The study protocol was performed according to the Declaration of Helsinki and was approved by the local ethics committee of Yuksek Ihtisas Training and Research Hospital, Ankara, Turkey and all the patients provided written informed consent before entering the study.

A post hoc survival analysis of the patients involved in a single-center prospective trial investigating the association between serum sclerostin levels and arteriovenous fistula calcification and patency, was performed [Balcı M, et al. Herz 2015;40:289–97]. The study was done at the RFM Dialysis Center, Ankara, Turkey between August 2011 and July 2013. The exclusion criteria in that work were vascular access other than mature native forearm arteriovenous fistula, a previous history of intervention like percutaneous angioplasty and/or stenting, a mature forearm arteriovenous fistula unable to provide spKt/V>1.2 in a 4-h dialysis session, active infection, active malignancy, active autoimmune disease, recent hospitalization or unstable hemodynamic status in the preceding 3 months, and patient refusal.

At August 1st 2011, 350 patients (155 women and 195 men, mean age 55±10 years, mean hemodialysis vintage: 56±23 months, on hemodialysis thrice a week for at least 6 months) had all work-up finished for the eligibility and allocation. The patients suffered from end stage renal disease due to diabetic nephropathy (n=124), hypertensive nephrosclerosis (n=102), chronic glomerulonephritis (n=77), chronic pyelonephritis (n=12) and polycystic disease (n=15). The renal diagnosis was unknown in 20 patients. Presence of residual renal function (RRF) was defined as residual glomerular filtration rate ≥1mL/min. Mean residual renal function was 0.7±1.1mL/min/1.73m2.

All patients were ≥18 years of age and were receiving conventional 4-h hemodialysis with synthetic polysulphone dialysers F6HPS and F7HPS (Fresenius AG, Bad Homburg, Germany) thrice a week, with bicarbonate dialysate, and low-molecular-weight heparin for standard anticoagulation. Mean blood flow rate was 300mL/min during the HD session (range 250–340mL/min). Dialysate fluid composition was sodium 140mEq/L, potassium 2–3mEq/L, calcium 3mEq/L and bicarbonate 33mEq/L. Targeted dialysis efficiency was based on single pool Kt/V urea nitrogen (spKt/V) values according to the National Kidney Foundation-Kidney Disease Outcomes Quality Initiative Guidelines.

Angiotensin-converting enzyme inhibitors (n=44), angiotensin receptor blockers (n=66), beta-blockers (n=82) and calcium channel blockers (n=76) were given for hypertension management. Patients were prescribed treatments including CaCO3 (32%), sevelamer-HCl (19%), Ca acetate (41%), calcitriol (69%), anti-platelet agents (70%), warfarin (5%) and erythropoietin (68%). The mean erythropoietin dose was 145U/kg/week achieving a mean hemoglobin (Hb) serum level of 11.2g/dL; <10% of patients had serum Hb <10g/dL. None of the patients received calcimimetic, glucocorticoid, statin, bisphosphonates or denosumab.

Information regarding baseline demographic characteristics, etiology of end stage renal disease, presence of diabetes mellitus were collected by reviewing medical records. Cardiovascular history was defined as history of myocardial infarction, percutaneous coronary artery intervention, cardiac by-pass or valvular surgery, peripheral artery disease or stroke.

Venous blood samples were drawn after an overnight fast. The blood sample was obtained directly through an arteriovenous fistula on a mid-week non-dialysis day. Biochemical serum parameters (creatinine, blood urea nitrogen, glucose, electrolytes, albumin and complete blood count calcium, phosphate, lipids, protein, cholesterol, and triglycerides) were performed by standard laboratory procedure using an automated analyzer. Serum C-reactive protein (CRP) level was detected by rate nephelometry (IMAGE). spKt/V value was calculated according to the Daugirdas second-generation formula.9 Normalized protein catabolic rate (nPCR) was calculated as a measure of daily protein intake of patients.10

The primary outcome was 24-months, all-cause mortality. The primary exposure variable was the baseline serum sclerostin level, measured at the initiation of outpatient hemodialysis. Serum sclerostin level was measured by a commercially available enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions and as used in previous studies by Cejka et al.7,11 The intra-assay and inter-assay coefficients of variation were 7.5% and 6.3%, respectively. Patients were classified in 3 groups according to value of baseline circulating sclerostin levels, corresponding to Group 1 (36–586pg/ml), Group 2 (587–1903pg/ml) and Group 3 (1904–8822pg/ml). Serum intact parathormone and alkaline phosphatase levels were studied by means of a computerized autoanalyser (Hitachi 717; Boehringer-Mannheim). Serum was available for radioimmunoassay of 25-dihydroxy vitamin D3 levels (DiaSorin Corporation, Stillwater, MN) in all patients.

A non-enhanced computed tomography (CT) was performed to measure to arteriovenous fistula calcium score on a 64-detector CT scanner (Aquilion; Toshiba Medical Systems, Tokyo, Japan) as a deputy marker of vascular calcification. Foci of arteriovenous fistula calcification was detected by an experienced radiologist and scored using semi-automatic commercial software (Vitrea 2, Vital images) by detection of at least 3 contiguous pixels (voxel size: 1.03mm3) of peak density ≥130Hounsfiels Units (HU). The lesion score was calculated with the area density method, by multiplying the lesion area by a density factor derived from the maximal HU within the area as described by Agatston.12 The density factor was assigned in the following manner: 1 for lesions whose maximal density was 130–199HU, 2 for lesions 200–299HU, 3 for lesions 300–399HU, and 4 for lesions >400HU. Based on these calculations, patients were divided into 2 groups as having non-calcified and calcified arteriovenous fistulas.

Normally distributed continuous variables are presented as mean±standard deviation, and non-normally distributed continuous variables are presented as median with 25th and 75th percentiles. Categorical data are presented as percentages. To test for associations between serum sclerostin levels and demographic, clinical and laboratory parameters, we examined serum sclerostin levels in quartiles according to the distribution of values in HD patients. Differences in clinical and biochemical parameters were compared using analysis of variance for continuous variables and chi-square test for categorical variables, whereas non-parametric Kruskal–Wallis test was used for non-normally distributed continuous variables. Spearman correlation coefficients were obtained for all potential predictor variables to look for confounding. Multivariate statistical analysis (logistic regression, enter method) was used to account for possible confounders for arteriovenous fistula calcification. Kaplan–Meier survival analysis was used to calculate cumulative incidence of the all-cause mortality. Variables that affected all-cause mortality (p<0.2) in univariate analysis were then included in multivariate Cox proportional hazards models to determine risk factors associated with all-cause mortality. Test results were presented as hazard ratios with 95% confidence intervals, and a p-value of <0.05 was considered to be statistically significant. Data analysis was performed by using SPSS for Windows, version 15.0 (SPSS Inc., Chicago, IL, USA).

Mean serum sclerostin level in hemodialysis patients was higher when compared to age- and gender-matched healthy controls (1519±1378 vs 128±49pg/ml, p<0.0001). Sclerostin levels were not significantly different in patients with a positive history of hypertension (1520±1479 vs 1506±1480pg/ml, p=0.870), dyslipidemia (1486±1432 vs 1541±1477, p=0.576) and smoking (1599±1356 vs 1534±1410, p=0.690) compared with those without. However, male patients had significantly higher serum sclerostin levels than female patients (1736±1109 vs 1212±975pg/ml, p=0.025). Moreover, patients with a positive cardiovascular history had higher serum levels than patients with not (2008±1635 vs 1431±1307pg/ml, p=0.018). Furthermore, patients with type 2 diabetes mellitus (DM) tended to have higher sclerostin levels than patients with not (1678±1476 vs 1404±1290pg/ml, p=0.071). No interaction between the use of calcitriol and serum sclerostin levels was found.

No significant correlations were observed between serum sclerostin levels and age, hemodialysis vintage, systolic and diastolic blood pressure, LDL- and HDL-cholesterol, triglycerides, hemoglobin, serum albumin, calcium, phosphate, alkaline phosphatase, uric acid and CRP levels. However, sclerostin levels were negatively correlated with log serum intact parathormone levels (r=−0.222, p=0.005).

The demographic and clinical characteristics of the study patients according to the sclerostin tertiles are depicted in Table 1. Patients in the highest quartile had lower albumin levels and log intact parathormone (Table 1), and higher frequency of arteriovenous fistula calcification on computerized tomography and cardiovascular history (Table 1). No significant difference was noted for any other demographic, clinical and laboratory variables (Table 1).

Patients with arteriovenous fistula calcification (Fig. 1) had higher sclerostin levels than patients without (2564±1841 vs 1063±706pg/ml, p<0.0001). According to univariate logistic regression analysis, diabetes mellitus, cardiovascular history, treatment with calcium-based phosphate binders and sclerostin levels were significant predictors for arteriovenous fistula calcification (Table 2). Increased age tended to influence arteriovenous fistula calcification among study patients. The multivariate logistic regression analysis corrected for calcitriol use revealed that presence of positive cardiovascular history, treatment with calcium-based phosphate binders and lower but not higher serum sclerostin levels predicted arteriovenous fistula calcification independently (Table 2).

Fig. 1.

Computed tomographic scan showing calcification of the arteriovenous fistula (AVF) (a) and a non-calcified AVF (b).

(0.11MB).

Overall, 84 of 350 patients (24%) died during the 24-months follow-up. When compared, patients who died were older (55.9±13 vs 55.9±13 years, p=0.022), had lower serum albumin (3.71±0.11 vs 3.9±0.34g/L, p=0.030) and 25-hydroxy vitamin D3 (11.2±5 vs 15.8±6ng/mL, p=0.010) and had higher serum alkaline phosphatase levels (132±28 vs 84±34U/L, p=0.03). Patients who died also had a tendency to have lower hemoglobin (11.20±1.20 vs 10.90±0.96g/dL, p=0.069) and log serum intact parathormone (2.01±0.11 vs 2.27±0.38, p=0.080). Gender, hemodialysis vintage, systolic and diastolic blood pressure, spKt/V, serum levels of calcium, phosphate, C-reactive protein, uric acid, LDL-cholesterol, HDL-cholesterol, triglycerides, cardiovascular history, presence of diabetes mellitus and treatment with calcium-based phosphate binders were not statistically different. However, patients who had arteriovenous fistula calcification did have worse survival than patients without (χ2=7.15; p=0.007, log-rank test, Fig. 2).

Fig. 2.

Kaplan–Meier curve depicting patient survival according to tertiles of serum sclerostin levels.

(0.18MB).

Patients who died had elevated serum sclerostin levels at baseline compared to patients who survived at the end of the 24-months follow-up (2143±1327 vs 1469±1373pg/ml; p=0.017). Kaplan–Meier curve shows that study patients with increasing serum sclerostin levels at baseline, had a worse survival (χ2=6.668; p=0.036, log-rank test, Fig. 3). Using the Cox proportional hazard model with multicovariate adjustments, age, albumin, and presence of arteriovenous fistula calcification were found to be associated with survival in maintenance hemodialysis patients (Table 3). Results were already corrected for calcitriol use and were unaffected when serum sclerostin level was considered as a continuous variable.

Fig. 3.

Kaplan–Meier curve depicting patient survival according to presence of arteriovenous fistula calcification.

(0.17MB).