Clinical and Experimental Nephrology

, Volume 16, Issue 3, pp 456–463

Serum cystatin C as a predictor for cardiovascular events in end-stage renal disease patients at the initiation of dialysis

Authors

  • Min Ji Shin
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
    • Biomedical Research InstitutePusan National University Hospital
  • Ihm Soo Kwak
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
    • Biomedical Research InstitutePusan National University Hospital
  • Soo Bong Lee
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
  • Dong Won Lee
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
  • Eun Young Seong
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
    • Biomedical Research InstitutePusan National University Hospital
  • Il Young Kim
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
  • Harin Rhee
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
    • Biomedical Research InstitutePusan National University Hospital
  • Naria Lee
    • Division of Nephrology, Department of Internal MedicinePusan National University Hospital School of Medicine
Original Article

DOI: 10.1007/s10157-011-0583-1

Cite this article as:
Shin, M.J., Song, S.H., Kwak, I.S. et al. Clin Exp Nephrol (2012) 16: 456. doi:10.1007/s10157-011-0583-1

Abstract

Background

There has been no study to investigate whether cystatin C could predict cardiovascular events in incident dialysis patients. We aimed to delineate the role of serum cystatin C and cystatin C-based estimated glomerular filtration rate (eGFRcysC) for prediction of cardiovascular events.

Methods

This study included 66 end-stage renal disease patients who survived for >3 months after the start of dialysis, and serum cystain C levels were measured at the point of dialysis initiation.

Results

Serum cystatin C was correlated with blood urea nitrogen (r = 0.537, p < 0.001), serum creatinine (r = 0.480, p < 0.001) and smoking (r = 0.284, p = 0.021). Cystatin C was inversely correlated with age (r = −0.316, p = 0.01) and eGFRCr by modification of diet in renal disease (r = −0.533, p < 0.001). Kaplan–Meier analysis for cardiovascular events revealed that patients in the group with lower cystatin C levels (<4.14 mg/L) had a better event-free survival rate than patients in the group with higher cystatin C levels (≥4.14 mg/L) (p = 0.039). In univariate analysis, cystatin C (hazard ratio (HR) 2.62, p = 0.011) and eGFRcysC (HR 0.64, p = 0.004) were significant factors for the prediction of cardiovascular events. After multivariate adjustment, serum cystatin C and eGFRcysC were independent determinants of cardiovascular events (HR 3.952, p = 0.001 and HR 0.640, p = 0.004, respectively).

Conclusion

Serum cystatin C might be an independent marker of cardiovascular events in incident dialysis patients. Furthermore, eGFRcysC based on measured serum cystatin C could have a new role in predicting cardiovascular events beyond the estimation of true GFR.

Keywords

Cardiovascular diseaseCystatin CEnd-stage renal diseaseGlomerular filtration rate

Introduction

Cardiovascular disease (CVD) is the leading cause of death in patients with end-stage renal disease (ESRD) [1]. Based upon these data, it has been considered that proper management of CVD is very important to improve the prognosis for ESRD patients. To date, the known major risk factors for CVD are hypertension, diabetes mellitus, dyslipidemia and atherosclerosis. Arterial stiffness has been a strong independent predictor of coronary events and cardiovascular mortality in various populations and increased arterial stiffness is closely associated with atherosclerosis [24]. Recent studies have demonstrated that dialysis patients have less compliant arteries compared with control subjects matched for age and blood pressure [5]. In addition, both pulse pressure and increased aortic pulse wave velocity, manifestations of large vessel stiffness, are independent risk factors for adverse outcomes in this population [6, 7]. Furthermore, our group previously reported that serum cystatin C was related to pulse wave velocity in subjects with normal serum creatinine [8, 9]. Thus, we thought that serum cystatin C may be related to cardiovascular events irrespective of serum creatinine.

Cystatin C is a small 13-kDa protein and is produced at a constant rate by all nucleated cells in the body. This is a consequence of cystatin C being a product of a housekeeping gene with stable and continuous expression [10, 11]. As a small-sized protein, cystatin C is filtrated freely in the glomerulus, with no known extrarenal excretion or degradation. In the proximal tubule of the nephron, there is a reuptake and complete degradation of cystatin C but without any reabsorption into the bloodstream; its properties making it a candidate for a good marker of glomerular filtration rate (GFR). In prior studies in the general population and in the elderly, cystatin C has been shown to be a better predictor of mortality and adverse cardiovascular events than serum creatinine [1214]. Another study showed that serum cystatin C level was associated with all-cause and CVD mortality in stage 3 or 4 chronic kidney disease (CKD) patients [15].

However, there has been no data to investigate whether cystatin C could predict cardiovascular events in incident dialysis patients. Furthermore, recent studies argue that cystatin C-based estimated GFR (eGFRcysC) is a better predictor of CVD than creatinine-based eGFR (eGFRCr) because of the non-GFR determinants of cystatin C [16, 17]. The aim of this study is to delineate the role of serum cystatin C and eGFRcysC for predicting cardiovascular events and compare other traditional variables with serum cystatin C in ESRD patients at the initiation of dialysis.

Materials and methods

Patients

A total of 66 ESRD patients [hemodialysis (HD) 46 patients, peritoneal dialysis (PD) 20 patients] were enrolled, who survived for >3 months after the start of dialysis from February 2007 to November 2010. The inclusion criteria were patients >18 years of age with measured serum cystatin C levels at the point of dialysis initiation. Because thyroid function could affect the levels of cystatin C [18], we excluded patients with thyroid dysfunction or taking medication due to thyroid disease for over 6 months. We also excluded patients with malignancy and receiving glucocorticoid therapy because patients in treatment with steroids could have higher levels of cystatin C [19]. Aside from the related diseases and medications, patients with liver cirrhosis, active infection and chronic inflammatory disorder were also excluded. We recorded confidential information of age, gender, height, weight, social histories, medical histories and cardiovascular events. This study was approved by the Ethical Committee of Pusan National University Hospital.

Laboratory methods and eGFR calculations

Serum cystain C levels were measured at the point of dialysis initiation. The following laboratory data were obtained: cystatin C, creatinine, blood urea nitrogen (BUN), lipid profile, hemoglobin, albumin and C-reactive protein (CRP). Serum cystatin C was measured by the latex agglutination test (Modular P800; Roche Diagnostics, Mannheim, Germany). Serum creatinine was measured by the modified Jaffe reaction (Modular D; Roche Diagnostics). The eGFRCr was calculated using modification of diet in renal disease (MDRD) formula: MDRD = 186.3 × (creatinine in mg/dL)−1.154 × (age)−0.203 × (0.742 if female) [20]. The eGFRcysC was calculated by the chronic kidney disease epidemiology equation: eGFR = 127.7 × (cystatin C in mg/L)−1.17 × (age)−0.13 × (0.91 if female) [21].

Statistical analysis

Statistical analyses were performed using SPSS for window version 17.0 (SPSS Inc., Chicago, IL, USA). Mean and standard deviation were reported for normally distributed data; otherwise, median and interquartile ranges were given. Student’s t test was used for continuous variables and chi-squared test was used for categorical variables. The Pearson’s correlation coefficient was performed to test the correlations between cystatin C level and different variables. After categorizing individuals into two groups on the basis of mean cystatin C level, the Kaplan–Meier survival analysis was performed. We compared the clinical and biochemical factors with and without cardiovascular events and we then used the Cox proportional hazards models to evaluate the predicting factors for cardiovascular outcomes; variables having p value <0.1 in univariate analysis were included in multivariate analysis. All results were considered significant by p < 0.05.

Results

Baseline clinical and biochemical characteristics of ESRD patients

Mean follow-up period was 14.9 months. Mean age of ESRD patients was 52.7 ± 16.3 years (18.0–86.0), and their clinical and biochemical characteristics are shown in Table 1. The main primary causes of ESRD were diabetes mellitus (56.0%), hypertension (15.2%) and chronic glomerulonephritis (10.6%). Serum cystatin C, creatinine, and BUN levels were 4.1 ± 0.7 mg/L, 9.1 ± 4.8 mg/dL and 94.4 ± 36.5 mg/dL, respectively. eGFRcysC and eGFRCr were 14.6 ± 2.9 mL/min/1.73 m2 and 7.1 ± 2.9 mL/min/1.73 m2, respectively.
Table 1

Characteristics of incident dialysis patients (n = 66)

Variables

Dermographics

 Age (years)

52.7 ± 16.3

 Male gender, n (%)

43 (65.2)

 Dialysis modality (HD), n (%)

44 (66.6)

 BMI (kg/m2)

23.3 (20.9–25.0)

Hemodynamic conditions

 SBP

147.5 ± 19.9

 DBP

79.6 ± 13.2

Causes of ESRD

 Diabetic nephropathy, n (%)

37 (56.0)

 Hypertension, n (%)

10 (15.2)

 Chronic glomerulonephritis, n (%)

7 (10.6)

 Others, n (%)

12 (18.2)

Laboratory data

 Serum cystatin C (mg/L)

4.1 ± 0.7

 Serum creatinine (mg/dL)

9.1 ± 4.8

 Serum BUN (mg/dL)

94.4 ± 36.5

 eGFRcysC (mL/min/1.73 m2)

14.6 ± 2.9

 eGFRCr (mL/min/1.73 m2)

7.1 ± 2.9

 Hemoglobin (g/dL)

9.0 ± 1.5

 Albumin (g/dL)

3.5 ± 0.6

 Total cholesterol (mg/dL)

149.9 ± 47.6

 LDL-C (mg/dL)

91.0 ± 40.6

 HDL-C (mg/dL)

38.9 ± 14.2

 CRP (mg/dL)

0.18 (0.04–0.76)

Data presented as mean ± SD, median (interquartile range), or percent frequency, as appropriate

HD Hemodialysis, BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, ESRD end-stage renal disease, BUN blood urea nitrogen, eGFRcysC cystatin C-based estimated glomerular filtration rate, eGFRCr creatinine-based estimated glomerular filtration rate, LDL-C low-density lipoprotein cholesterol, HDL-C high-density lipoprotein cholesterol, CRP C-reactive protein

Variables associated with serum cystatin C level in ESRD patients

Cystatin C was correlated with BUN (r = 0.537, p < 0.001), creatinine (r = 0.480, p < 0.001) and smoking (r = 0.284, p = 0.021). Cystatin C was inversely correlated with age (r = −0.316, p = 0.01) and eGFRCr by MDRD (r = −0.533, p < 0.001). There were no correlations between cystatin C and body mass index (BMI) (r = 0.113, not significant [NS]), total cholesterol (r = 0.179, NS), low-density lipoprotein (LDL) cholesterol (r = 0.094, NS), high-density lipoprotein (HDL) cholesterol (r = −0.108, NS) or CRP (r = 0.410, NS) (Table 2).
Table 2

Correlation analysis of variables associated with serum cystatin C level

 

Correlation coefficient

Age

−0.316*

Smoking

0.284*

Serum BUN

0.537*

Serum creatinine

0.480*

eGFRCr

−0.533*

Gender

0.079

BMI

0.113

SBP

−0.125

DBP

−0.162

Hemoglobin

−0.119

Albumin

0.067

Total cholesterol

0.179

LDL-C

0.094

HDL-C

−0.108

CRP

0.410

BUN Blood urea nitrogen, eGFRCr creatinine-based estimated glomerular filtration rate, BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, LDL-C low-density lipoprotein cholesterol, HDL-C high-density lipoprotein cholesterol, CRP C-reactive protein

*p < 0.05

Comparison of clinical and biochemical factors with and without cardiovascular events

During the follow-up periods of 14.9 months, 11 patients (16.7%) presented new cardiovascular events (1 with myocardial infarction, 4 with heart failure and 6 with stroke) and 3 patients (4.5%) died as a result of cardiovascular causes (Table 3). In comparison of the clinical and biochemical variables according to cardiovascular events, cystatin C (4.7 ± 0.8 vs 4.0 ± 0.7, p < 0.01), total cholesterol (182.64 ± 48.1 vs 143.3 ± 45.1, p = 0.011) and LDL cholesterol (117.3 ± 45.8 vs 85.8 ± 38.0, p = 0.033) were significantly higher, and eGFRcysC (12.2 ± 2.7 vs 15.0 ± 2.8, p < 0.01) was significantly lower in patients with than without cardiovascular events. Other significant differences between the two groups were not detected for HDL cholesterol, BUN, creatinine, eGFRCr, BMI, CRP and medication (Table 4).
Table 3

Summary of cardiovascular events (n = 11, 16.7%)

 

n (%)

Myocardial infarction

1 (1.5)

Heart failure

4 (6.1)

Stroke

6 (9.1)

 Ischemic

4

 Hemorrhagic

2

Peripheral arterial occlusive disease

0 (0)

Table 4

Comparison of clinical and biochemical factors according to cardiovascular events

 

Patients with cardiovascular events (n = 11)

Patients without cardiovascular events (n = 55)

p value

Age (year)

56.62 ± 16.58

58.27 ± 14.81

NS

BMI (kg/m2)

22.0 (21.0–24.3)

23.4 (20.9–25.3)

NS

DM, n (%)

31 (56.4%)

6 (54.5%)

NS

Medications

  

NS

 ACEIs or ARBs, n (%)

7 (63.6%)

38 (69.0%)

NS

 Calcium antagonists, n (%)

8 (72.7%)

45 (81.8%)

NS

 Beta blockers, n (%)

3 (27.2%)

14 (25.4%)

NS

 Diuretics, n (%)

9 (81.8%)

49 (89.0%)

NS

 Statins, n (%)

3 (27.2%)

14 (25.4%)

NS

 Antiplatelet agents, n (%)

3 (27.2%)

14 (25.4%)

NS

Serum BUN (mg/dL)

103.8 ± 39.4

92.5 ± 36.0

NS

Serum creatinine (mg/dL)

11.0 ± 5.2

8.7 ± 4.7

NS

eGFRCr (mL/min/1.73 m2)

5.6 ± 2.4

7.5 ± 3.0

NS

Serum cystatin C (mg/L)

4.7 ± 0.8

4.0 ± 0.7

<0.01

eGFRcysC (mL/min/1.73 m2)

12.2 ± 2.7

15.0 ± 2.8

<0.01

Total cholesterol (mg/dL)

182.6 ± 48.1

143.3 ± 45.1

0.011

LDL-C (mg/dL)

117.3 ± 45.8

85.8 ± 38.0

0.033

HDL-C (mg/dL)

39.7 ± 16.1

38.7 ± 14.0

NS

Albumin (g/dL)

3.4 ± 0.3

3.5 ± 0.6

NS

CRP (mg/dL)

0.57 (0.18–2.10)

0.11 (0.04–0.64)

NS

eGFRcysC Cystatin C-based estimated glomerular filtration rate, LDL-C low-density lipoprotein cholesterol, HDL-C high-density lipoprotein cholesterol, CRP C-reactive protein, BUN blood urea nitrogen, eGFRCr creatinine-based estimated glomerular filtration rate, BMI body mass index, CRP C-reactive protein, ACEIs angiotensin-converting enzyme inhibitors, ARBs angiotensin II receptor blockers

Cardiovascular outcomes according to cystatin C level and predictors for cardiovascular outcomes

Enrolled patients were classified into two groups according to the mean level of cystatin C (4.14 mg/L) for Kaplan–Meier analysis. The results showed that patients in the lower cystatin C level group (<4.14 mg/L) had a better event-free survival rate than patients in the higher cystatin C level group (≥4.14 mg/L) (p = 0.039) (Fig. 1). In univariate analysis, cystatin C (hazard ratio [HR] 2.62, 95% confidence interval [CI] 1.24–5.53, p = 0.011) and eGFRcysC (HR 0.64, 95% CI 0.47–0.87, p = 0.004) were significant factors for the prediction of cardiovascular events. Age of the subjects also contributed to cardiovascular events, even though the significance did not reach p < 0.05 (HR 1.022, 95% CI 0.986–1.060, p = 0.084) (Table 5). Because the correlation coefficient between cystatin C and eGFRcysC was >0.9, multivariate analysis was performed with the two different models. In model I, after cystatin C and age were adjusted, both age and cystatin C were independent determinants of cardiovascular events. However, in model II, which adjusted age and eGFRcysC, only eGFRcysC was an independent determinant of cardiovascular events (HR 0.64, 95% CI 0.47–0.87, p = 0.04) (Table 6).
https://static-content.springer.com/image/art%3A10.1007%2Fs10157-011-0583-1/MediaObjects/10157_2011_583_Fig1_HTML.gif
Fig. 1

Cardiovascular event-free survival curves according to mean cystatin C level (4.14 mg/L)

Table 5

Univariate analysis for the risk of cardiovascular events

Predictor variable

Cardiovascular events

HR (95% CI)

p value

Age (per year)

1.022 (0.986–1.060)

0.084

BMI (per 1.0 kg/m2 increment)

1.006 (0.812–1.246)

0.956

Smoking (vs non smoking)

2.201 (0.642–7.545)

0.210

Diabetes (vs non-diabetes)

1.416 (0.426–4.706)

0.570

BUN (per 1.0 mg/dL increment)

1.000 (0.987–1.014)

0.945

Creatinine (per 1.0 mg/dL increment)

1.014 (0.929–1.106)

0.761

eGFRCr (per 1.0 mL/min/1.73 m2 increment)

0.886 (0.704–1.115)

0.303

Cystatin C (per 1.0 mg/L increment)

2.620 (1.241–5.528)

0.011

eGFRcysC (per 1.0 mL/min/1.73 m2 increment)

0.640 (0.471–0.870)

0.004

Albumin (per 1.0 g/dL increment)

1.234 (0.454–3.351)

0.680

Total cholesterol (per 1.0 mg/dL increment)

1.008 (0.998–1.019)

0.128

LDL-C (per 1.0 mg/dL increment)

1.010 (0.996–1.025)

0.155

HDL-C (per 1.0 mg/dL increment)

0.995 (0.956–1.036)

0.821

CRP (per 1.0 mg/dL increment)

0.992 (0.785–1.254)

0.947

CI Confidence interval, HR hazard ratio, eGFRcysC cystatin C-based estimated GFR, BMI body mass index, BUN blood urea nitrogen, eGFRCr creatinine-based estimated GFR, LDL-C low-density lipoprotein cholesterol, HDL-C high-density lipoprotein cholesterol, CRP C-reactive protein

Table 6

Multivariate analysis for the risk of cardiovascular events

Models

HR (95% CI)

p value

Model I

 Cystatin C (per 1.0 mg/L increment)

3.952 (1.693–9.226)

0.001

 Age (per year)

1.654 (1.004–1.107)

0.034

Model II

 eGFRcysC (per 1.0 mL/min/1.73 m2 increment)

0.640 (0.471–0.879)

0.004

 Age (per year)

1.003 (0.991–1.076)

0.131

The following parameters were included in the multivariate analysis model. Model I: adjusted for age, cystatin C, model II: adjusted for age, eGFRcysC

eGFRcysC Cystatin C-based estimated GFR

Discussion

In the current study, we found that the lower cystatin C group (<4.14 mg/L) had a better cardiovascular events-free survival rate than the higher cystatin C group (≥4.14 mg/L). Furthermore, eGFRcysC was a significant predictor for cardiovascular events in incident dialysis patients. Although the lower cystatin C group could account for an early dialysis starter group, creatinine and eGFRCr by MDRD were not significant factors for the prediction of cardiovascular events and these findings might suggest another role for serum cystatin C. In part, these results are consistent with previous studies demonstrating that cystatin C level predicts cardiovascular outcomes in another population. In the general population and the elderly, serum cystatin C has been shown to be a better predictor of mortality and adverse cardiovascular events than serum creatinine [1214]. An elevated cystatin C level (>1 mg/L) in subjects with eGFRCr >60 mL/min/1.73 m2 has been used to classify subjects as having preclinical kidney disease, which portended an increased risk of CVD, CKD, and death [22]. A recent study showed that serum cystatin C level was also associated with all-cause and CVD mortality in stage 3 and 4 CKD patients [15].

Serum creatinine concentration is affected by several factors that are independent of GFR, such as age, race, muscle mass, gender, medication use and catabolism [23]. On the other hand, serum cystatin C possesses many of the characteristics required of an ideal marker of GFR. It is an endogenous substance with constant production, and is freely filtered in the glomeruli. There is no tubular secretion or reabsorption, and cystatin C can nowadays be easily measured [10]. In a prior meta-analysis study, it suggested that serum cystatin C was clearly superior to serum creatinine as a marker of GFR measured by correlation or mean receiver operating characteristic-plot area under the curve [24].

Although it has been known that serum cystatin C was less influenced by factors other than GFR compared with creatinine, evidence of an association between cystatin C and factors unrelated to GFR have been reported [25, 26]. A correlation between cystatin C and age is to be expected, as GFR declines with age [27]. The association between cystatin C and inflammatory markers has been described by some authors [25, 28, 29], although this may be expected as inflammation is activated in renal insufficiency. However, the only study adjusted for measured GFR still suggested a weak relationship between CRP and cystatin C [30, 31]. A number of prior studies have found cystatin C to be associated with BMI [25, 32]. Knight et al. [25], in an analysis of the Prevention of Renal and Vascular End-Stage Disease (PREVEND) study, adjusted for 24 h creatinine clearance and found an association between cystatin C and BMI and CRP. Our study showed that serum cystatin C level was associated with age, smoking, BUN, creatinine, and eGFRCr by MDRD; however, there was no correlation between cystatin C and BMI or CRP. These results were inconsistent with the results of other studies in non-ESRD patients. Further studies are needed to confirm these associations.

To date, attempts to convert cystatin C values to corresponding GFR have resulted in multiple equations in prior studies [21, 33, 34]. eGFRcysC has been found to predict CVD mortality better than eGFRCr in the general adult population of the USA [14]. Furthermore, eGFRcysC has had a clinical role in distinguishing between higher risk and lower risk individuals for cardiovascular and mortality outcomes with eGFRCr <60 mL/min/1.73 m2 [35]. These results are consistent with our study demonstrating that eGFRcysC might be an independent predictor of cardiovascular events in incident dialysis patient. The reason for the different performance of the two estimates of GFR may be either different abilities to estimate GFR or a dependence on non-GFR confounders [36]. Mathisen and colleagues recently described that an eGFR based on either creatinine or cystatin C was influenced by traditional cardiovascular risk factors even after adjusting for GFR by iohexol clearance, indicating that the mechanism for these associations were independent of GFR, especially true for eGFR based on cystatin C [37]. One of these non-GFR determinants of cystatin C is possibly its anti-atherosclerotic activity. Anti-atherosclerotic activity of cystatin C was identified in an animal model study that showed cystatin C-deficient mice had larger subvalvular plaques compared with control mice [38, 39]. Previously, our group reported that cystatin C was related to pulse wave velocity, a marker representing arterial stiffness, independently of serum creatinine and creatinine-based GFR [8]. Increased arterial stiffness is one of the pathological states of vascular damage and is closely associated with atherosclerosis. This hypothesis has been supported by radiological findings that higher serum cystatin C concentrations were correlated with early-stage coronary atherosclerotic plaques among patients without established CKD on multidetector computed tomography and the results suggested that atherosclerosis could lead to increased production of cystatin [40]. Previous studies suggested that cystatin C was associated with an increased risk of coronary artery disease [4144]. Furthermore, a high level of cystatin C was related to suspected or confirmed acute coronary syndrome [45]. In 2007, the European Society of Cardiology recommended the use of cystatin C for predicting myocardial infarction and long-term mortality in patients with non-ST elevation acute coronary syndrome [46]. As mentioned earlier, cystatin C seems to be a related inflammatory process. A possible implication of cystatin C in the inflammatory process was seen by a significant correlation with CRP, apolipoprotein A1, and monocytes, which play an important role in chronic inflammation and atherosclerosis [4749]. Ferraro and colleagues recently described that inflammatory cytokines mainly contribute to perturb the cystatin C-cathepsins equilibrium in pathological mechanisms promoting several diseases [50]. These findings may provide a plausible link for cystatin C and inflammatory markers in predicting cardiovascular events related to non-GFR determinants, and support the important role of serum cystatin C in predicting cardiovascular events. Based on the above investigations, we thought that eGFRcysC as well as serum cystatin C, might have a function of predicting cardiovascular events beyond the estimation of true GRF in incident HD patients of the current study.

Our current study has some limitations. First, the number of enrolled subjects was small. Second, measuring cystatin C only once was rationally not sufficient to expect cardiovascular event rates for several months afterwards. Third, because enrolled patients were all Korean from a single center, we were not able to elucidate the effect of ethnicity and generalize to all populations. Fourth, owing to the retrospective study, it was difficult to clarify the exact mechanism by which serum cystatin C was associated with cardiovacular risk factor, being either related to GFR or non-GFR determinants. Moreover, since the true GFR measurements were not conducted, we could not delineate non-GFR contribution to cardiovascular event-related cystatin C and eGFRcysC. Nonetheless, it is valuable to demonstrate that cystatin C and eGFRcysC might be important in predicting future cardiovascular outcomes in incident dialysis patients.

In conclusion, serum cystatin C might be an independent marker of cardiovascular events in incident dialysis patients. Furthermore, eGFRcysC based on measured serum cystatin C, could have a new role in predicting cardiovascular events beyond the estimation of true GFR. Further investigations with a larger sample size and prospective designs are required to confirm the potential application of serum cystatin C and eGFRcysC as useful predictors for cardiovascular events.

Conflict of interest

The authors have declared that no conflict of interest exists.

Copyright information

© Japanese Society of Nephrology 2012