Abstract
The prevalence of chronic inflammation is high in dialysis patients. Moreover, it is associated with an increased mortality risk, yet the origin of chronic inflammation in dialysis patients remains unclear. The aim of this study was to determine the effect of a hemodialysis session (HD) on C-reactive protein (CRP) levels and to study the relation with survival. As part of a large, prospective, multicenter study in the Netherlands (Netherlands Cooperative Study on the Adequacy of Dialysis), patients who were started on dialysis treatment between September 1997 and May 1999 were included. Demographic data, clinical data, and serum samples were collected at regularly timed intervals. From this cohort, a random sample of patients was taken. CRP levels were determined before and after an HD session and before the next session. Date of death or censoring was recorded until September 2002. A total of 186 HD patients were included. Mean age was 65 yr (SD, 13); 56% were male. A total of 71 patients had a CRP level below the detection limit (3 mg/L), 68 patients showed no increase in CRP during an HD session (no-increase group), and 47 (25%) patients showed an increase in CRP level during an HD session (increase-group). No statistically difference in mean CRP levels before the dialysis session was found between the increase group (22.3 mg/L) and the no-increase group (19.4 mg/L). In the subsequent interdialytic period, CRP levels returned to the levels of the initial CRP value. Two-year survival was 44% in the increase group and 66% in the no-increase group (P = 0.09). Independent of CRP level before the session and adjusted for age, comorbidity, nutritional status, and primary kidney disease, a raise of 1 mg/L CRP during a session was associated with a 9% increased mortality risk (adjusted hazard ratio, 1.09; 95% CI, 1.02 to 1.16). The present study showed an increase in CRP level during a single dialysis session in 25% of the patients; during the succeeding interdialytic period, CRP level returned to its original value. More important, however, an increase in CRP level during an HD session was independently associated with a higher mortality risk.
The prevalence of chronic inflammation, as reflected by increased levels of proinflammatory cytokines or acute-phase proteins, such as C-reactive protein (CRP), is high in dialysis patients. The reported prevalence of inflammation varies between 35 and 65% (1). Moreover, elevated levels of inflammatory factors have been associated with an increased risk of mortality and morbidity (2–5⇓⇓⇓). A single determination of CRP level was a powerful indicator of mortality even after 4 to 5 yr follow-up. In a study by Iseki et al. (6) a 3.5-times higher mortality risk was observed for dialysis patients with a CRP level >10 mg/L after 5 yr of follow-up. In a German study, a nearly two-times higher mortality risk was found for patients with an initial CRP level >8 mg/L (7).
So far, the origin of inflammation in patients with chronic renal disease remains unclear. It has been suggested that some dialysis-related alteration in the immune and host-defense system may be relevant, as they seem to correlate to the high production of proinflammatory cytokines (8). Moreover, the type of dialysis membrane has been suggested to play a role (9) as well as time on dialysis, as it correlated negatively with cytokine release (10). Next, the hemodialysis (HD) procedure itself may be involved. Several mechanisms have been proposed to induce cytokine production during HD: the generation of complement fractions as a result of plasma protein-membrane contact, the backfiltration of contaminated dialysate to the blood compartment, and the direct contact of blood cells with the dialysis membrane (11,12⇓). Some studies observed an inflammatory response of the HD procedure. In a study by Malaponte et al. (10), the concentration of cytokines was higher at the end of a dialysis session compared with the beginning of the session, and in a study by Caglar et al. (13), an increase in cytokine concentration was noticed in a 2-h period after the HD session. However, it remains to be established which factors of the dialysis procedure may be responsible for the inflammatory response. Furthermore, the clinical impact of the change in inflammatory factors during a dialysis session is still unclear. We therefore studied the impact of an HD session on an inflammatory factor, CRP; we studied factors associated with a change in CRP during a dialysis session and studied the relation of a change in CRP with survival in the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD), a large, multicenter, prospective study of Dutch dialysis patients.
Materials and Methods
Patients
All new ESRD patients from 38 Dutch dialysis units were consecutively invited to participate in the study. These patients participated in NECOSAD. The aim of this study was to monitor the quality and adequacy of dialysis treatment in the Netherlands. Special attention was paid to the determination of risk factors for poor outcome. The role of chronic inflammation as a risk factor is one of the aims of this study. As it is well known that the dialysis procedure itself may result in an inflammatory response, the first step was to study the influence of a dialysis session on CRP concentration in serum. Patients had to be 18 yr or older with dialysis as the first renal replacement therapy. The inclusion period was between January 1997 and May 1999. All medical ethics committees approved the study, and all patients gave informed consent before inclusion.
Procedures
Demographic data and clinical data were collected and serum samples were stored centrally at the start of dialysis treatment, 3 and 6 mo after the start, and after that every 6 mo onward. The following data were collected at the start: age, gender, primary kidney disease, and comorbidity. Primary kidney disease was classified according to the codes of the European Renal Association-European Dialysis and Transplantation Association. Comorbidity was defined according to the risk criteria of Khan et al. (14). This classification relies on a combination of the number of comorbidities and age, resulting in three risk categories: low, medium, and high.
At each time of measurement, the following data were collected: A plasma sample and a sample from the urine collection; urine was sampled during the preceding interdialytic period. Plasma creatinine and plasma urea levels were determined. Urea and creatinine were analyzed in the urine sample. Renal function was expressed as GFR, calculated as the mean of creatinine and urea clearance, corrected for body surface area. Nutritional status was scored on the seven-point scale of the subjective global assessment, a standardized method based on clinical judgment (15). HD-related data were collected: Blood and dialysate flow rate [blood flow (ml/min) {Qb} and dialysate flow (ml/min) {Qd}], needle type, membrane type, treatment with hemofiltration or hemodiafiltration, total hours of dialysis per week, and dialysis dose, calculated as Kt/Vurea according to the second-generation Daugirdas formula (16). Finally, at each time of measurement, a plasma sample was collected and stored centrally. HD treatment was predominantly in-center HD; the vast majority of chronic dialysis patients in the Netherlands are treated with an arteriovenous fistula, and patients were treated according to usual local care. No dialysis reuse was used.
A random sample of the centrally stored plasma samples was taken from this cohort. This resulted in a cross-sectional sample of HD patients. Only one time of measurement per patient was selected. CRP levels and serum albumin levels were determined before and after the selected dialysis session (index session) and again, when available, before the next session. The index session had to be at least 3 mo after the start of HD treatment. CRP levels were determined by an immunoturbidimetric assay. Detection level of CRP was 3 mg/L. Inter- and intra-assay variability were for within-run coefficient of variation (CV) of 1.8%, run-to-run CV of 1.7%, and day-to-day CV of 2.8%. The normal range expected was <10 mg/L; linearity was between 3.0 and 250 mg/L. Serum albumin levels were determined by immunonephelometry.
The date and time of the index session and the date of death, when applicable, were documented. Surviving patients were considered to have their survival times censored at the date of leaving the study as a result of a renal transplant, a transfer to a nonparticipating unit, or the end of the follow-up period in September 2002.
Statistical Analyses
Differences between groups were tested by ANOVA and t tests for continuous variables when normally distributed. If not, then they were tested with Kruskal-Wallis or Wilcoxon rank test. For categorical data, the χ2 test was applied. A multivariate linear regression model was build with a stepwise selection procedure (entry level, 0.40; removal level, 0.10) to determine the factors related to a change in CRP level during a dialysis session. Factors presented to the model were age, gender, comorbidity, primary kidney disease, time on dialysis before index session, duration of index session, total hours of dialysis per week, Qd, Qb, needle type, artificial kidney type, nutritional status, hemoglobin, cholesterol level, and medication (e.g., epo, antihypertensives, statins).
CRP levels after dialysis were corrected for dialysis-induced changes in blood volume by multiplying CRP after dialysis with the ratio between serum albumin before and after dialysis (CRPadj) (17). Serum albumin levels were taken to adjust for changes in blood volume, as albumin is not cleared during a dialysis session or broken down in such a short time. In case the blood volume would not change during a dialysis session, serum albumin levels would remain similar. Therefore, any change in albumin level indicates a change in blood volume.
CRP levels were rounded off to the nearest integer by our laboratory. After adjustment for the change in blood volume, we obtained for the CRP level after the index session a level with digits. Consequently, a difference of <0.5 mg/L could be due to the rounding off of the CRP level before the index session, instead of a true difference. Therefore, we considered an increase of <0.5 mg/L as no increase.
Unadjusted survival curves for the total time of follow-up were estimated using the Kaplan-Meier method. Multivariate Cox proportional hazards regression analyses were performed to calculate the hazard ratios for the change in CRP level during a dialysis session. In all models, adjustments for initial CRP level and change in blood volume were made. In the first model, no further adjustments were made to determine the effect of change in CRP levels. In the second model, adjustments were made for factors that are known to influence survival: age, primary kidney disease, comorbidity, nutritional status, and serum albumin level.
Results
A total of 186 patients were selected. Mean age was 65.0 yr, 56% were male, and mean time on dialysis was 9.0 mo. The average duration of the index session was 3.8 h. Seventy-one (38%) patients had a CRP concentration below detection level (≤3 mg/L), 68 patients showed an increase <0.5 mg/L during the index HD session (no-increase group), and 47 (25%) patients showed an increase in CRP level of at least 0.5 mg/L (increase group; Table 1). Patients with a CRP concentration below detection level were younger, had a higher serum albumin level, had less comorbidity, had relatively more glomerulonephritis as primary kidney disease, and were less often malnourished compared with the patients with a CRP level >3 mg/L. Patients with a CRP level below detection limit are refrained from further analyses, because their CRP level and thus the effect of a dialysis session on CRP could not be determined.
Table 1. Clinical data of the patientsa
In the no-increase group, CRP level before the index session was 19.4 mg/L (21.9), and in the increase group, this concentration was 22.3 mg/L (22.1). Median change in CRP during the index session was 2.3 in the increase group, and −0.6 in the no-increase group (P < 0.05). CRP level just before the next session was not available for all patients; in the increase group, this was determined in 40 of the 47 patients. In these 40 patients, CRP level just before the next session (24.6 mg/L) was similar to the level of before the index session: CRP before index session, 24.2 mg/L; CRPadj after index session, 27.1 mg/L. Patients in whom CRP increased during the index session were older compared with the no-increase group (P < 0.05; Table 1).
After adjusting for the change in blood volume and initial CRP level, patients with a higher dialysis flow rate, with lower cholesterol levels, dialyzing with a double-needle system, and older patients, had a significantly higher increase in CRP level during a dialysis session (Table 2). Each year older was associated with an increase in CRP level of 0.07 mg/L, and each increase of 10 ml/min in dialysate flow rate was associated with an increase of 0.20 mg/L in CRP during an HD session. No associations between a change in CRP level and dialysis parameters such as type of artificial kidney, blood flow rate, duration of the index session, time on dialysis before index session, renal function, dialysis dose, or the use of angiotensin-converting enzyme (ACE) inhibitors was noticed. Moreover, no effect of treating center was observed.
Table 2. Multivariate analyses of factors that influence the change in CRP level during a dialysis session, adjusted for a change in blood volume (N = 115 patients) limited to patients with a CRP level above detection level
At the end of the follow-up period, 61 of 115 patients died; median follow-up time was 22.5 mo. With 115 patients above detection level, we had 70% power to detect a two-times worse survival for the patients whose CRP increased during an HD session compared with patients whose CRP did not increase. Crude 2-yr survival in the no-increase group was 65% and in the increase group was 44%. A survival plot of the two groups is shown in Figure 1; overall mortality differed not significantly (log-rank test statistics P = 0.09). Causes of death were 34% cardiovascular, 18% infection, 18% discontinuation of treatment, and 30% other. No differences in cause of death between increase and no-increase groups was observed (data not shown).
Figure 1. Survival of patients in whom C-reactive protein (CRP) level increased by at least 0.5 mg/L during the index session (increase group) compared with the patients of the no-increase group (CRP level increased by <0.5 mg/L during the index session), limited to the patients with a CRP value above detection level (n = 115; P = 0.09).
In Table 3, the adjusted hazard ratios for a change in CRP level during a dialysis session are presented. After adjusting for initial CRP level and change in blood volume, an increase of 1 mg/L during a dialysis session was associated with an 8% increased mortality risk (model 1). After adjustment for the aforementioned factors combined with factors known to influence mortality—age, primary kidney disease, comorbidity, serum albumin concentration, and nutritional status—the adjusted hazard ratio for a change in CRP level became 1.09 (95% CI, 1.02 to 1.16). So, both the absolute CRP level before the dialysis session and the change in CRP during the session are independent predictors of survival.
Table 3. Adjusted hazard ratios for a change in CRP level during the index session limited to the patients with a CRP value above detection level (N = 115)a
Without adjustment for the dialysis-induced change in blood volume, 40% of the patients (instead of 25%) would have in increase in CRP level ≥0.5 mg/L. Repeating the Cox regression analysis without correcting CRP level for a change in blood volume resulted in findings similar to CRP levels with a correction for blood volume (adjusted hazards ratio, 1.11 [95% CI, 1.05 to 1.17]).
Discussion
This prospective multicenter study on the impact of an HD session on CRP level showed an increase in CRP level in 25% of the patients during a dialysis session. Moreover, independent of the predialysis CRP level, the change in CRP level during an HD session was associated with an increased mortality risk; an increase of 1 mg/L in CRP level was associated with a 9% raised mortality risk. This suggests that a patient with an increase of 3 mg/L in CRP level during an HD session, that is the average elevation in CRP of the increase group, will have a raised mortality risk of 30%.
CRP is a marker of inflammation, which involves a number of complex processes that can be induced by any trauma or infection. As soon as the inflammatory stimulus has been eliminated, the CRP level declines. The increase in CRP level observed in the present study could have been the result of such acute trauma or infection instead of a response to a dialysis session. However, this does not seem likely as the CRP level in the subsequent interdialytic period returned to the level of the initial CRP value. So, the fast increase seems to be related to the HD session itself. The correlation of an increase in CRP levels with the dialysate flow rate during dialysis is in favor of this hypothesis. Moreover, even after 5 yr of follow-up, the increase in CRP during the HD session was associated with an increased mortality risk. This makes it less likely that the increase was related to an acute trauma or infection.
The question that could be put forward is whether CRP level is able to increase in just 4 h. Plasma CRP is produced by hepatocytes, predominantly under transcriptional control by IL-6, although other sites of local CRP synthesis and possibly secretion have been suggested (18). De novo hepatic synthesis starts rapidly after a single stimulus; a raise in serum concentrations to >5 mg/L was noticed after 6 h (18). Such a rapid increase was observed, for example, during operations; after 290 min of operation time in patients with colorectal cancer, an average increase of 15.4 mg/L was observed (19), whereas in patients who underwent surgery for hernia, a mean increase of 4.4 mg/L was found 2 h after the operation (20). In addition, an increase in CRP level during an HD session was observed in some recent studies. One study reported an increase of 6.7 mg/L after 4 h of HD (21). Another study noticed an increase of 4 mg/L or more in 42 (36%) of 118 patients (22). These studies reported a higher percentage of patients with an increase in CRP level compared with the present study, yet no adjustments were made for a change in blood volume. Without adjusting for this change, we found an increase in CRP level more in concordance with these previous studies: an increase in 40% of the patients. In a study from the United States, no significant increase in CRP level was observed in the period of 30 min before an HD session and 2 h after the session (13). However, an increase in cytokine level was reported in this study.
Age, dialysate flow rate, cholesterol level, the use of a double-needle system, and initial CRP level were predictive for the change in CRP level during a dialysis session. We could not demonstrate an effect of the use of ACE inhibitors, dialyzer type, or duration of the dialysis session on the increase in CRP level. Van Tellingen et al. (23) also found no effect of dialyzer type, contamination, material, or flux characteristics; only clinical events, primarily infections, were predictive for the plasma CRP level. In a study by Panichi et al. (24), an increase in CRP level was observed in dialysis patients who were exposed to contaminated dialysate, suggesting that backfiltration may induce chronic inflammation. We do not know the quality of the dialysate used in this multicenter study. However, we could not find a center effect, making it less likely that contaminated dialysate caused the increase in CRP level.
Results from the present study demonstrated that CRP level itself and an increase in CRP level during a dialysis session are independently associated with an increased mortality risk. The higher the baseline CRP and the larger the increase in CRP during a dialysis session, the larger the mortality risk. Several mechanisms may be involved. It has been suggested that the rapidly developing atherosclerosis in ESRD patients could be caused by a synergism of different mechanisms such as malnutrition, inflammation, and oxidative stress (3,4⇓). Results from a study by Kaysen et al. (25) point in the same direction, as inflammation and dietary protein intake exerted competing effects on serum albumin. Moreover, Zimmermann et al. (26) demonstrated an increased risk of cardiovascular death in ESRD patients with elevated CRP levels. In addition, elevated CRP levels associated with an HD session were predictive for proatherogenic reactivity and cardiac hypertrophy (22). Even among apparently healthy men, baseline CRP levels predicted future risk of developing symptomatic peripheral arterial disease (27). Some therapeutic approaches have been suggested, such as antioxidant therapy and ACE inhibitors (28). In addition, there is some evidence that statins might be able to reduce CRP levels (29).
The present study showed an increase in CRP level during a single HD session in 25% of the patients. This increase was associated with a higher mortality risk independent of the CRP level before the dialysis session: an increase of 3 mg/L during a dialysis session was associated with a 30% higher risk of death. The impact on patient outcome underlines the importance of understanding which factors might cause this inflammatory response, yet so far, the mechanisms still remain unclear, and therapeutic approaches are not confirmed.
Acknowledgments
This work was supported by grants from The Dutch Kidney Foundation (E.018) and the Dutch National Health Insurance Board (OG97/005).
The nursing staff of the 38 different dialysis units, who collected most of the data, are gratefully acknowledged for their assistance. Moreover, we thank the staff of the NECOSAD trial office for assistance in the logistics of this study.
Members of the NECOSAD Study Group: A.J. Apperloo, J.N.M. Barendregt, R.J. Birnie, M. Boekhout, W.H. Boer, H.R. Büller, F.T.H. de Charro, C.J. Doorenbos, W.J. Fagel, G.W. Feith, C.F.M. Franssen, L.A.M. Frenken, W. Geerlings, P.G.G. Gerlag, J.P.M.C. Gorgels, W. Grave, R.M. Huisman, K.J. Jager, M.A. Jansen, K. Jie, W.A.H. Koning-Mulder, M.I. Koolen, T.K. Kremer Hovinga, A.T.J. Lavrijssen, A.W. Mulder, K.J. Parlevliet, J.B. Rosman, E.F.H. van Bommel, W.T. van Dorp, A. van Es, J.A.C.A. van Geelen, J.L.C.M. van Saase, M.J.M. Schonk, M.M.J. Schuurmans, P. Stevens, J.G.P. Tijssen, R.M. Valentijn, G. Vastenburg, C.A. Verburg, V.M.C. Verstappen, H.H. Vincent, and P. Vos.
- © 2004 American Society of Nephrology
References
