- Email: [email protected]
Influence of initiation of maintenance hemodialysis on biomarkers of inflammation and oxidative stress
Kidney International, Vol. 65 (2004), pp. 2371–2379
Influence of initiation of maintenance hemodialysis on biomarkers of inflammation and oxidative stress LARA B. PUPIM, JONATHAN HIMMELFARB, ELLEN MCMONAGLE, YU SHYR, and T. ALP IKIZLER Department of Medicine, Division of Nephrology, Vanderbilt University Medical Center, Nashville, Tennessee; Department of Medicine, Division of Nephrology, Maine Medical Center, Portland, Maine; and Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee
Influence of initiation of maintenance hemodialysis on biomarkers of inflammation and oxidative stress. Background. Chronic inflammation and oxidative stress are highly prevalent in patients with chronic kidney disease and end-stage renal disease (ESRD). These conditions contribute to high mortality rates associated with cardiovascular disease, the leading cause of death in this patient population. To our knowledge, no prospective studies have examined how initiation of maintenance hemodialysis (MHD) affects biomarkers of inflammation and oxidative stress status. Methods. This was a prospective cohort study evaluating time-dependent changes in C-reactive protein (CRP), interleukin-6 (IL-6), interleukin-10 (IL-10), and protein carbonyl content before and after initiation of MHD over a 12-month period. Fifty incident hemodialysis patients [57.6 ± 17.2 years, 60% male, 38% Caucasians, 32% insulin-dependent diabetes mellitus (IDDM) were studied, with 50 healthy subjects for comparison. The study variables were assessed before the initial outpatient hemodialysis treatment, and every 3 months thereafter for 12 months. Results. At baseline, CRP, IL-6, and carbonyl content levels were significantly higher in MHD patients compared with healthy subjects (P < 0.001 for each). After initiation of MHD, there were no significant changes in any of the study variables. Patients who initiated MHD with the highest levels of all the study variables had a significant decrease over the next year, but the variables were still higher than normal at the end of the 12-month study period. Conclusion. Our data show that initiation of MHD does not have significant influence on plasma concentrations of CRP, IL6, and IL-10, as well as plasma protein carbonyl content. These findings suggest that MHD is ineffective in controlling inflammation and oxidative stress in uremia.
Patients with advanced chronic kidney disease develop a variety of signs and symptoms related to uremic toxicity,
Key words: hemodialysis, cytokines, C-reactive protein, inflammation, oxidative stress. Received for publication April 9, 2003 and in revised form October 6, 2003, and December 15, 2003 Accepted for publication January 6, 2004 C
2004 by the International Society of Nephrology
including gastrointestinal [1], neurologic [2], and cardiac symptomatology [3]. Uremia is also characterized by a variety of metabolic perturbations, including metabolic acidosis [4], hyperphosphatemia [5], azotemia, and disorders of volume homeostasis [6]. Maintenance hemodialysis (MHD) therapy is initiated to lessen uremic symptoms, improve nutritional status, improve the metabolic milieu, and to sustain and improve the overall quality of life. Unfortunately, however, mortality and morbidity rates of patients receiving maintenance dialysis therapy remain unacceptably high despite demonstrable improvements in metabolic parameters, volume homeostasis, and nutritional status [7, 8] that occur after dialysis is initiated [9]. The HEMO Study, the largest and most comprehensive randomized trial ever undertaken in the MHD population, was designed to test the hypothesis that either increasing the delivered dose of dialysis, or increasing dialysis membrane flux would improve all-cause mortality in a prevalent MHD population [10]. The primary results of this comprehensive trial have recently been reported and support the null hypothesis for both dose of dialysis and membrane flux. In light of the HEMO Study primary results, a systematic re-examination of the relative benefits or lack of benefits of the dialysis procedure on determinants of mortality in uremic patients is indicated [11]. To critically evaluate how initiation of dialysis affects uremic factors that contribute to mortality, prospective cohort studies of the incident dialysis population should be particularly valuable. Cardiovascular disease remains the leading cause of morbidity and mortality in the maintenance dialysis population [12]. Data increasingly suggest that a cytokinedriven acute-phase inflammatory response and an increase in oxidative stress status are highly prevalent in the maintenance dialysis population, and closely associated with the pathogenesis of cardiovascular disease [13, 14, 15]. To our knowledge, no prospective studies have examined how initiation of MHD affects biomarkers of inflammation and oxidative stress status. We therefore performed a prospective cohort in a group of incident
2371
2372
Pupim et al: Inflammation and oxidative stress in MHD
Table 1. Demographic characteristics of the patients studied MHD patients Gender M/F Race % Caucasian African American Age years (mean ± SD) Cause of ESRD % Diabetes Hypertension Glomerulonephritis Polycystic kidney disease Other Unknown
Healthy subjects
30 (60%)/20 (40%) 18 (36%)/32 (64%) 40 60 57.6 ± 17.2 (range 24–90)
94 06 49.7 ± 16.3 (range 24–85)
30 24 24 8 12 4
N/A N/A N/A N/A N/A N/A
Abbreviations are: MHD, maintenance hemodialysis; M, male; F, female; ESRD, end-stage renal disease.
hemodialysis patients evaluating time-dependent changes in biomarkers of inflammation (IL-6, IL-10, and CRP levels) and oxidative stress (plasma protein carbonyl content). METHODS Patient characteristics The study was conducted at the Vanderbilt University Outpatient Dialysis Unit. All newly initiated MHD patients were asked to participate in the study. The study period was from December 1997 through November 2000, and included all consenting patients surveyed from December 1997 to November 1999. There were no exclusion criteria (except age <18 years), and all incident patients were referred to enrollment in the study. Between December 1997 and November 1999, a total of 73 new patients were initiated on MHD at the Dialysis Unit. Of those 73 patients, 50 patients (68%) consented to participate in the study and were followed for 12 months. The only cause of nonenrollment was refusal by the patient. A group of 50 healthy subjects served as the control group only for the initial measurement of the study. In addition, CRP, IL-6, and carbonyl content were measured in 29 patients with diabetes and other comorbid conditions, such as cardiovascular disease and/or hypertension but no end-stage renal disease (ESRD), for a baseline comparison. The Institutional Review Board approved the study protocol and written consent was obtained from all study subjects. Of the 50 study patients, 30 (60%) were male and 20 (40%) were female. Specific demographic characteristics are shown in Table 1. Forty-one (82%) patients had complete follow-up until the end of the study, 4 patients died during the study, and 5 patients were lost to follow-up (4 patients transferred to another facility, and 1 patient received a liver transplant). The causes of death in the 4 patients were infection (1), cardiovascular (2), and unknown (1).
Study design This was a prospective cohort study. There were no interventions specific to the study protocol. Each patient’s attending nephrologist determined the timing of initiation of MHD and the dialysis prescription was according to the facility protocol (see below). The study variables consisted of serial measurements of biomarkers of inflammation and oxidative stress every 3 months for a period of 12 months after initiation of MHD. In a previous study with similar patient population we reported the trends in nutritional markers, as well as certain dialytic characteristics, over 12 months [7]. Table 2 shows serum albumin, serum prealbumin, and dialysis dose (Kt/V) values in the study patients. Baseline blood samples were collected immediately before the first outpatient hemodialysis treatment. Biochemical indices followed during the study included plasma IL-6 (IL-6), plasma IL-10 (IL-10), plasma protein carbonyl content, and serum CRP levels. The most recent measurements of 24-hour urine creatinine excretion before the initiation of MHD (range of 1 to 90 days) were used to calculate baseline creatinine clearance and were repeated at the 12-month interval. Twentyfour–hour creatinine excretion was available in 41 patients (82%). Dialysis adequacy was assessed by kinetically determined Kt/V and urea reduction ratio (URR) measurements at each study time point. Blood samples were obtained prehemodialysis in a nonfasting state at the beginning of the week, after two consecutive days without treatment, except for the initial sampling, which occurred on the day the patient initiated MHD therapy. During the study period, none of the study patients had a known acute inflammatory event at the time of blood draws (no hospitalization event within 1 week of scheduled sampling time, excluding the patients who were hospitalized for initiation of hemodialysis at the baseline measurement and not on antibiotic therapy). Blood was drawn into Vacutainer tubes containing ethyldiaminetetraacetic acid (EDTA) supplemented with 1000 U/mL catalase and centrifuged at 1700g for 15 minutes or into serum separator tubes. Serum CRP levels were measured using highsensitivity immunoturbidimetric assay by Roche Modular System (Indianapolis, IN, USA). Plasma IL-6 and IL-10 concentrations were determined by enzyme-linked immunosorbent assay (ELISA) with kits from BioSource International (Camarillo, CA, USA). Plasma protein carbonyl content was measured using the Zentech PC Test (Protein carbonyl Enzyme Immuno-assay Kit; Zenith Technology, Dunedin, New Zealand) as described previously. This kit follows the method outlined by Buss et al [16] as amended by Winterbourn and Buss [17], which utilizes derivatization of protein carbonyl in samples and oxidized protein standards with dinitrophenylhydrazine, followed by ELISA with an anti-DNP antibody and standard ELISA techniques for
2373
Pupim et al: Inflammation and oxidative stress in MHD
Table 2. Serum proteins and dialysis dose at follow-up time points during 12 months of study
Serum albumin g/dLa Serum prealbumin mg/dLa Kt/Va
Baseline N = 50
3 months N = 50
6 months N = 46
9 months N = 41
12 months N = 41
3.36 ± 0.08 30.1 ± 1.25 1.37 ± 0.3
3.59 ± 0.07 33.1 ± 1.49 1.45 ± 0.3
3.62 ± 0.07 34.5 ± 1.70 1.49 ± 0.3
3.72 ± 0.06 34.6 ± 1.34 1.66 ± 0.4
3.63 ± 0.06 32.6 ± 1.51 1.57 ± 0.3
Values reported are the mean ± SEM of each period for each variable. a Denotes significant trends over time, P < 0.001, by mixed-model procedure.
labeling and visualizing labeled molecules [16]. Absorbance is read at 450 nm on an MRX microplate reader (Dynex Technologies, Chantilly, VA, USA). A standard curve is plotted, and the carbonyl concentration of samples is read off the curve, using the MRX Revelation Software (Dynex Technologies). Assay analytical sensitivity was 2.0 pg/mL, 1.0 pg/mL for IL-6 and IL-10, respectively. Interassay variability was 6% and 3%, and intra-assay variability was 8% and 4% for IL-6 and IL-10, respectively. Carbonyl content assay had the following characteristics: 0.021 nmol/mg of analytical sensitivity and 3.8% of intra-assay variability. The interassay variability for carbonyl assay was 5.8% for high values, 7.7% for medium values, and 24.7% for low values. Analytical sensitivity of CRP was 0.003 mg/dL, with an interassay variation of 5.9%, 3.0%, and 2.4%, respectively for low, medium, and highest values, and an intraassay variation of 1.3%, 0.4%, and 0.3% for low, medium, and highest values. All patients underwent formal urea kinetic modeling at each data collection during the study period. The minimum dialysis dose delivered at the dialysis unit was Kt/V greater than or equal to 1.4 (single pool). All patients were dialyzed with high-flux biocompatible membranes (Fresenius F80; Fresenius USA, Lexington, MA, USA). The dialysate used was identical for all treatments and consisted of sodium 139 mEq/L, potassium 2 mEq/L, calcium 2.5 mEq/L, glucose 200 mg/dL, and bicarbonate 39 mEq/L. Dialyzers were reused with formaldehyde and bleach for a maximum of 30 times, with an average of 18 reuses. The process of reuse involved an initial exposure to a diluted bleach solution (0.375%), rinsing of the bleach, and sterilizing with formaldehyde diluted to a concentration of 1.5% using the Fresenius DRS-4 machine (Fresenius USA). Dialyzers were then stored for a minimum of 24 hours before the subsequent patient’s blood contact. Dialysate quality was checked on a quarterly basis, and the bacteria and endotoxin levels have all been below highest acceptable limits (less than 50 CFU/mL and less than 0.99 EU/mL). Study patients’ calcium, phosphorus, total bicarbonate, and hematocrit levels were included in regular patient monthly laboratory reports. Every effort was made to keep laboratory parameters within acceptable ranges through established protocols. Recombinant human ery-
thropoetin and intravenous iron were prescribed to keep hematocrit within the range of 33% to 36% and transferrin saturation greater than 20%.
Statistical analysis Analyses of results focused on estimating the profile changes of biomarkers of inflammation and oxidative stress in ESRD upon initiation of MHD and over the 12-month study period. The null hypothesis was that initiation of MHD would not change levels of biomarkers of inflammation and oxidative stress in ESRD patients. The primary end point was circulating levels of IL-6. Fifty incident hemodialysis patients accepted and were included in this study during the enrollment period. With a sample size of 50, we estimated to have 93% power to detect a 10% difference in IL-6 means from baseline to the end of 1 year (i.e., an average change of 2.0 pg/mL in the means of IL-6 from baseline to 12-month time-point), assuming a standard deviation of the differences of 4.0 pg/mL (effect size of 0.5), using a paired t test approach and a two-sided 0.05 alpha level [18]. We anticipated having a 20% dropout rate in 1 year. Indeed, our sample size was 41 at the end of the study period, which provided 87% of power to detect the differences estimated above. Correlations among the study variables were performed by correlation matrix mixed procedure [19]. Comparisons of study variables between healthy control patients and the study population were done with the Student t test for normally distributed data and the Mann-Whitney U test for non-normally distributed data. The trends of study variables over the 12-month study period were analyzed using the restricted/residual maximum likelihood-based mixed effect model to adjust the intracorrelation effect for patients who had multiple measurements [20]. For this analysis, logarithmic transformation was used to account for the skewed distribution of IL-6, IL-10, and plasma carbonyl content. The model was selected based on Akaike’s Information Criterion and Schwarz’s Bayesian Criterion [21]. All tests of significance were two-sided, and differences were considered statistically significant when P value was less than 0.05. All data are expressed as mean ± SEM unless otherwise stated. SAS version 8.2 (Cary, NC, USA) was used for all analyses.
2374
Pupim et al: Inflammation and oxidative stress in MHD
Table 3. Biomarkers of inflammation and carbonyl levels at baseline comparing MHD patients and healthy subjects
IL-6 pg/mL CRP mg/dL IL-10 pg/mL Carbonyl nmol/mg protein
MHD patients
Healthy subjects
P value
20.17 ± 4.32 2.01 ± 0.06 2.91 ± 0.86 0.154 ± 0.014
4.02 ± 0.47 0.28 ± 0.06 1.06 ± 0.16 0.029 ± 0.004
<0.001 <0.001 0.249 <0.001
Abbreviations are: IL-6, interleukin-6; IL-10, interleukin-10; CRP, C-reactive protein. Values reported are the mean ± SEM of each period for all variables.
RESULTS Biomarkers of inflammation and oxidative stress are increased at initiation of hemodialysis We initially compared the study variables in the hemodialysis study population with 50 healthy subjects. The patients had significantly higher levels of IL-6 (20.17 ± 4.32 pg/mL vs. 4.02 ± 0.47 pg/mL, P < 0.001), CRP (2.01 ± 0.06 vs. 0.28 ± 0.06 mg/dL, P < 0.001), and carbonyl content (0.154 ± 0.014 nmol/mg of protein vs. 0.029 ± 0.004 nmol/mg of protein, P < 0.001) at the time of initiation of MHD therapy (Table 3). IL-10 levels were numerically lower in the healthy population compared with the ESRD population, but the difference did not reach statistical significance (1.06 ± 0.16 pg/mL vs. 2.91 ± 0.86 pg/mL, respectively, P = 0.249). When compared with diabetics without renal failure (63 ± 14 years old, 70% males), CHD patients had significantly higher levels of carbonyl content (0.154 ± 0.014 nmol/mg of protein vs. 0.021 ± 0.004 nmol/mg of protein, P < 0.001), IL-6 (20.17 ± 4.32 pg/mL vs. 6.05 ± 1.85 pg/mL, P = 0.015), and CRP (2.01 ± 0.06 mg/dL vs. 0.76 ± 0.2 mg/dL, P = 0.030) at the time of initiation of MHD therapy. Biomarkers of inflammation and oxidative stress are stable during the first year of MHD Directional changes of plasma IL-6, plasma IL-10, serum CRP, and plasma carbonyl content were analyzed over a period of 12 months. There were no significant changes in any of the study variables over the study period. This was consistent for overall trends as well as comparisons of each time point to baseline. Specifically, plasma IL-6 levels changed from 20.17 ± 4.32 pg/mL at baseline to 15.94 ± 4.22 pg/mL at 6 months, and 18.22 ± 5.27 pg/mL at 12 months’ follow-up (P = 0.983 for the trend) (Fig. 1). Serum CRP levels changed from 2.01 ± 0.06 mg/dL to 1.79 ± 0.07 mg/dL at 6 months, and to 1.25 ± 0.08 mg/dL at 12 months (P = 0.651) (Fig. 2). IL-10 concentrations were 2.91 ± 0.86 pg/mL at baseline, 0.94 ± 0.84 pg/mL at 6 months, and 1.00 ± 0.09 pg/mL at 12 months (P = 0.121). Plasma carbonyl content also decreased from 0.154 ± 0.014 nmol/kg of protein to 0.122 ± 0.013 nmol/kg of protein in the first 6 months, but dis-
played a subsequent increase to 0.162 ± 0.017 nmol/kg of protein in 12 months (P = 0.887) (Fig. 3). Changes in biomarkers of inflammation are dependent on baseline concentrations An analysis of trends over time based on initial quartiles for study variables was conducted. The trends were dependent on the initial baseline values of study variables. Figure 4 shows this analysis for IL-6. As can be seen, patients who initiated MHD with highest levels of IL-6 at baseline (Q4) experienced a significant decrease in IL-6 concentrations over the study period (58.02 ± 14.06 pg/mL to 19.15 ± 4.33 pg/mL, P = 0.036). On the contrary, patients who initiated MHD therapy in the two lowest quartiles of IL-6 (Q1 and Q2) increased their levels significantly (from 2.59 ± 0.30 pg/mL to 12.77 ± 1.30 pg/ mL for Q1, P < 0.001, and from 6.48 ± 0.38 to 20.77 ± 4.97 pg/mL for Q2, P = 0.029). Patients in the Q3 group had a trend for increase but it did not reach statistical significance (13.57 ± 1.43 pg/mL to 19.19 ± 4.65 pg/mL, P = 0.647). Similar trends were observed for IL-10, CRP, and carbonyl content. Figures 5 and 6 show these trends for CRP and carbonyl content, respectively. Most important, none of the quartiles for any study variable reached normal values at the end of the 12-month study period. For example, at the end of 1 year of MHD, levels of IL-6 were 12.77 pg/mL, 20.77 pg/mL, 19.19 pg/mL, and 19.15 pg/mL for Q1, Q2, Q3, and Q4, respectively. Levels of CRP, IL10, and carbonyl content remained similarly elevated at the end of the study period. Lack of influence of residual renal function and other clinical parameters To elucidate any potential interaction of residual renal function on the observed outcomes, we analyzed the adjusted association of residual renal function, assessed by 24-hour urine creatinine clearance, with the study variables, by mixed model procedure. No significant associations were found between the baseline creatinine clearance and the future trends of CRP (P = 0.384), IL-6 (P = 0.863), IL-10 (P = 0.721), or carbonyl content (P = 0.391). We additionally analyzed the trends in biomarkers of inflammation and oxidative stress according to the type of vascular access, concurrent illnesses, and the presence of diabetes. Our results suggested that the presence of diabetes, when analyzed separately, was associated with significant increases in carbonyl content over the study period (P = 0.040), although the trends in nondiabetics (P = 0.756) and the overall trends were not statistically significant. Analyses of the other study variables were not affected by the presence of diabetes. Hypertension, cardiovascular disease, and type of vascular access did not influence trends in any of the study variables.
Pupim et al: Inflammation and oxidative stress in MHD
30
2375
P = 0.204
25 21.95
IL-6, pg/mL
20
20.17
15
15.94
17.04
18.22
10 5 4.02 0 Control (N = 50)
Baseline 3 months (N = 50) (N = 50)
6 months 9 months 12 months (N = 46) (N = 41) (N = 41)
Fig. 1. Black bars represent changes in plasma interleukin (IL)-6 concentration at follow-up time points during 12 months of study. White bar represents one time-point measurement in healthy control patients. Bars and error bars represent mean ± SEM for each time point.
4 3.5
P = 0.651
Serum CRP, mg/dL
3 2.5 2
2.01
2.06 1.79
1.5
1.30 1
1.25
0.5 0
0.28 Control (N = 50)
Baseline (N = 50)
3 months 6 months 9 months 12 months (N = 50) (N = 46) (N = 41) (N = 41)
Correlations among study variables The associations among the four study variables were analyzed in a multivariate matrix correlation by mixedmodel procedure. There were significant positive correlations comparing CRP with IL-6 (estimate = 0.546, P = 0.012), and CRP with IL-10 (estimate = 1.893, P < 0.001). CRP correlated substantially with carbonyl content, but it did not reach statistical significance (estimate = 1.330, P = 0.599). IL-6 was negatively correlated with IL-10 but this correlation was not statistically significant (estimate = −0.067, P = 0.678). DISCUSSION The current study was designed to determine the influence, if any, of initiation of conventional MHD on cir-
Fig. 2. Black bars represent changes in serum C-reactive protein (CRP) concentration at follow-up time points during 12 months of study. White bar represents one time-point measurement in healthy control patients. Bars and error bars represent mean ± SEM for each time point.
culating levels of biomarkers of inflammation (cytokines and CRP), as well as a marker of oxidative stress (plasma protein carbonyl content). Our results showed that the levels of CRP, IL6, and carbonyl content were significantly higher in patients than healthy control patients at baseline. IL-10 was numerically higher in patients than control patients, but the difference did not reach statistical significance. In addition, there were no significant changes in any study variable over a 12-month period after initiation of MHD. After subdividing patients according to their baseline level of each study variable, highest values decreased significantly for all variables, and lowest values increased significantly for IL-6, IL-10, and carbonyl content. These findings are in accordance with the statistical occurrence known as regression to the mean, suggesting that the study intervention (i.e., hemodialysis)
2376
Pupim et al: Inflammation and oxidative stress in MHD
P = 0.887
Carbonyl content, nmol/mg protein
0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02
0.029
0 Control (N = 50)
Baseline (N = 50)
3 months (N = 50)
Q1 60
Q2
6 months 9 months 12 months (N = 46) (N = 41) (N = 41)
Q3
Fig. 3. Black bars represent changes in plasma carbonyl content at follow-up time points during 12 months of study. White bar represents one time-point measurement in healthy control patients. Bars and error bars represent mean ± SEM for each time point.
Q4
Q4, P = 0.367
55 50 IL-6 quartiles, pg/mL
45 40 35 30 25 20 15 Q3, P = 0.647 10 Q2, P = 0.029 5 Q1, P = 0.001 0 Baseline
12 months
has no significant effect on inflammation or oxidative stress in MHD patients. The [lack of] changes in biomarkers of inflammation and oxidative stress over time were independent of clinical covariates, such as presence or absence of diabetes with cardiovascular disease and/or hypertension, as well as the type of vascular access used. Acute-phase inflammation is a common feature of ESRD and it is estimated that 30% to 50% of ESRD patients have serologic evidence of an activated inflammatory response [13]. The plasma levels of the prototypical acute-phase protein CRP have been demonstrated in numerous studies to be a powerful predictor of subsequent cardiovascular events and all-cause mortality in the dialysis population [22, 23]. Plasma IL-6 levels are also increasingly being recognized as an independent predictor
Fig. 4. Changes in plasma interleukin (IL)-6 concentration over the study period according to quartiles. Each line represents one quartile.
of cardiovascular events, progression of atherosclerosis, and all-cause mortality in the dialysis population [24–28]. In this study, we also measured plasma levels of the anti-inflammatory cytokine IL-10. IL-10 is known to counter-regulate the cascade of proinflammatory cytokines, including IL-6, as part of the acute-phase reaction. The importance of IL-10 as an anti-inflammatory cytokine is emphasized by a recent study in which a lowproducing IL-10 gene polymorphism was associated with a higher cardiovascular morbidity and mortality in a cohort of MHD patients [29]. Early data suggested that plasma levels of IL-6 and IL-10 were inversely related in the hemodialysis population [30]. Other studies have demonstrated overproduction of IL-10 in the hemodialysis population [31]. A recent study using a sensitive
Pupim et al: Inflammation and oxidative stress in MHD
Q1
Q2
Q3
2377
Q4
7.0 6.5 Q4, P = 0.008 6.0
CRP quartiles, mg/dL
5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0
Q3, P = 0.973
0.5 Q2, P = 0.587 Q1, P = 0.155 0.0 Baseline
12 months
Q1
Q2
Q3
Fig. 5. Changes in serum C-reactive protein (CRP) concentration over the study period according to quartiles. Each line represents one quartile.
Q4
0.300
Carbonyl quartiles, nmol/mg protein
Q4, P = 0.003 0.250
0.200
0.150
0.100
Q3, P = 0.531
Q2, P = 0.244 Q1, P = 0.028
0.050
0.000 Baseline
12 months
technique measuring monokine production at the single cell level demonstrated that monocytes from hemodialysis patients tend to overproduce both IL-6 and IL-10 simultaneously, while healthy subjects’ monocytes produce IL-6 and IL-10 in separate monocyte subpopulations [32]. In the present study, a trend toward higher plasma levels of IL-10 was detected in dialysis patients compared
Fig. 6. Changes in plasma carbonyl protein content over the study period according to quartiles. Each line represents one quartile.
with healthy subjects with a significant relationship to CRP levels. Similar to IL-6, plasma IL-10 levels did not change significantly over the course of time after initiation of MHD therapy. The underlying etiology of the accentuated inflammatory response in uremia remains poorly understood. While multiple dialysis-related factors can contribute to
2378
Pupim et al: Inflammation and oxidative stress in MHD
maintenance of a chronic inflammatory response, recent data demonstrating that plasma IL-6 and CRP levels are already elevated in earlier stages of chronic kidney disease suggest a more proximate relationship to loss of kidney function [33]. Indeed, our results clearly show that plasma levels of cytokines and carbonyl content were significantly elevated at the time of initiation of maintenance hemodialysis. Loss of renal function may be directly associated with reduced clearance of cytokines, complement peptides, oxidants, and other proinflammatory solutes. Additionally, the hemodialysis procedure itself may contribute to a proinflammatory milieu by increasing the synthesis of fibrinogen, IL-6, and other proinflammatory peptides, likely by bioincompatibility from membrane and dialysate exposure during the extracorporeal procedure [34]. However, in this study we did not observe a significant increase in inflammatory markers during the initial 12 months of MHD. This is in contrast to a report by Haubitz et al in which significant increases in CRP levels were reported after initiation of MHD in 8 patients. While the exact reasons for this discrepancy are not clear, it should be noted that the sample size in the former study was relatively small and the patients were dialyzed with a bioincompatible membrane, which might have initiated an inflammatory response, per se [35]. An increase in oxidative stress is increasingly recognized as an important metabolic accompaniment to ESRD [36]. Reactive oxygen species are known to activate phagocytic cells and, by way of NF-jB, up-regulate the release of IL-6 and increase production of CRP [37]. Our group has recently described a linkage between increased oxidative stress and acute-phase inflammation in severely hypoalbuminemic MHD patients [38]. In the present study, we have measured plasma protein carbonyl content as a biomarker of oxidative stress status [39–41]. The importance of increased carbonyl is evidenced by their important role in the pathogenesis of atherosclerosis [42]. The present study confirms high levels of plasma protein carbonyl content in patients with advanced kidney disease, and demonstrates no significant effect of initiating MHD therapy on plasma carbonyl content over time. Similar to the data on biomarkers of inflammation, these data suggest that initiation of MHD therapy is not adequate to control oxidative stress in uremia. A corollary interpretation of the data in this study is that the hemodialysis procedure itself is not the predominant cause of the increased oxidative stress and inflammation observed in the ESRD population. Thus, hemodialysis may not accelerate atherogenesis, so much as simply not abate its development. This suggestion is supported by a growing body of evidence demonstrating that many proatherogenic metabolic alterations in uremia are already present before the initiation of dialysis therapy. Furthermore, previous work by our group and by other investigators has demonstrated that the net effect
of a single hemodialysis treatment is to improve overall redox balance [39]. Even when bioincompatible membranes, which vigorously activate the alternative pathway of complement are used, it is extraordinarily difficult to detect an increase in oxidative stress over the course of the dialysis procedure. Thus, an important potential implication of this study and other similar studies is that improvements in the proatherogenic metabolic milieu in uremia will require the removal or biologic inactivation of solutes not readily cleared by the diffusive process in hemodialysis therapy. Either alternative types of renal replacement therapy (such as hemodiafiltration or the renal tubular assist device) or pharmacologic therapies (such as antioxidants or anti-inflammatory agents) may be required to control the uremia-associated oxidative stress and inflammation. The present study in incident MHD patients should not be viewed as an attempt to quantitatively measure the potentially competing effects of hemodialysis and uremia on inflammation and oxidative stress biomarkers. Instead, it should be viewed rather as an analysis of changes observed in these parameters after definitive initiation of conventional MHD therapy in patients with advanced ESRD. While a randomized controlled trial is always preferable, such design is not applicable for the hypothesis tested in this study. The current data must also be evaluated in the context of recent publications suggesting that initiation of MHD is associated with improvements in several nutritionally and metabolically relevant markers to some extent [7, 8]. In a similar group of patients, we observed an increase in serum albumin and serum prealbumin in the setting of increased solute clearance at adequate levels [7]. The results from this study, on the other hand, show that such an effect is not observed for inflammation or oxidative stress markers, important contributors to the high cardiovascular morbidity and mortality observed in ESRD patients. Together with the context of the recent HEMO Study results, we can postulate that conventional dialysis is “inadequate” as an anti-inflammatory and antioxidant therapy.
CONCLUSION We have shown that ESRD patients have substantially increased inflammatory and oxidative stress burden at the initiation of MHD treatment with currently prescribed treatment parameters. Our results suggest that conventional MHD is not effectively treating the highly prevalent inflammation and oxidative stress in ESRD patients. The lack of effective therapy for these metabolic perturbations may contribute to an ongoing high cardiovascular morbidity and mortality. Further research is needed to elucidate the mechanistic pathways of inflammation and oxidative stress, both before and after MHD is initiated.
Pupim et al: Inflammation and oxidative stress in MHD
ACKNOWLEDGMENTS We thank Karen A. Kinne for providing administrative assistance in the production of this manuscript, Pamela Kent, R.D., for assistance with data collection, and Joel Botler, M.D., William Ervin, M.D., Carrie Brewer, and Gary Banker for their assistance in obtaining blood samples from diabetic patients. This work is supported by National Institutes of Health grants DK45604, DK62849, and HL070938; and FDA Orphan Products Development grant FD-R-000943. Dr. Pupim is partly supported by the Marilyn Charitable Trust Young Investigator Grant of the National Kidney Foundation and Vanderbilt University School of Medicine Clinician Scientist Award, as part of the Vanderbilt Physician-Scientist Development Program. Reprint requests to T. Alp Ikizler, M.D., Associate Professor of Medicine, Division of Nephrology, S-3223 MCN, Vanderbilt University Medical Center, 1161 21st Ave. South & Garland, Nashville, TN 372322372. E-mail: [email protected]
REFERENCES 1. HAMMER J, OESTERREICHER C, HAMMER K, et al: Chronic gastrointestinal symptoms in hemodialysis patients. Wien Klin Wochenschr 110:287–291, 1998 2. REESE GN, APPEL SH: Neurologic complications of renal failure. Semin Nephrol 1:137–142, 1981 3. FOLEY RN, PARFREY PS: Cardiovascular disease and mortality in ESRD. J Nephrol 11:239–245, 1998 4. BAILEY JL, MITCH WE: Metabolic acidosis as a uremic toxin. Semin Nephrol 16:160–166, 1996 5. FIASCHI E, MIONI G, MASCHIO G, et al: Calcium and phosphorus metabolism in chronic uremia. Nephron 14:163–180, 1975 6. OBRADOR GT, PEREIRA BJ: Systemic complications of chronic kidney disease. Pinpointing clinical manifestations and best management. Postgrad Med 111:115–122 7. PUPIM LB, KENT P, CAGLAR K, et al: Improvement in nutritional parameters after initiation of chronic hemodialysis. Am J Kidney Dis 40:143–151, 2002 8. MEHROTRA R, BERMAN N, ALISTWANI A, et al: Improvement of nutritional status after initiation of maintenance hemodialysis. Am J Kidney Dis 40:133–142, 2002 9. AGODOA L: United States Renal Data System (USRDS). Nefrologia 20(Suppl 5):13–16, 2000 10. EKNOYAN G, BECK GJ, CHEUNG AK, et al: Effect of dialysis dose and membrane flux in maintenance hemodialysis. N Engl J Med 347:2010–2019, 2002 11. HIMMELFARB J: Success and challenge in dialysis therapy. N Engl J Med 347:2068–2070, 2002 12. FOLEY RN, HERZOG CA, COLLINS AJ: Blood pressure and long-term mortality in United States hemodialysis patients: USRDS Waves 3 and 4 Study. Kidney Int 62:1784–1790, 2002 13. KAYSEN GA: The microinflammatory state in uremia: Causes and potential consequences. J Am Soc Nephrol 12:1549–1557, 2001 14. BECKER BN, HIMMELFARB J, HENRICH WL, et al: Reassessing the cardiac risk profile in chronic hemodialysis patients: A hypothesis on the role of oxidant stress and other non-traditional cardiac risk factors. J Am Soc Nephrol 8:475–486, 1997 15. MAGGI E, BELLAZZI R, FALASCHI F, et al: Enhanced LDL oxidation in uremic patients: An additional mechanism for accelerated atherosclerosis? Kidney Int 45:876–883, 1994 16. BUSS IH, WINTERBOURN CC: Protein carbonyl measurement by ELISA. Methods Mol Biol 186:123–128, 2002 17. WINTERBOURN CC, BUSS IH: Protein carbonyl measurement by enzyme-linked immunosorbent assay. Methods Enzymol 300:106– 111, 1999 18. DUPONT WD, PLUMER WD: PS power and sample size program available for free on the Internet. Controlled Clin Trials 18:274, 1997 19. JENNRICH RI, SCHLUCHTER MD: Unbalanced repeated-measures models with structured covariance matrices. Biometrics 42:805–820, 1986
2379
20. DIGGLE P, LIANG KY, ZEGER SL: Analysis of Longitudinal Data, Oxford, Claredon Press, 1994 21. LIANG KY, ZEGER SL: Longitudinal data analysis using generalized linear models. Biometrika 73:13–22, 1986 22. ARICI M, WALLS J: End-stage renal disease, atherosclerosis, and cardiovascular mortality: Is C-reactive protein the missing link? Kidney Int 59:407–414, 2001 23. STENVINKEL P, HEIMBURGER O, PAULTRE F, et al: Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney Int 55:1899–1911, 1999 24. KATO A, ODAMAKI M, TAKITA T, et al: Association between interleukin-6 and carotid atherosclerosis in hemodialysis patients. Kidney Int 61:1143–1152, 2002 25. MEMOLI B, MINUTOLO R, BISESTI V, et al: Changes of serum albumin and C-reactive protein are related to changes of interleukin-6 release by peripheral blood mononuclear cells in hemodialysis patients treated with different membranes. Am J Kidney Dis 39:266– 273, 2002 26. PECOITS-FILHO R, BARANY P, LINDHOLM B, et al: Interleukin-6 is an independent predictor of mortality in patients starting dialysis treatment. Nephrol Dial Transplant 17:1684–1688, 2002 27. STENVINKEL P, HEIMBURGER O, JOGESTRAND T: Elevated interleukin6 predicts progressive carotid artery atherosclerosis in dialysis patients: Association with Chlamydia pneumoniae seropositivity. Am J Kidney Dis 39:274–282, 2002 28. KIMMEL PL, PHILLIPS TM, SIMMENS SJ, et al: Immunologic function and survival in hemodialysis patients. Kidney Int 54:236–244, 1998 29. GIRNDT M, KAUL H, SESTER U, et al: Anti-inflammatory interleukin10 genotype protects dialysis patients from cardiovascular events. Kidney Int 62:949–955, 2002 30. GIRNDT M, KOHLER H, SCHIEDHELM-WEICK E, et al: Production of interleukin-6, tumor necrosis factor alpha and interleukin-10 in vitro correlates with the clinical immune defect in chronic hemodialysis patients. Kidney Int 47:559–565, 1995 31. MORITA Y, YAMAMURA M, KASHIHARA N, et al: Increased production of interleukin-10 and inflammatory cytokines in blood monocytes of hemodialysis patients. Res Commun Mol Pathol Pharmacol 98:19– 33, 1997 32. GIRNDT M, SESTER U, KAUL H, et al: Production of proinflammatory and regulatory monokines in hemodialysis patients shown at a single-cell level. J Am Soc Nephrol 9:1689–1696, 1998 33. OBERG BP, MCMENAMIN E, LUCAS FL, et al: Increased prevalence of oxidant stress and inflammation in patients with moderate to severe chronic kidney disease. Kidney Int 65:1009–1016, 2004 34. CAGLAR K, PENG Y, PUPIM LB, et al: Inflammatory signals associated with hemodialysis. Kidney Int 62:1408–1416, 2002 35. HAUBITZ M, SCHULZE M, KOCH KM: Increase of C-reactive protein serum values following haemodialysis. Nephrol Dial Transplant 5:500–503, 1990 36. HIMMELFARB J, STENVINKEL P, IKIZLER TA, et al: The elephant in uremia: Oxidant stress as a unifying concept of cardiovascular disease in uremia. Kidney Int 62:1524–1538, 2002 37. BARNES PJ, KARIN M: Nuclear factor-kappaB: A pivotal transcription factor in chronic inflammatory diseases. N Engl J Med 336:1066– 1071, 1997 38. DANIELSKI M, IKIZLER TA, MCMONAGLE E, et al: Linkage of hypoalbuminemia, inflammation, and oxidative stress in patients receiving maintenance hemodialysis therapy. Am J Kidney Dis 42:286–294, 2003 39. HIMMELFARB J, MCMONAGLE E, MCMENAMIN E: Plasma protein thiol oxidation and carbonyl formation in chronic renal failure. Kidney International 58:2571–2578, 2000 40. MIYATA T, VAN YPERSELE DE STRIHOU C, KUROKAWA K, et al: Alterations in nonenzymatic biochemistry in uremia: Origin and significance of “carbonyl stress” in long-term uremic complications. Kidney Int 55:389–399, 1999 41. WEISS MF, ERHARD P, KADER-ATTIA FA, et al: Mechanisms for the formation of glycoxidation products in end-stage renal disease. Kidney Int 57:2571–2585, 2000 42. SAKATA N, IMANAGA Y, MENG J, et al: Increased advanced glycation end products in atherosclerotic lesions of patients with end-stage renal disease. Atherosclerosis 142:67–77, 1999
Influence of initiation of maintenance hemodialysis on biomarkers of inflammation and oxidative stress LARA B. PUPIM, JONATHAN HIMMELFARB, ELLEN MCMONAGLE, YU SHYR, and T. ALP IKIZLER Department of Medicine, Division of Nephrology, Vanderbilt University Medical Center, Nashville, Tennessee; Department of Medicine, Division of Nephrology, Maine Medical Center, Portland, Maine; and Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee
Influence of initiation of maintenance hemodialysis on biomarkers of inflammation and oxidative stress. Background. Chronic inflammation and oxidative stress are highly prevalent in patients with chronic kidney disease and end-stage renal disease (ESRD). These conditions contribute to high mortality rates associated with cardiovascular disease, the leading cause of death in this patient population. To our knowledge, no prospective studies have examined how initiation of maintenance hemodialysis (MHD) affects biomarkers of inflammation and oxidative stress status. Methods. This was a prospective cohort study evaluating time-dependent changes in C-reactive protein (CRP), interleukin-6 (IL-6), interleukin-10 (IL-10), and protein carbonyl content before and after initiation of MHD over a 12-month period. Fifty incident hemodialysis patients [57.6 ± 17.2 years, 60% male, 38% Caucasians, 32% insulin-dependent diabetes mellitus (IDDM) were studied, with 50 healthy subjects for comparison. The study variables were assessed before the initial outpatient hemodialysis treatment, and every 3 months thereafter for 12 months. Results. At baseline, CRP, IL-6, and carbonyl content levels were significantly higher in MHD patients compared with healthy subjects (P < 0.001 for each). After initiation of MHD, there were no significant changes in any of the study variables. Patients who initiated MHD with the highest levels of all the study variables had a significant decrease over the next year, but the variables were still higher than normal at the end of the 12-month study period. Conclusion. Our data show that initiation of MHD does not have significant influence on plasma concentrations of CRP, IL6, and IL-10, as well as plasma protein carbonyl content. These findings suggest that MHD is ineffective in controlling inflammation and oxidative stress in uremia.
Patients with advanced chronic kidney disease develop a variety of signs and symptoms related to uremic toxicity,
Key words: hemodialysis, cytokines, C-reactive protein, inflammation, oxidative stress. Received for publication April 9, 2003 and in revised form October 6, 2003, and December 15, 2003 Accepted for publication January 6, 2004 C
2004 by the International Society of Nephrology
including gastrointestinal [1], neurologic [2], and cardiac symptomatology [3]. Uremia is also characterized by a variety of metabolic perturbations, including metabolic acidosis [4], hyperphosphatemia [5], azotemia, and disorders of volume homeostasis [6]. Maintenance hemodialysis (MHD) therapy is initiated to lessen uremic symptoms, improve nutritional status, improve the metabolic milieu, and to sustain and improve the overall quality of life. Unfortunately, however, mortality and morbidity rates of patients receiving maintenance dialysis therapy remain unacceptably high despite demonstrable improvements in metabolic parameters, volume homeostasis, and nutritional status [7, 8] that occur after dialysis is initiated [9]. The HEMO Study, the largest and most comprehensive randomized trial ever undertaken in the MHD population, was designed to test the hypothesis that either increasing the delivered dose of dialysis, or increasing dialysis membrane flux would improve all-cause mortality in a prevalent MHD population [10]. The primary results of this comprehensive trial have recently been reported and support the null hypothesis for both dose of dialysis and membrane flux. In light of the HEMO Study primary results, a systematic re-examination of the relative benefits or lack of benefits of the dialysis procedure on determinants of mortality in uremic patients is indicated [11]. To critically evaluate how initiation of dialysis affects uremic factors that contribute to mortality, prospective cohort studies of the incident dialysis population should be particularly valuable. Cardiovascular disease remains the leading cause of morbidity and mortality in the maintenance dialysis population [12]. Data increasingly suggest that a cytokinedriven acute-phase inflammatory response and an increase in oxidative stress status are highly prevalent in the maintenance dialysis population, and closely associated with the pathogenesis of cardiovascular disease [13, 14, 15]. To our knowledge, no prospective studies have examined how initiation of MHD affects biomarkers of inflammation and oxidative stress status. We therefore performed a prospective cohort in a group of incident
2371
2372
Pupim et al: Inflammation and oxidative stress in MHD
Table 1. Demographic characteristics of the patients studied MHD patients Gender M/F Race % Caucasian African American Age years (mean ± SD) Cause of ESRD % Diabetes Hypertension Glomerulonephritis Polycystic kidney disease Other Unknown
Healthy subjects
30 (60%)/20 (40%) 18 (36%)/32 (64%) 40 60 57.6 ± 17.2 (range 24–90)
94 06 49.7 ± 16.3 (range 24–85)
30 24 24 8 12 4
N/A N/A N/A N/A N/A N/A
Abbreviations are: MHD, maintenance hemodialysis; M, male; F, female; ESRD, end-stage renal disease.
hemodialysis patients evaluating time-dependent changes in biomarkers of inflammation (IL-6, IL-10, and CRP levels) and oxidative stress (plasma protein carbonyl content). METHODS Patient characteristics The study was conducted at the Vanderbilt University Outpatient Dialysis Unit. All newly initiated MHD patients were asked to participate in the study. The study period was from December 1997 through November 2000, and included all consenting patients surveyed from December 1997 to November 1999. There were no exclusion criteria (except age <18 years), and all incident patients were referred to enrollment in the study. Between December 1997 and November 1999, a total of 73 new patients were initiated on MHD at the Dialysis Unit. Of those 73 patients, 50 patients (68%) consented to participate in the study and were followed for 12 months. The only cause of nonenrollment was refusal by the patient. A group of 50 healthy subjects served as the control group only for the initial measurement of the study. In addition, CRP, IL-6, and carbonyl content were measured in 29 patients with diabetes and other comorbid conditions, such as cardiovascular disease and/or hypertension but no end-stage renal disease (ESRD), for a baseline comparison. The Institutional Review Board approved the study protocol and written consent was obtained from all study subjects. Of the 50 study patients, 30 (60%) were male and 20 (40%) were female. Specific demographic characteristics are shown in Table 1. Forty-one (82%) patients had complete follow-up until the end of the study, 4 patients died during the study, and 5 patients were lost to follow-up (4 patients transferred to another facility, and 1 patient received a liver transplant). The causes of death in the 4 patients were infection (1), cardiovascular (2), and unknown (1).
Study design This was a prospective cohort study. There were no interventions specific to the study protocol. Each patient’s attending nephrologist determined the timing of initiation of MHD and the dialysis prescription was according to the facility protocol (see below). The study variables consisted of serial measurements of biomarkers of inflammation and oxidative stress every 3 months for a period of 12 months after initiation of MHD. In a previous study with similar patient population we reported the trends in nutritional markers, as well as certain dialytic characteristics, over 12 months [7]. Table 2 shows serum albumin, serum prealbumin, and dialysis dose (Kt/V) values in the study patients. Baseline blood samples were collected immediately before the first outpatient hemodialysis treatment. Biochemical indices followed during the study included plasma IL-6 (IL-6), plasma IL-10 (IL-10), plasma protein carbonyl content, and serum CRP levels. The most recent measurements of 24-hour urine creatinine excretion before the initiation of MHD (range of 1 to 90 days) were used to calculate baseline creatinine clearance and were repeated at the 12-month interval. Twentyfour–hour creatinine excretion was available in 41 patients (82%). Dialysis adequacy was assessed by kinetically determined Kt/V and urea reduction ratio (URR) measurements at each study time point. Blood samples were obtained prehemodialysis in a nonfasting state at the beginning of the week, after two consecutive days without treatment, except for the initial sampling, which occurred on the day the patient initiated MHD therapy. During the study period, none of the study patients had a known acute inflammatory event at the time of blood draws (no hospitalization event within 1 week of scheduled sampling time, excluding the patients who were hospitalized for initiation of hemodialysis at the baseline measurement and not on antibiotic therapy). Blood was drawn into Vacutainer tubes containing ethyldiaminetetraacetic acid (EDTA) supplemented with 1000 U/mL catalase and centrifuged at 1700g for 15 minutes or into serum separator tubes. Serum CRP levels were measured using highsensitivity immunoturbidimetric assay by Roche Modular System (Indianapolis, IN, USA). Plasma IL-6 and IL-10 concentrations were determined by enzyme-linked immunosorbent assay (ELISA) with kits from BioSource International (Camarillo, CA, USA). Plasma protein carbonyl content was measured using the Zentech PC Test (Protein carbonyl Enzyme Immuno-assay Kit; Zenith Technology, Dunedin, New Zealand) as described previously. This kit follows the method outlined by Buss et al [16] as amended by Winterbourn and Buss [17], which utilizes derivatization of protein carbonyl in samples and oxidized protein standards with dinitrophenylhydrazine, followed by ELISA with an anti-DNP antibody and standard ELISA techniques for
2373
Pupim et al: Inflammation and oxidative stress in MHD
Table 2. Serum proteins and dialysis dose at follow-up time points during 12 months of study
Serum albumin g/dLa Serum prealbumin mg/dLa Kt/Va
Baseline N = 50
3 months N = 50
6 months N = 46
9 months N = 41
12 months N = 41
3.36 ± 0.08 30.1 ± 1.25 1.37 ± 0.3
3.59 ± 0.07 33.1 ± 1.49 1.45 ± 0.3
3.62 ± 0.07 34.5 ± 1.70 1.49 ± 0.3
3.72 ± 0.06 34.6 ± 1.34 1.66 ± 0.4
3.63 ± 0.06 32.6 ± 1.51 1.57 ± 0.3
Values reported are the mean ± SEM of each period for each variable. a Denotes significant trends over time, P < 0.001, by mixed-model procedure.
labeling and visualizing labeled molecules [16]. Absorbance is read at 450 nm on an MRX microplate reader (Dynex Technologies, Chantilly, VA, USA). A standard curve is plotted, and the carbonyl concentration of samples is read off the curve, using the MRX Revelation Software (Dynex Technologies). Assay analytical sensitivity was 2.0 pg/mL, 1.0 pg/mL for IL-6 and IL-10, respectively. Interassay variability was 6% and 3%, and intra-assay variability was 8% and 4% for IL-6 and IL-10, respectively. Carbonyl content assay had the following characteristics: 0.021 nmol/mg of analytical sensitivity and 3.8% of intra-assay variability. The interassay variability for carbonyl assay was 5.8% for high values, 7.7% for medium values, and 24.7% for low values. Analytical sensitivity of CRP was 0.003 mg/dL, with an interassay variation of 5.9%, 3.0%, and 2.4%, respectively for low, medium, and highest values, and an intraassay variation of 1.3%, 0.4%, and 0.3% for low, medium, and highest values. All patients underwent formal urea kinetic modeling at each data collection during the study period. The minimum dialysis dose delivered at the dialysis unit was Kt/V greater than or equal to 1.4 (single pool). All patients were dialyzed with high-flux biocompatible membranes (Fresenius F80; Fresenius USA, Lexington, MA, USA). The dialysate used was identical for all treatments and consisted of sodium 139 mEq/L, potassium 2 mEq/L, calcium 2.5 mEq/L, glucose 200 mg/dL, and bicarbonate 39 mEq/L. Dialyzers were reused with formaldehyde and bleach for a maximum of 30 times, with an average of 18 reuses. The process of reuse involved an initial exposure to a diluted bleach solution (0.375%), rinsing of the bleach, and sterilizing with formaldehyde diluted to a concentration of 1.5% using the Fresenius DRS-4 machine (Fresenius USA). Dialyzers were then stored for a minimum of 24 hours before the subsequent patient’s blood contact. Dialysate quality was checked on a quarterly basis, and the bacteria and endotoxin levels have all been below highest acceptable limits (less than 50 CFU/mL and less than 0.99 EU/mL). Study patients’ calcium, phosphorus, total bicarbonate, and hematocrit levels were included in regular patient monthly laboratory reports. Every effort was made to keep laboratory parameters within acceptable ranges through established protocols. Recombinant human ery-
thropoetin and intravenous iron were prescribed to keep hematocrit within the range of 33% to 36% and transferrin saturation greater than 20%.
Statistical analysis Analyses of results focused on estimating the profile changes of biomarkers of inflammation and oxidative stress in ESRD upon initiation of MHD and over the 12-month study period. The null hypothesis was that initiation of MHD would not change levels of biomarkers of inflammation and oxidative stress in ESRD patients. The primary end point was circulating levels of IL-6. Fifty incident hemodialysis patients accepted and were included in this study during the enrollment period. With a sample size of 50, we estimated to have 93% power to detect a 10% difference in IL-6 means from baseline to the end of 1 year (i.e., an average change of 2.0 pg/mL in the means of IL-6 from baseline to 12-month time-point), assuming a standard deviation of the differences of 4.0 pg/mL (effect size of 0.5), using a paired t test approach and a two-sided 0.05 alpha level [18]. We anticipated having a 20% dropout rate in 1 year. Indeed, our sample size was 41 at the end of the study period, which provided 87% of power to detect the differences estimated above. Correlations among the study variables were performed by correlation matrix mixed procedure [19]. Comparisons of study variables between healthy control patients and the study population were done with the Student t test for normally distributed data and the Mann-Whitney U test for non-normally distributed data. The trends of study variables over the 12-month study period were analyzed using the restricted/residual maximum likelihood-based mixed effect model to adjust the intracorrelation effect for patients who had multiple measurements [20]. For this analysis, logarithmic transformation was used to account for the skewed distribution of IL-6, IL-10, and plasma carbonyl content. The model was selected based on Akaike’s Information Criterion and Schwarz’s Bayesian Criterion [21]. All tests of significance were two-sided, and differences were considered statistically significant when P value was less than 0.05. All data are expressed as mean ± SEM unless otherwise stated. SAS version 8.2 (Cary, NC, USA) was used for all analyses.
2374
Pupim et al: Inflammation and oxidative stress in MHD
Table 3. Biomarkers of inflammation and carbonyl levels at baseline comparing MHD patients and healthy subjects
IL-6 pg/mL CRP mg/dL IL-10 pg/mL Carbonyl nmol/mg protein
MHD patients
Healthy subjects
P value
20.17 ± 4.32 2.01 ± 0.06 2.91 ± 0.86 0.154 ± 0.014
4.02 ± 0.47 0.28 ± 0.06 1.06 ± 0.16 0.029 ± 0.004
<0.001 <0.001 0.249 <0.001
Abbreviations are: IL-6, interleukin-6; IL-10, interleukin-10; CRP, C-reactive protein. Values reported are the mean ± SEM of each period for all variables.
RESULTS Biomarkers of inflammation and oxidative stress are increased at initiation of hemodialysis We initially compared the study variables in the hemodialysis study population with 50 healthy subjects. The patients had significantly higher levels of IL-6 (20.17 ± 4.32 pg/mL vs. 4.02 ± 0.47 pg/mL, P < 0.001), CRP (2.01 ± 0.06 vs. 0.28 ± 0.06 mg/dL, P < 0.001), and carbonyl content (0.154 ± 0.014 nmol/mg of protein vs. 0.029 ± 0.004 nmol/mg of protein, P < 0.001) at the time of initiation of MHD therapy (Table 3). IL-10 levels were numerically lower in the healthy population compared with the ESRD population, but the difference did not reach statistical significance (1.06 ± 0.16 pg/mL vs. 2.91 ± 0.86 pg/mL, respectively, P = 0.249). When compared with diabetics without renal failure (63 ± 14 years old, 70% males), CHD patients had significantly higher levels of carbonyl content (0.154 ± 0.014 nmol/mg of protein vs. 0.021 ± 0.004 nmol/mg of protein, P < 0.001), IL-6 (20.17 ± 4.32 pg/mL vs. 6.05 ± 1.85 pg/mL, P = 0.015), and CRP (2.01 ± 0.06 mg/dL vs. 0.76 ± 0.2 mg/dL, P = 0.030) at the time of initiation of MHD therapy. Biomarkers of inflammation and oxidative stress are stable during the first year of MHD Directional changes of plasma IL-6, plasma IL-10, serum CRP, and plasma carbonyl content were analyzed over a period of 12 months. There were no significant changes in any of the study variables over the study period. This was consistent for overall trends as well as comparisons of each time point to baseline. Specifically, plasma IL-6 levels changed from 20.17 ± 4.32 pg/mL at baseline to 15.94 ± 4.22 pg/mL at 6 months, and 18.22 ± 5.27 pg/mL at 12 months’ follow-up (P = 0.983 for the trend) (Fig. 1). Serum CRP levels changed from 2.01 ± 0.06 mg/dL to 1.79 ± 0.07 mg/dL at 6 months, and to 1.25 ± 0.08 mg/dL at 12 months (P = 0.651) (Fig. 2). IL-10 concentrations were 2.91 ± 0.86 pg/mL at baseline, 0.94 ± 0.84 pg/mL at 6 months, and 1.00 ± 0.09 pg/mL at 12 months (P = 0.121). Plasma carbonyl content also decreased from 0.154 ± 0.014 nmol/kg of protein to 0.122 ± 0.013 nmol/kg of protein in the first 6 months, but dis-
played a subsequent increase to 0.162 ± 0.017 nmol/kg of protein in 12 months (P = 0.887) (Fig. 3). Changes in biomarkers of inflammation are dependent on baseline concentrations An analysis of trends over time based on initial quartiles for study variables was conducted. The trends were dependent on the initial baseline values of study variables. Figure 4 shows this analysis for IL-6. As can be seen, patients who initiated MHD with highest levels of IL-6 at baseline (Q4) experienced a significant decrease in IL-6 concentrations over the study period (58.02 ± 14.06 pg/mL to 19.15 ± 4.33 pg/mL, P = 0.036). On the contrary, patients who initiated MHD therapy in the two lowest quartiles of IL-6 (Q1 and Q2) increased their levels significantly (from 2.59 ± 0.30 pg/mL to 12.77 ± 1.30 pg/ mL for Q1, P < 0.001, and from 6.48 ± 0.38 to 20.77 ± 4.97 pg/mL for Q2, P = 0.029). Patients in the Q3 group had a trend for increase but it did not reach statistical significance (13.57 ± 1.43 pg/mL to 19.19 ± 4.65 pg/mL, P = 0.647). Similar trends were observed for IL-10, CRP, and carbonyl content. Figures 5 and 6 show these trends for CRP and carbonyl content, respectively. Most important, none of the quartiles for any study variable reached normal values at the end of the 12-month study period. For example, at the end of 1 year of MHD, levels of IL-6 were 12.77 pg/mL, 20.77 pg/mL, 19.19 pg/mL, and 19.15 pg/mL for Q1, Q2, Q3, and Q4, respectively. Levels of CRP, IL10, and carbonyl content remained similarly elevated at the end of the study period. Lack of influence of residual renal function and other clinical parameters To elucidate any potential interaction of residual renal function on the observed outcomes, we analyzed the adjusted association of residual renal function, assessed by 24-hour urine creatinine clearance, with the study variables, by mixed model procedure. No significant associations were found between the baseline creatinine clearance and the future trends of CRP (P = 0.384), IL-6 (P = 0.863), IL-10 (P = 0.721), or carbonyl content (P = 0.391). We additionally analyzed the trends in biomarkers of inflammation and oxidative stress according to the type of vascular access, concurrent illnesses, and the presence of diabetes. Our results suggested that the presence of diabetes, when analyzed separately, was associated with significant increases in carbonyl content over the study period (P = 0.040), although the trends in nondiabetics (P = 0.756) and the overall trends were not statistically significant. Analyses of the other study variables were not affected by the presence of diabetes. Hypertension, cardiovascular disease, and type of vascular access did not influence trends in any of the study variables.
Pupim et al: Inflammation and oxidative stress in MHD
30
2375
P = 0.204
25 21.95
IL-6, pg/mL
20
20.17
15
15.94
17.04
18.22
10 5 4.02 0 Control (N = 50)
Baseline 3 months (N = 50) (N = 50)
6 months 9 months 12 months (N = 46) (N = 41) (N = 41)
Fig. 1. Black bars represent changes in plasma interleukin (IL)-6 concentration at follow-up time points during 12 months of study. White bar represents one time-point measurement in healthy control patients. Bars and error bars represent mean ± SEM for each time point.
4 3.5
P = 0.651
Serum CRP, mg/dL
3 2.5 2
2.01
2.06 1.79
1.5
1.30 1
1.25
0.5 0
0.28 Control (N = 50)
Baseline (N = 50)
3 months 6 months 9 months 12 months (N = 50) (N = 46) (N = 41) (N = 41)
Correlations among study variables The associations among the four study variables were analyzed in a multivariate matrix correlation by mixedmodel procedure. There were significant positive correlations comparing CRP with IL-6 (estimate = 0.546, P = 0.012), and CRP with IL-10 (estimate = 1.893, P < 0.001). CRP correlated substantially with carbonyl content, but it did not reach statistical significance (estimate = 1.330, P = 0.599). IL-6 was negatively correlated with IL-10 but this correlation was not statistically significant (estimate = −0.067, P = 0.678). DISCUSSION The current study was designed to determine the influence, if any, of initiation of conventional MHD on cir-
Fig. 2. Black bars represent changes in serum C-reactive protein (CRP) concentration at follow-up time points during 12 months of study. White bar represents one time-point measurement in healthy control patients. Bars and error bars represent mean ± SEM for each time point.
culating levels of biomarkers of inflammation (cytokines and CRP), as well as a marker of oxidative stress (plasma protein carbonyl content). Our results showed that the levels of CRP, IL6, and carbonyl content were significantly higher in patients than healthy control patients at baseline. IL-10 was numerically higher in patients than control patients, but the difference did not reach statistical significance. In addition, there were no significant changes in any study variable over a 12-month period after initiation of MHD. After subdividing patients according to their baseline level of each study variable, highest values decreased significantly for all variables, and lowest values increased significantly for IL-6, IL-10, and carbonyl content. These findings are in accordance with the statistical occurrence known as regression to the mean, suggesting that the study intervention (i.e., hemodialysis)
2376
Pupim et al: Inflammation and oxidative stress in MHD
P = 0.887
Carbonyl content, nmol/mg protein
0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02
0.029
0 Control (N = 50)
Baseline (N = 50)
3 months (N = 50)
Q1 60
Q2
6 months 9 months 12 months (N = 46) (N = 41) (N = 41)
Q3
Fig. 3. Black bars represent changes in plasma carbonyl content at follow-up time points during 12 months of study. White bar represents one time-point measurement in healthy control patients. Bars and error bars represent mean ± SEM for each time point.
Q4
Q4, P = 0.367
55 50 IL-6 quartiles, pg/mL
45 40 35 30 25 20 15 Q3, P = 0.647 10 Q2, P = 0.029 5 Q1, P = 0.001 0 Baseline
12 months
has no significant effect on inflammation or oxidative stress in MHD patients. The [lack of] changes in biomarkers of inflammation and oxidative stress over time were independent of clinical covariates, such as presence or absence of diabetes with cardiovascular disease and/or hypertension, as well as the type of vascular access used. Acute-phase inflammation is a common feature of ESRD and it is estimated that 30% to 50% of ESRD patients have serologic evidence of an activated inflammatory response [13]. The plasma levels of the prototypical acute-phase protein CRP have been demonstrated in numerous studies to be a powerful predictor of subsequent cardiovascular events and all-cause mortality in the dialysis population [22, 23]. Plasma IL-6 levels are also increasingly being recognized as an independent predictor
Fig. 4. Changes in plasma interleukin (IL)-6 concentration over the study period according to quartiles. Each line represents one quartile.
of cardiovascular events, progression of atherosclerosis, and all-cause mortality in the dialysis population [24–28]. In this study, we also measured plasma levels of the anti-inflammatory cytokine IL-10. IL-10 is known to counter-regulate the cascade of proinflammatory cytokines, including IL-6, as part of the acute-phase reaction. The importance of IL-10 as an anti-inflammatory cytokine is emphasized by a recent study in which a lowproducing IL-10 gene polymorphism was associated with a higher cardiovascular morbidity and mortality in a cohort of MHD patients [29]. Early data suggested that plasma levels of IL-6 and IL-10 were inversely related in the hemodialysis population [30]. Other studies have demonstrated overproduction of IL-10 in the hemodialysis population [31]. A recent study using a sensitive
Pupim et al: Inflammation and oxidative stress in MHD
Q1
Q2
Q3
2377
Q4
7.0 6.5 Q4, P = 0.008 6.0
CRP quartiles, mg/dL
5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0
Q3, P = 0.973
0.5 Q2, P = 0.587 Q1, P = 0.155 0.0 Baseline
12 months
Q1
Q2
Q3
Fig. 5. Changes in serum C-reactive protein (CRP) concentration over the study period according to quartiles. Each line represents one quartile.
Q4
0.300
Carbonyl quartiles, nmol/mg protein
Q4, P = 0.003 0.250
0.200
0.150
0.100
Q3, P = 0.531
Q2, P = 0.244 Q1, P = 0.028
0.050
0.000 Baseline
12 months
technique measuring monokine production at the single cell level demonstrated that monocytes from hemodialysis patients tend to overproduce both IL-6 and IL-10 simultaneously, while healthy subjects’ monocytes produce IL-6 and IL-10 in separate monocyte subpopulations [32]. In the present study, a trend toward higher plasma levels of IL-10 was detected in dialysis patients compared
Fig. 6. Changes in plasma carbonyl protein content over the study period according to quartiles. Each line represents one quartile.
with healthy subjects with a significant relationship to CRP levels. Similar to IL-6, plasma IL-10 levels did not change significantly over the course of time after initiation of MHD therapy. The underlying etiology of the accentuated inflammatory response in uremia remains poorly understood. While multiple dialysis-related factors can contribute to
2378
Pupim et al: Inflammation and oxidative stress in MHD
maintenance of a chronic inflammatory response, recent data demonstrating that plasma IL-6 and CRP levels are already elevated in earlier stages of chronic kidney disease suggest a more proximate relationship to loss of kidney function [33]. Indeed, our results clearly show that plasma levels of cytokines and carbonyl content were significantly elevated at the time of initiation of maintenance hemodialysis. Loss of renal function may be directly associated with reduced clearance of cytokines, complement peptides, oxidants, and other proinflammatory solutes. Additionally, the hemodialysis procedure itself may contribute to a proinflammatory milieu by increasing the synthesis of fibrinogen, IL-6, and other proinflammatory peptides, likely by bioincompatibility from membrane and dialysate exposure during the extracorporeal procedure [34]. However, in this study we did not observe a significant increase in inflammatory markers during the initial 12 months of MHD. This is in contrast to a report by Haubitz et al in which significant increases in CRP levels were reported after initiation of MHD in 8 patients. While the exact reasons for this discrepancy are not clear, it should be noted that the sample size in the former study was relatively small and the patients were dialyzed with a bioincompatible membrane, which might have initiated an inflammatory response, per se [35]. An increase in oxidative stress is increasingly recognized as an important metabolic accompaniment to ESRD [36]. Reactive oxygen species are known to activate phagocytic cells and, by way of NF-jB, up-regulate the release of IL-6 and increase production of CRP [37]. Our group has recently described a linkage between increased oxidative stress and acute-phase inflammation in severely hypoalbuminemic MHD patients [38]. In the present study, we have measured plasma protein carbonyl content as a biomarker of oxidative stress status [39–41]. The importance of increased carbonyl is evidenced by their important role in the pathogenesis of atherosclerosis [42]. The present study confirms high levels of plasma protein carbonyl content in patients with advanced kidney disease, and demonstrates no significant effect of initiating MHD therapy on plasma carbonyl content over time. Similar to the data on biomarkers of inflammation, these data suggest that initiation of MHD therapy is not adequate to control oxidative stress in uremia. A corollary interpretation of the data in this study is that the hemodialysis procedure itself is not the predominant cause of the increased oxidative stress and inflammation observed in the ESRD population. Thus, hemodialysis may not accelerate atherogenesis, so much as simply not abate its development. This suggestion is supported by a growing body of evidence demonstrating that many proatherogenic metabolic alterations in uremia are already present before the initiation of dialysis therapy. Furthermore, previous work by our group and by other investigators has demonstrated that the net effect
of a single hemodialysis treatment is to improve overall redox balance [39]. Even when bioincompatible membranes, which vigorously activate the alternative pathway of complement are used, it is extraordinarily difficult to detect an increase in oxidative stress over the course of the dialysis procedure. Thus, an important potential implication of this study and other similar studies is that improvements in the proatherogenic metabolic milieu in uremia will require the removal or biologic inactivation of solutes not readily cleared by the diffusive process in hemodialysis therapy. Either alternative types of renal replacement therapy (such as hemodiafiltration or the renal tubular assist device) or pharmacologic therapies (such as antioxidants or anti-inflammatory agents) may be required to control the uremia-associated oxidative stress and inflammation. The present study in incident MHD patients should not be viewed as an attempt to quantitatively measure the potentially competing effects of hemodialysis and uremia on inflammation and oxidative stress biomarkers. Instead, it should be viewed rather as an analysis of changes observed in these parameters after definitive initiation of conventional MHD therapy in patients with advanced ESRD. While a randomized controlled trial is always preferable, such design is not applicable for the hypothesis tested in this study. The current data must also be evaluated in the context of recent publications suggesting that initiation of MHD is associated with improvements in several nutritionally and metabolically relevant markers to some extent [7, 8]. In a similar group of patients, we observed an increase in serum albumin and serum prealbumin in the setting of increased solute clearance at adequate levels [7]. The results from this study, on the other hand, show that such an effect is not observed for inflammation or oxidative stress markers, important contributors to the high cardiovascular morbidity and mortality observed in ESRD patients. Together with the context of the recent HEMO Study results, we can postulate that conventional dialysis is “inadequate” as an anti-inflammatory and antioxidant therapy.
CONCLUSION We have shown that ESRD patients have substantially increased inflammatory and oxidative stress burden at the initiation of MHD treatment with currently prescribed treatment parameters. Our results suggest that conventional MHD is not effectively treating the highly prevalent inflammation and oxidative stress in ESRD patients. The lack of effective therapy for these metabolic perturbations may contribute to an ongoing high cardiovascular morbidity and mortality. Further research is needed to elucidate the mechanistic pathways of inflammation and oxidative stress, both before and after MHD is initiated.
Pupim et al: Inflammation and oxidative stress in MHD
ACKNOWLEDGMENTS We thank Karen A. Kinne for providing administrative assistance in the production of this manuscript, Pamela Kent, R.D., for assistance with data collection, and Joel Botler, M.D., William Ervin, M.D., Carrie Brewer, and Gary Banker for their assistance in obtaining blood samples from diabetic patients. This work is supported by National Institutes of Health grants DK45604, DK62849, and HL070938; and FDA Orphan Products Development grant FD-R-000943. Dr. Pupim is partly supported by the Marilyn Charitable Trust Young Investigator Grant of the National Kidney Foundation and Vanderbilt University School of Medicine Clinician Scientist Award, as part of the Vanderbilt Physician-Scientist Development Program. Reprint requests to T. Alp Ikizler, M.D., Associate Professor of Medicine, Division of Nephrology, S-3223 MCN, Vanderbilt University Medical Center, 1161 21st Ave. South & Garland, Nashville, TN 372322372. E-mail: [email protected]
REFERENCES 1. HAMMER J, OESTERREICHER C, HAMMER K, et al: Chronic gastrointestinal symptoms in hemodialysis patients. Wien Klin Wochenschr 110:287–291, 1998 2. REESE GN, APPEL SH: Neurologic complications of renal failure. Semin Nephrol 1:137–142, 1981 3. FOLEY RN, PARFREY PS: Cardiovascular disease and mortality in ESRD. J Nephrol 11:239–245, 1998 4. BAILEY JL, MITCH WE: Metabolic acidosis as a uremic toxin. Semin Nephrol 16:160–166, 1996 5. FIASCHI E, MIONI G, MASCHIO G, et al: Calcium and phosphorus metabolism in chronic uremia. Nephron 14:163–180, 1975 6. OBRADOR GT, PEREIRA BJ: Systemic complications of chronic kidney disease. Pinpointing clinical manifestations and best management. Postgrad Med 111:115–122 7. PUPIM LB, KENT P, CAGLAR K, et al: Improvement in nutritional parameters after initiation of chronic hemodialysis. Am J Kidney Dis 40:143–151, 2002 8. MEHROTRA R, BERMAN N, ALISTWANI A, et al: Improvement of nutritional status after initiation of maintenance hemodialysis. Am J Kidney Dis 40:133–142, 2002 9. AGODOA L: United States Renal Data System (USRDS). Nefrologia 20(Suppl 5):13–16, 2000 10. EKNOYAN G, BECK GJ, CHEUNG AK, et al: Effect of dialysis dose and membrane flux in maintenance hemodialysis. N Engl J Med 347:2010–2019, 2002 11. HIMMELFARB J: Success and challenge in dialysis therapy. N Engl J Med 347:2068–2070, 2002 12. FOLEY RN, HERZOG CA, COLLINS AJ: Blood pressure and long-term mortality in United States hemodialysis patients: USRDS Waves 3 and 4 Study. Kidney Int 62:1784–1790, 2002 13. KAYSEN GA: The microinflammatory state in uremia: Causes and potential consequences. J Am Soc Nephrol 12:1549–1557, 2001 14. BECKER BN, HIMMELFARB J, HENRICH WL, et al: Reassessing the cardiac risk profile in chronic hemodialysis patients: A hypothesis on the role of oxidant stress and other non-traditional cardiac risk factors. J Am Soc Nephrol 8:475–486, 1997 15. MAGGI E, BELLAZZI R, FALASCHI F, et al: Enhanced LDL oxidation in uremic patients: An additional mechanism for accelerated atherosclerosis? Kidney Int 45:876–883, 1994 16. BUSS IH, WINTERBOURN CC: Protein carbonyl measurement by ELISA. Methods Mol Biol 186:123–128, 2002 17. WINTERBOURN CC, BUSS IH: Protein carbonyl measurement by enzyme-linked immunosorbent assay. Methods Enzymol 300:106– 111, 1999 18. DUPONT WD, PLUMER WD: PS power and sample size program available for free on the Internet. Controlled Clin Trials 18:274, 1997 19. JENNRICH RI, SCHLUCHTER MD: Unbalanced repeated-measures models with structured covariance matrices. Biometrics 42:805–820, 1986
2379
20. DIGGLE P, LIANG KY, ZEGER SL: Analysis of Longitudinal Data, Oxford, Claredon Press, 1994 21. LIANG KY, ZEGER SL: Longitudinal data analysis using generalized linear models. Biometrika 73:13–22, 1986 22. ARICI M, WALLS J: End-stage renal disease, atherosclerosis, and cardiovascular mortality: Is C-reactive protein the missing link? Kidney Int 59:407–414, 2001 23. STENVINKEL P, HEIMBURGER O, PAULTRE F, et al: Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney Int 55:1899–1911, 1999 24. KATO A, ODAMAKI M, TAKITA T, et al: Association between interleukin-6 and carotid atherosclerosis in hemodialysis patients. Kidney Int 61:1143–1152, 2002 25. MEMOLI B, MINUTOLO R, BISESTI V, et al: Changes of serum albumin and C-reactive protein are related to changes of interleukin-6 release by peripheral blood mononuclear cells in hemodialysis patients treated with different membranes. Am J Kidney Dis 39:266– 273, 2002 26. PECOITS-FILHO R, BARANY P, LINDHOLM B, et al: Interleukin-6 is an independent predictor of mortality in patients starting dialysis treatment. Nephrol Dial Transplant 17:1684–1688, 2002 27. STENVINKEL P, HEIMBURGER O, JOGESTRAND T: Elevated interleukin6 predicts progressive carotid artery atherosclerosis in dialysis patients: Association with Chlamydia pneumoniae seropositivity. Am J Kidney Dis 39:274–282, 2002 28. KIMMEL PL, PHILLIPS TM, SIMMENS SJ, et al: Immunologic function and survival in hemodialysis patients. Kidney Int 54:236–244, 1998 29. GIRNDT M, KAUL H, SESTER U, et al: Anti-inflammatory interleukin10 genotype protects dialysis patients from cardiovascular events. Kidney Int 62:949–955, 2002 30. GIRNDT M, KOHLER H, SCHIEDHELM-WEICK E, et al: Production of interleukin-6, tumor necrosis factor alpha and interleukin-10 in vitro correlates with the clinical immune defect in chronic hemodialysis patients. Kidney Int 47:559–565, 1995 31. MORITA Y, YAMAMURA M, KASHIHARA N, et al: Increased production of interleukin-10 and inflammatory cytokines in blood monocytes of hemodialysis patients. Res Commun Mol Pathol Pharmacol 98:19– 33, 1997 32. GIRNDT M, SESTER U, KAUL H, et al: Production of proinflammatory and regulatory monokines in hemodialysis patients shown at a single-cell level. J Am Soc Nephrol 9:1689–1696, 1998 33. OBERG BP, MCMENAMIN E, LUCAS FL, et al: Increased prevalence of oxidant stress and inflammation in patients with moderate to severe chronic kidney disease. Kidney Int 65:1009–1016, 2004 34. CAGLAR K, PENG Y, PUPIM LB, et al: Inflammatory signals associated with hemodialysis. Kidney Int 62:1408–1416, 2002 35. HAUBITZ M, SCHULZE M, KOCH KM: Increase of C-reactive protein serum values following haemodialysis. Nephrol Dial Transplant 5:500–503, 1990 36. HIMMELFARB J, STENVINKEL P, IKIZLER TA, et al: The elephant in uremia: Oxidant stress as a unifying concept of cardiovascular disease in uremia. Kidney Int 62:1524–1538, 2002 37. BARNES PJ, KARIN M: Nuclear factor-kappaB: A pivotal transcription factor in chronic inflammatory diseases. N Engl J Med 336:1066– 1071, 1997 38. DANIELSKI M, IKIZLER TA, MCMONAGLE E, et al: Linkage of hypoalbuminemia, inflammation, and oxidative stress in patients receiving maintenance hemodialysis therapy. Am J Kidney Dis 42:286–294, 2003 39. HIMMELFARB J, MCMONAGLE E, MCMENAMIN E: Plasma protein thiol oxidation and carbonyl formation in chronic renal failure. Kidney International 58:2571–2578, 2000 40. MIYATA T, VAN YPERSELE DE STRIHOU C, KUROKAWA K, et al: Alterations in nonenzymatic biochemistry in uremia: Origin and significance of “carbonyl stress” in long-term uremic complications. Kidney Int 55:389–399, 1999 41. WEISS MF, ERHARD P, KADER-ATTIA FA, et al: Mechanisms for the formation of glycoxidation products in end-stage renal disease. Kidney Int 57:2571–2585, 2000 42. SAKATA N, IMANAGA Y, MENG J, et al: Increased advanced glycation end products in atherosclerotic lesions of patients with end-stage renal disease. Atherosclerosis 142:67–77, 1999