockers and statins compared with those in healthy controls (all p< 0.05). PD patients had a lower GNRI than that of healthy controls, and serum irisin levels were significantly lower in the nondiabetic PD group (17.02 (11.27-20.09) ng/ml) compared with the control group (22.17 (17.00-26.57) ng/ml, P<0.001) (Fig 1).

Correlations between serum irisin levels and other study parameters
Univariate correlations revealed no significant correlation between serum irisin levels and age, BMI, blood pressure, BUN, serum fasting glucose, creatinine, ALB, cholesterol, TGs, uric acid, GNRI or HOMA-IR in PD patients ( Table 2).

Relationship between irisin levels and bone turnover markers
Intact parathyroid hormone (iPTH) and total alkaline phosphatase (tALP) were bone turnover markers evaluated for CKD-MBD and neither correlated with serum irisin levels. Similarly, no correlations were observed between serum irisin levels and serum levels of corrected calcium, phosphate, or calcium-phosphorus products (Table 4). These results applied to male and female patients individually, as well as the overall PD group. Thus, the association between irisin levels and bone metabolism was not present in PD patients.

Discussion
In this cross-sectional study, we revealed significantly lower serum irisin levels in nondiabetic PD patients compared with age-and sex-matched healthy controls. Multivariate regression  analyses also revealed that peritoneal Kt/V urea and peritoneal C Cr were positively correlated with serum irisin levels, suggesting that adequate dialysis may improve irisin secretion. Moreover, fasting glucose levels were inversely correlated with serum irisin levels in PD patients; however, no association between serum irisin levels and HOMA-IR was observed. No association was observed between serum irisin levels and GNRI in PD patients.
Irisin was first identified in 2012 and thought to be involved in energy regulation [1]. Since energy imbalance is prevalent in ESRD, evaluation of circulating irisin levels in ESRD patients helped elucidate the mechanisms of energy dysfunction in ESRD. In the present study, serum irisin levels were significantly lower in nondiabetic PD patients compared with age-and sexmatched healthy controls. These results are in agreement with previous studies showing lower irisin levels in CKD patients, including hemodialysis patients [4][5][6][7]. Recently, Lee et al. and Rodriguez et al. also illustrated that patients undergoing peritoneal dialysis tend to have lower plasma irisin levels compared with healthy controls [8,9]. While diabetic patients were included in the PD groups in their studies, none was included in the control group. Clinical studies have shown diabetic patients have lower irisin levels than those of subjects with normal glucose tolerance or nondiabetic controls [10,11]. Kaluzna et al. recently also reported irisin levels were lower in ESRD patients compared to controls. However the study had not analyzed the correlation between irisin levels and dialysis adequacy [22]. Since their study included only seven PD patients, it is hard to make the conclusion in PD patients. To address these questions, our study especially included nondiabetic PD patients and controls; therefore, as the first study to identify a correlation between serum irisin and adequacy of peritoneal dialysis. Since irisin is a myokine excreted by muscle, sarcopenia, which is common among ESRD patients, may partially account for this finding. Mid-arm muscle circumference was thought to be an independent predictor of serum irisin levels in PD patients [8]; however, studies have shown that no correlation between irisin levels and anthropometric parameters among the CKD population exist [5,7]. Discrepancies among study populations, the methods used for assessment of lean body mass and the assay kits used to detect irisin may have contributed to the conflicting conclusions. It is worth noting that indoxyl sulfate, a uremic toxin, can affect the production of FNDC5 and irisin in human skeletal muscle cells [4]. Recently, irisin was found to be secreted by adipose tissue as well, and therefore irisin is an adipo-myokine [23]. To determine whether the expression and secretion of FNDC5/irisin are down-regulated in CKD patients, both subcutaneous adipose tissue and skeletal muscle samples should be assessed. Moreover, the activities of BAT should also be determined in CKD patients to assess whether they are in line with circulating irisin levels.
In this study, peritoneal Kt/V urea or peritoneal C Cr were identified as parameters independently associated with elevated serum irisin levels in PD patients, which suggest that irisin is associated with dialysis adequacy. Lee et al. suggested that a higher peritoneal Kt/V urea may result in lower irisin levels, which is different from the results of this study. However, the association between irisin and RRF in that study was determined without adjusting for peritoneal dialysis doses [8]. It should be noted that both RRF and peritoneal dialysis may contribute to the metabolism of irisin. In this study, we addressed the question by performing multivariate Table 2. Univariate correlations with circulation irisin levels.

Correlation with irisin (r) P-value Correlation with irisin (r) P-value
Age ( Table 4. Associations of serum irisin levels and bone turnover markers. regression analyses and found that the more uremic toxins removed by peritoneal dialysis, the higher the circulating irisin levels will be. Our results also suggest that uremic toxins decrease the secretion of irisin in skeletal muscle in a dose-dependent manner [4], and the clearance of uremic toxins may enhance the production of FNDC5/irisin. Conversely, it has been reported that irisin levels are correlated with insulin resistance among type 2 diabetes mellitus patients [24,25], and our results also revealed that irisin levels were inversely correlated with fasting blood glucose levels. As insulin resistance is frequently seen in CKD patients, peritoneal dialysis may partially correct this phenomenon [26]. Thus, we suggest that adequate dialysis may improve insulin resistance and increase the levels of circulating irisin. Unfortunately, the circulating irisin levels in PD patients did not correlate with HOMA-IR in our study. Since diabetes patients were excluded from this study, the severity of insulin resistance in our PD patients was minimal.

Correlation with irisin (r) P-value Correlation with irisin (r) P-value Correlation with irisin (r) P-value
No association between serum irisin levels and GNRI was observed in this study. As an objective nutritional evaluation score, GNRI predicts the risks of nutrition-related diseases [27]. Our results suggest that serum irisin levels may not indicate PEW in PD patients. To our knowledge, this was the first study to investigate the association between irisin and GNRI. A previous study showed that lower serum irisin levels were associated with sarcopenia in PD patients [8]; however, the different study populations may account for this discrepancy. Importantly, the results of this study should be interpreted with caution, since GNRI is merely a risk index. Machined-based assessments, including dual energy X-ray absorptiometry and body impedance analysis, are validated reference methods for nutritional evaluations and should be used in future studies [28].
There is controversy concerning the association between circulating irisin and fasting glucose levels. While Timmons et al. revealed no association in a diabetic population [29], Huh et al. found a positive association among healthy women [30]. Our study showed circulating irisin levels were inversely associated with fasting glucose levels in PD patients, and the differences may be explained by the different study populations evaluated. In a study by Kurdiova, glucose decreased the in vitro expression of Fndc5 in myotubes in which the levels of Fndc5 mRNA were lower in adipose tissue and plasma in type 2 diabetes patients versus pre-diabetic individuals [31]. As irisin is the cleavage product of FNDC5, the levels of irisin may be downregulated when Fndc5 expression is reduced.
Increasing evidence suggests that irisin is a vital link between skeletal muscle and bone formation. Previous results were based mainly on postmenopausal osteoporosis [32]. As renal osteodystrophy is characterized by bone remodeling disorders, we explored the association between the expression of bone turnover markers and serum irisin levels in PD patients. No associations were observed, and it should be noted that CKD-MBD is more complicated than osteoporosis. Both tALP and iPTH are bone turnover markers and not completely agreeable with histomorphometric analyses. Thus, bone biopsies are ideal; however, they are invasive and not practical.
The strength of this study was that diabetic patients were excluded from both groups. This is also the first study to explore the relationship between peritoneal dialysis adequacy and irisin after adjusting for RRF. However, this study also has several limitations. This was a cross-sectional, observational study and thus could not provide causal relationships for the findings. The study had a small sample size, and there were confounding factors that were not addressed. Moreover, anthropometry parameters and bone histomorphometric analyses from biopsies were unavailable, and therefore, those correlative analyses were not performed.
Furthermore, the ELISA kits used in our study have been found to reveal lower irisin levels as compared with some other kits in investigations by Choi  In conclusion, this study revealed that circulating irisin levels were lower in nondiabetic PD patients compared with healthy controls, and that peritoneal Kt/V urea and creatinine clearance were positively correlated with serum irisin levels in nondiabetic PD patients. These findings are contrary to previous viewpoints that irisin is partially dialyzable. No associations between irisin and HOMA-IR and GNRI were observed, and no association between bone turnover markers and irisin was found among PD patients.