Skip Navigation



NDT Advance Access published online on July 10, 2007
Nephrology Dialysis Transplantation, doi:10.1093/ndt/gfm381
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
22/11/3277    most recent
gfm381v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Grootendorst, D. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Grootendorst, D. C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2007]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Excellent agreement between C-reactive protein measurement methods in end-stage renal disease patients—no additional power for mortality prediction with high-sensitivity CRP

Diana C. Grootendorst1, Dinanda J. de Jager1, Vincent M. Brandenburg2, Elisabeth W. Boeschoten3, Raymond T. Krediet4, Friedo W. Dekker1 and The NECOSAD Study Group5

1Department of Clinical Epidemiology, Leiden University Medical Center, 2300 RC Leiden, The Netherlands, 2Department of Nephrology and Clinical Immunology, University Hospital of the RWTH Aachen, Aachen, Germany, 3Hans Mak Insitute, Naarden and 4Department of Nephrology, Amsterdam Medical Center, Amsterdam, The Netherlands and 5The NECOSAD Study Group consisting of: Apperloo AJ, Bijlsma JA, Boekhout M, Boer WH, van der Boog PJM, Büller HR, van Buren M, de Charro FTh, Doorenbos CJ, van den Dorpel MA, van Es A, Fagel WJ, Feith GW, de Fijter CWH, Frenken LAM, van Geelen JACA, Gerlag PGG, Gorgels JPMC, Grave W, Huisman RM, Jager KJ, Jie K, Koning-Mulder WAH, Koolen MI, Kremer Hovinga TK, Lavrijssen ATJ, Luik AJ, van der Meulen J, Parlevliet KJ, Raasveld MHM, van der Sande FM, Schonck MJM, Schuurmans MMJ, Siegert CEH, Stegeman CA, Stevens P, Thijssen JGP, Valentijn RM, Vastenburg GH, Verburgh CA, Vincent HH and Vos PF

Correspondence and offprint requests to: Diana C. Grootendorst, Leiden University Medical Center, Department of Clinical Epidemiology C9-29, P.O. Box 9600 2300 RC Leiden, The Netherlands. Email: d.c.grootendorst{at}lumc.nl



   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Acknowledgements
 References
 
Background. The conventional method for C-reactive protein (CRP) measurement is an immunoturbidimetric assay (imCRP, detection limit ≥3 mg/l). However, high-sensitivity CRP (hsCRP, detection limit >0.1 mg/l) has been advocated as preferable biomarker for cardiovascular risk assessment. The aim of this study was to determine agreement between imCRP and hsCRP in end-stage renal disease (ESRD) patients, and to examine whether the association between CRP and mortality is comparable when using imCRP or hsCRP.

Methods. Patients from a prospective follow-up study among incident ESRD patients (NECOSAD) with serum CRP available at 3 months of follow-up were included [n = 840, 60% male, mean (SD) age 59 (15) years]. Agreement between imCRP and hsCRP was determined by intraclass correlation coefficient (ICC) and by Cohen's kappa ({kappa}) for CRP dichotomized to the presence (CRP >10 mg/l) or absence of systemic inflammation. The association between CRP and mortality was determined by Cox regression analysis and c-statistic.

Results. ICC between imCRP and hsCRP was 0.78, which improved to 0.86 after correction for systematic differences between measurement methods. Systemic inflammation was present in 28.2% and absent in 67.6% of patients according to both methods (discordant in 4.2%), resulting in good agreement between the two methods ({kappa} = 0.90). Patients with systemic inflammation had a significantly increased mortality risk compared with patients without systemic inflammation [HRim,adj = 1.49 (95%CI 1.14–1.93) and HRhs,adj = 1.53 (1.18–2.0)]. Predictive capacity of mortality was similar for both CRP methods [c-statisticadj 0.83 (0.79–0.86)].

Conclusion. The agreement between imCRP and hsCRP in patients with ESRD is very good. Furthermore, the association between CRP and mortality in ESRD patients is similar when using imCRP and hsCRP. These data suggest that there is no need to use a high-sensitivity method for the determination of inflammatory status in ESRD patients.

Keywords: dialysis; end-stage renal disease; inflammation; mortality; reliability; survival analysis



   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Acknowledgements
 References
 
C-reactive protein (CRP) is considered as a biomarker of chronic, systemic inflammation [1], as well as a mediator of atherosclerosis [2]. The conventional method for CRP measurement is by an immunoturbidimetric assay (imCRP). This method is available in routine testing laboratories and is suitable for measurement of CRP concentrations during infection. However, it is regarded as a relatively insensitive method for determination of (changes in) concentrations within the normal range (<10 mg/l) [3]. Therefore, hsCRP assays have been developed which can measure CRP at concentrations as low as 0.1 mg/l. Measurement of hsCRP can demonstrate subclinical inflammatory states, which may reflect vascular inflammation [4]. However, hsCRP measurements may not routinely be available and, in general, produce higher costs compared with imCRP. Nevertheless, official guidelines for non-renal patient populations recommend the use of hsCRP, since this is currently regarded as the best available biomarker for assessment of systemic inflammation [5].

In primarily non-renal patient populations with coronary artery disease or myocardial infarction, high levels of hsCRP have been associated with an increased risk for cardiovascular events [6–8]. Recently, hsCRP has been proposed for risk assessment in individuals at intermediate risk of coronary heart disease [9]. Indeed, cardiovascular risk prediction from global risk prediction models improves when hsCRP is used on top of more traditional parameters such as total high-density lipoprotein (HDL), low-density lipoprotein (LDL), age, smoking and blood pressure [10]. In the Women's Health Study, subtle differences in hsCRP provided important prognostic information upon cardiovascular risk [11].

Patients with end-stage renal disease (ESRD) are at a high risk to develop cardiovascular disease and cardiovascular mortality is an important cause of death [12]. Part of their increased risk is caused by highly prevalent cardiovascular risk factors such as hypertension, diabetes and anaemia. Inflammatory processes have recently emerged as additional cardiovascular risk factors of great importance in ESRD patients [13]. Twenty to 50% of ESRD patients have increased CRP levels [14], consistent with the presence of systemic inflammation. Previous prospective studies in ESRD patients have shown that high CRP levels are associated with increased cardiovascular mortality [15–18], probably through the effect of chronicly elevated CRP on atherogenesis. Therefore, also in ESRD patients, CRP measurement is considered an important tool for cardiovascular risk assessment [19].

At present, it is unknown whether hsCRP rather than imCRP should be the preferred marker in the risk assessment of ESRD patients. The need for hsCRP measurement could be particularly of interest due to the high incidence of cardiovascular events and the high prevalence of systemic inflammation in these patients. The aim of this study was to determine the agreement between imCRP and hsCRP in incident patients with ESRD. Furthermore, it was examined whether the association between CRP and mortality in ESRD patients was consistent when using imCRP vs hsCRP.



   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Acknowledgements
 References
 
Patients
The study population was part of the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD), a multi-centre prospective follow-up study among ESRD patients. Incident dialysis patients from 38 dialysis centres in the Netherlands were enrolled from 1997. Inclusion criteria were age above 18 years and starting renal replacement therapy for the first time. Additional criteria for the present analysis included enrolment prior to 1 October 2001, at least 90 days of follow-up and having stored serum available at 3 months after start of dialysis for determination of both imCRP and hsCRP. The study was approved by the Medical Ethics Committees of all participating hospitals. All patients gave written informed consent.

Design
At the start of dialysis (inclusion) patient demographics including gender, age, height, weight, body mass index (BMI), primary kidney disease (PKD), comorbidities and medication use were recorded. Patients were followed-up at 3 and 6 months after start of dialysis, and thereafter every 6 months until time of death, transplantation, 1 January 2006 or a maximum follow-up of 5 years, whichever was earliest. At 3 months after start of dialysis, blood was drawn and serum samples were stored locally at –20°C. All samples were transferred to –80°C until measurement of CRP in a central laboratory. Recorded comorbidities were used to calculate a comorbidity score according to Khan et al. [20].

CRP
Serum CRP was measured by immunoturbidimetry (imCRP, low-sensitive method) and high sensitivity method (hsCRP) in sera obtained at 3 months after start of dialysis. Serum analyses for imCRP were performed by means of an immunoturbidimetric test (Roche; detection limit ≥3 mg/l, precision 1 mg/l). The between-assay coefficient of variation (CV) was 1.8%. Within-run CV was 1.8%, run-to-run CV 1.7% and day-to-day CV 2.8%. Serum analyses for hsCRP were performed by means of particle-enhanced immunonephelometry using a standard CardioPhase hsCRP for BNII (Dade Behring Holding GmbH, Liederbach, Germany. Detection limit 0.1 mg/l, precision 0.1 mg/l) [21]. CRPI or CRPII assay protocols were used when appropriate. Interday precision controls revealed CV <9%. Testing for precision with a 20-fold repetitive measurement of a single sample showed a CV of 7%. The lowest possible CRP value in patients was either 3 mg/l (imCRP) or 0.1 mg/l (hsCRP).

CRP values were used to classify patients into groups with systemic inflammation present (CRP>10 mg/l) or absent (CRP≤10 mg/l). The cut-off point to define systemic inflammation has been used previously [15,22] and corresponds well to findings that 90% of all adults in a large population-based study displayed CRP levels below that threshold [23].

Statistical analysis
Intraclass correlation coefficients (ICCs) were calculated following an analysis of variance (ANOVA). The ICC was subsequently computed by using the formula: [mean square between groups (MSB)–mean square within groups (MSW)/(MSB + MSW)]. In addition, in order to correct for systematic differences between measurement methods, ICC was also calculated by variance components. Bland–Altman analysis was performed to assess the limits of agreement between the two measurement methods of CRP [24]. Cohen's kappa statistic ({kappa}) was calculated for classification of patients into inflamed and non-inflamed groups, at different cut-off-points of CRP. Both indicators of agreement (ICC and {kappa}) can have a maximum value of 1, indicating perfect agreement. Sensitivity, specificity, positive predictive value and negative predictive value of imCRP were computed not only for the cut-off point of 10 mg/l CRP for defining inflammation, but also for other cut-off points used by others [25–27]. Survival of patients with and without inflammation (using the cut-off point of 10 mg/l CRP) as determined by imCRP or hsCRP was analysed by Kaplan–Meier and Cox-regression analysis. Moreover, the association of CRP with mortality was further explored with patients categorized according to tertiles of CRP, but also using CRP on a continuous scale. Kaplan–Meier and Cox regression results were compared between imCRP and hsCRP. The concordance c-statistic was calculated in order to assess the predictive capacity of imCRP and hsCRP with respect to mortality [28]. This analysis is comparable to a receiver operating characteristic analysis, but allows adjustment for confounders. CRP on a continuous scale and mortality (yes/no) are modelled, together with confounders. The area under the curve represents the predictive capacity for mortality of CRP and can vary between 0.5 (the model's prediction is no better than chance) and 1. Finally, influence of storage time on serum CRP level was investigated. Patients were grouped according to storage time of serum prior to measurement of CRP. Levels of imCRP were compared with those of hsCRP within storage time periods of 1 year by ANOVA. Agreement ({kappa}) between imCRP and hsCRP was assessed for patients in whom both CRP measurements were performed in serum stored <5 years or ≥5 years. All analyses were performed in the Statistical Package for Social Sciences (SPSS, version 12).



   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Acknowledgements
 References
 
On 1 October 2001, a total of 1533 patients starting dialysis treatment were enrolled in the NECOSAD study. Of those, 68 patients were excluded because follow-up was <90 days. Serum samples were available at 3 months after start of dialysis in 844 patients. Demographic data were not different between the patients who were included in the analysis and those who were not (data not shown). In two patients, hsCRP could not be determined, whereas in two other patients the imCRP and hsCRP measurements were extreme and outlying (one with consistent but extremely high CRP levels, one with extremely differing values for imCRP and hsCRP). Since these values (or the combination of imCRP with hsCRP) seemed biologically impossible, these two patients were excluded from further analysis. A total of 840 participants were, therefore, available for the present analysis (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Patient characteristics

 
Levels of CRP
Median (minimum, maximum) CRP levels were significantly higher when measured by imCRP [5.0 (1.0–376.0)] as compared with hsCRP [4.5 (0.14–375.0), P < 0.0001]. When measured by imCRP, 40.4% of patients had a CRP level of 3 mg/l, which is the detection limit of the immunoturbidimetric assay. The hsCRP method showed that in fact these patients had CRP values as low as 0.1 mg/l (Figure 1). Three patients had imCRP values below the detection limit of the low-sensitivity method. It was not possible to determine whether the serum samples of these patients were diluted prior to measurement of imCRP and thus true values, or that the data were incorrect. Therefore, these patients were excluded from further analyses.


Figure 1
View larger version (30K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Distribution of CRP as measured by low-sensitive (immunoturbidimetry; imCRP) and high-sensitive (hsCRP) method in incident dialysis patients. Lines indicate geometric mean. Note: imCRP values at (and below) 3 mg/ml comprise 40.4% of the total cohort of patients.

 
Agreement
The correlation between imCRP and hsCRP was excellent (Spearman rank rS = 0.95, P < 0.001, Figure 2). The mean difference between the two measurement methods, as calculated by Bland–Altman analysis, was 1.8 mg/l (Figure 3). In general, imCRP levels were higher than hsCRP, indicating a systematic difference between the two methods. The limits of agreement for the two methods were between –5.8 mg/l and 2.2 mg/l.


Figure 2
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. Correlation between imCRP and hsCRP as assessed by Spearman rank correlation in incident dialysis patients.

 

Figure 3
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. Bland–Altman plot of the individual difference between hsCRP and imCRP against the mean of the two CRP measurements in incident dialysis patients. Dashed line represents the mean difference between the measurement methods. Lines (in grey) indicate the upper (mean difference+2 SD) and lower (mean difference–2 SD) limit of agreement.

 
ICC for imCRP vs hsCRP was 0.78. Following correction for systematic differences between the two methods, ICC was 0.86. When focusing on the range of CRP values which can be measured adequately by both methods (>3 mg/l) ICC improved to 0.98 and 0.99, respectively.

Patients were classified as having systemic inflammation (CRP>10 mg/l) or not having systemic inflammation according to both CRP measurement methods (Table 2). Only in 35 of 837 patients (4.2%), the two measurement methods were not concordant in this classification. Therefore, the agreement between measurement methods for classification of inflammation was good ({kappa} = 0.901). Sensitivity, specificity, positive predictive value and negative predictive value for imCRP compared with hsCRP were very good when using >10 mg/l of CRP as cut-off point for the definition of systemic inflammation, but also when using other cut-off-points such as CRP > 3.4 mg/l [26], >16.8 mg/l [27] or >50 mg/l [25] (Table 3).


View this table:
[in this window]
[in a new window]

  		Table 2. The number of patients with (CRP > 10 mg/l) and without (CRP ≤ 10 mg/l) systemic inflammation according to both im- and hsCRP

 

View this table:
[in this window]
[in a new window]

  		Table 3. Cohen's kappa, sensitivity, specificity, positive predictive value and negative predictive value for all serum imCRP values and values >3 mg/l at different cut-off-points, with hsCRP as golden standard

 
Association with mortality
During follow-up, 38.2% of patients died (incidence 11.8 per 100 person years) of whom 46.9% died of cardiovascular causes. Patients with inflammation (>10 mg/l) had a significantly increased mortality risk compared with those without inflammation [HRim = 1.86 (95%CI 1.46–2.38); HRhs = 1.99 (95%CI 1.55, 2.56)]; Figure 4)], which remained significant after adjustment for age, gender, PKD, Khan comorbidity score, BMI, smoking habit and GFR [HRim,adj = 1.49 (95%CI 1.14–1.93), HRhs,adj = 1.53 (95%CI 1.18–2.0)].


Figure 4
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. Cumulative survival of incident dialysis patients with (grey) vs without (black) systemic inflammation according to immunoturbidimetric (dashed lines) and high sensitivity analysis of CRP (solid lines). With both measurement methods, patients with systemic inflammation (CRP>10 mg/l) were at increased risk for mortality during follow-up (HRhsCRP, adj = 1.53, HRimCRP,adj = 1.49).

 
Patients in the middle tertile of imCRP (CRP 4–9 g/ml) did not have a significantly increased mortality risk [HRadj 1.3 (0.9–1.8)] compared with patients in the lowest tertile of imCRP (reference group, CRP 3 mg/ml). In contrast, patients in the highest tertile of imCRP (CRP ≥ 10 mg/ml) had an increased mortality risk [HRadj 1.8 (1.3–2.4)]. These results were comparable for the tertiles of hsCRP, even though cut-off values for each of the tertiles were slightly different from those for imCRP (Table 4).


View this table:
[in this window]
[in a new window]

  		Table 4. Mortality risks (HR with 95%CI) associated with tertiles of CRP

 
When analysed as continuous variable, the crude mortality risk for each 1 mg/l increase in serum CRP was 1.009 (95% CI 1.007–1.015) for both imCRP and hsCRP. Adjustment for confounders (see above) slightly influenced the mortality risk, but to the same extent for both measurement methods [HRim,adj = 1.006 (95% CI 1.002–1.011) and HRhs,adj = 1.006 (95% CI 1.002–1.010)]. This corresponds with an increase in (adjusted) mortality risk of 6.2% for an increase in either imCRP or hsCRP of 10 mg/l. To further explore the added value of hsCRP over imCRP, subjects who had imCRP values of 3 mg/l, the lowest detectable concentration by imCRP, were studied. Within this group of patients (n = 336), a 1 mg/l raise of serum hsCRP was not significantly associated with increased mortality risk [HRhs,adj = 1.03 (95% CI 0.85–1.25)].

The discriminative capacity of imCRP and hsCRP for mortality was very similar. The c-statistic for imCRP and hsCRP was 0.65 (95%CI 0.61–0.69) and 0.66 (0.62–0.69), respectively, which improved to 0.83 (0.79–0.86) and 0.83 (0.79–0.86), respectively, after adjustment for age, gender, PKD, Khan comorbidity score, BMI, smoking habit and GFR.

Influence of storage time
ImCRP and hsCRP were measured at different points in time in the present study. Therefore, the effect of storage time on CRP levels was assessed. Patients were grouped according to the storage time of serum prior to measurement of CRP. Within storage periods, imCRP was significantly higher than hsCRP levels (P < 0.001, Figure 5). However, duration of serum storage did not affect the level of imCRP or hsCRP (P = 0.245 and P = 0.065, respectively). CRP was assessed by both methods within a storage time of <5 years in 146 patients, and within a storage time ≥5 years in 112 patients. The ICCs within each group corrected for systematic differences between the methods were good (ICC<5yrs = 0.87 and ICC≥5yrs = 0.84).


Figure 5
View larger version (17K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5. Levels of imCRP (black bars) and hsCRP (open bars) by storage time of serum prior to measurement of CRP.

 


   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Acknowledgements
 References
 
This study showed a very good agreement between CRP as measured by low-sensitive and high-sensitive assays in patients with ESRD. Furthermore, the present data clearly indicate that ESRD patients with systemic inflammation had an increased mortality risk when compared with those without inflammation. This finding was comparable for the two measurement methods. For both methods, every 10 mg/l increase in serum CRP was associated with an increase in mortality risk of 6.2%. Finally, the predictive capacity for mortality was highly similar for CRP determined by low-sensitive or high-sensitive assay. These results suggest that in ESRD patients, assessment of mortality risk is equally effective with imCRP as with hsCRP.

This is the first study showing very good agreement between a low-sensitive and a hsCRP assay in ESRD patients when analysed both for the overall range of CRP concentrations and dichotomized into present or absent systemic inflammation. At different cut-off-points indicating presence of inflammation, the agreement ({kappa}) as well as sensitivity, specificity, positive predictive value and negative predictive value remained constant and very good. The results of our study are in accordance with findings of Clarke et al. [29] who showed an excellent correlation between a medium-sensitive and hsCRP assay in patients suspected of coronary artery disease [29].

In our group of ESRD patients, ~30% had systemic inflammation with CRP >10 mg/l, which is in accordance with previous studies—reviewed by Arici and Walls [14]. Median levels of CRP in dialysis patients are much higher than in the general population, in which levels of 1.5–2 mg/ml have been observed [11,30]. Previous follow-up studies in ESRD patients have demonstrated that elevated serum CRP is associated with increased (cardiovascular) mortality [15–18], which is in line with our data. It could be argued that our observed (adjusted) mortality risks associated with high imCRP or high hsCRP were relatively low at ~1.5. However, overall mortality risks were calculated during a follow-up period with a maximum of 5 years. When restricting our analysis to the first year of follow-up, adjusted mortality risks (95% CI) were 4.1 (2.2, 7.8) for high imCRP and 3.7 (2.0, 6.9) for high hsCRP, as compared with the reference group with normal CRP levels. Furthermore, CRP determined by medium-sensitive and high-sensitive assay have similar predictive values for myocardial infarction and death in patients suspected of coronary artery disease, when comparing the highest quartile of CRP to the lowest [29]. In the present study, these findings were extended as it was demonstrated that, when using 10 mg/l as cut-off-point for determination of inflammatory status, the mortality risk for ESRD patients with systemic inflammation compared with patients without systemic inflammation was similar when assessed by low-sensitive or high-sensitive method. Our data are especially valid since they were obtained in a large, countrywide patient cohort. Moreover, laboratory data were obtained at homogeneous time points in incident dialysis patients.

The agreement of the association of low-sensitive CRP and hsCRP measurement method with mortality was analysed in four different ways. First, patients were categorized into those with and without systemic inflammation based on an arbitrary cut-off point for CRP (10 mg/l). Second, patients were categorized according to tertile of CRP, as has been done before [31]. Even though creating groups by tertiles of CRP yielded slight differences in the levels of CRP covered by the tertiles between imCRP and hsCRP, the mortality risks associated with each of the tertiles were comparable between imCRP and hsCRP. Third, CRP levels on a continuous scale were modelled in relation to mortality. Finally, the discriminative capacity of CRP for mortality was examined by means of receiver operating characteristic analysis. All analyses resulted in comparable results (HRs and c-statistic, respectively) for imCRP and hsCRP, supporting agreement of the two methods in their association with mortality. In order to further explore the added value of hsCRP in the lowest concentration range (<3 mg/l), the mortality risk for hsCRP was calculated separately for patients with imCRP <3 mg/l. Within this group of 336 individuals, hsCRP was not significantly associated with mortality. The latter data indicate that the added value of the lower detection limit of hsCRP is limited in mortality risk assessment in patients with ESRD.

The present study has potential limitations. A systematic difference was observed between the low-sensitive and hsCRP measurement method. As shown by Bland–Altman plot, imCRP measures slightly higher concentrations than hsCRP. This can be explained by the fact that imCRP has a poorer precision than hsCRP and measures in concentration steps of 1.0 mg/l rather than 0.1 mg/l, as is the case for hsCRP. Nevertheless, the ICC for comparison of the two methods was good. Correction of ICCs for such systematic differences, as suggested by Fleiss and Shrout [32], indeed resulted in even better agreement between the low- and high-sensitive measurement method for CRP. Furthermore, the detection limits of the low- and hsCRP method are different, 3 and 0.1 mg/l, respectively, which may affect the agreement between the two methods. In the present study, ~40% of the patients had serum CRP values ≤3 mg/l. When the agreement between the methods was assessed within a detection range which is assessable for both methods (>3 mg/l), the agreement further improved. Two patients with extreme CRP values were excluded from all analyses since their CRP values (or the combination of imCRP and hsCRP) seemed biologically impossible. A post hoc analysis including these patients yielded similar reliability results, further substantiating the agreement between the two measurement methods. Finally, imCRP levels were measured earlier in time than hsCRP levels (2003 vs 2005). However, storage time of serum did not influence the absolute level of imCRP or hsCRP. This is in accordance with earlier findings, showing that in healthy individuals CRP levels in plasma or serum do not change over time in material stored as long as 20 years at –70°C [33]. Moreover, agreement between the methods was similar in groups of patients in whom both CRP measurements were performed in samples stored, 5 vs > 5 years. Overall, it seems unlikely that these methodological issues had major influences on the presented results.

What are the clinical implications? In ESRD patients, agreement between a low-sensitive and high-sensitive measurement method for CRP is very good, in particular after correction for systematic differences between the measurement methods, outliers and detection range. ImCRP and hsCRP seem equally predictive in the assessment of mortality risk in these ESRD patients who recently started dialysis treatment. HsCRP has been advocated as the preferred method for assessment of inflammatory status [5], given the much higher precision of this method (0.1 mg/l) and low detection limit (0.1 mg/l). However, in ESRD patients with relatively high CRP levels, the added value of this precision may be questioned. This is supported by our finding that in patients with imCRP≤3 mg/l, hsCRP was not associated with mortality. The overall mortality risk is very high in ESRD patients [12]. Therefore, it seems likely that precise CRP values at the lower end of the concentration range, as obtained by high-sensitive measurement methods, do not add to the mortality risk, whereas elevated levels of CRP do. Moreover, low-sensitive CRP assays are widely available in clinical laboratories which facilitates (and thereby favours) the use of this tool not only for clinical but also for research application in patients with ESRD.

In conclusion, although imCRP has been considered as a less sensitive method to detect CRP levels than hsCRP, it is equally effective for determination of presence of inflammation in ESRD patients. The increase in precision by CRP determination by high-sensitivity method does not improve mortality risk stratification, neither in patients with imCRP≤3 mg/l, nor with imCRP>3 mg/l. This implies that there is no need to use a high-sensitivity method instead of imCRP for determination of inflammatory status in ESRD patients.



   Acknowledgements
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Acknowledgements
References
 
This study was supported by grants from the Dutch Kidney Foundation (E.018) and the Dutch National Health Insurance Board (OG97/005). The nursing staffs of the participating dialysis centres are gratefully acknowledged for collecting most of the clinical data. The authors also wish to thank the staff of the NECOSAD trial office for their assistance in the logistics of this study.

Conflict of interest statement. None declared.



   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Acknowledgements
 References
 

  1. Lacson E Jr, Levin NW. C-reactive protein and end-stage renal disease. Semin Dial (2004) 17:438–448.[CrossRef][Web of Science][Medline]
  2. Maggiore U, Cristol JP, Canaud B, et al. Comparison of 3 automated assays for C-reactive protein in end-stage renal disease: clinical and epidemiological implications. J Lab Clin Med (2005) 145:305–308.[CrossRef][Web of Science][Medline]
  3. Ledue TB, Rifai N. Preanalytic and analytic sources of variations in C-reactive protein measurement: implications for cardiovascular disease risk assessment. Clin Chem (2003) 49:1258–1271.[Abstract/Free Full Text]
  4. Wilson AM, Ryan MC, Boyle AJ. The novel role of C-reactive protein in cardiovascular disease: risk marker or pathogen. Int J Cardiol (2006) 106:291–297.[CrossRef][Web of Science][Medline]
  5. Pearson TA, Mensah GA, Alexander RW, et al. Markers of inflammation and cardiovascular disease: application to clinical and public health practice: A statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation (2003) 107:499–511.[Free Full Text]
  6. Lindahl B, Toss H, Siegbahn A, Venge P, Wallentin L. Markers of myocardial damage and inflammation in relation to long-term mortality in unstable coronary artery disease. FRISC Study Group. Fragmin during instability in coronary artery disease. N Engl J Med (2000) 343:1139–1147.[Abstract/Free Full Text]
  7. Liuzzo G, Biasucci LM, Gallimore JR, et al. The prognostic value of C-reactive protein and serum amyloid a protein in severe unstable angina. N Engl J Med (1994) 331:417–424.[Abstract/Free Full Text]
  8. Morrow DA, Rifai N, Antman EM, et al. C-reactive protein is a potent predictor of mortality independently of and in combination with troponin T in acute coronary syndromes: a TIMI 11A substudy. Thrombolysis in myocardial infarction. J Am Coll Cardiol (1998) 31:1460–1465.[Abstract/Free Full Text]
  9. Rao M, Jaber BL, Balakrishnan VS. Inflammatory biomarkers and cardiovascular risk: association or cause and effect? Semin Dial (2006) 19:129–135.[CrossRef][Web of Science][Medline]
  10. Cook NR, Buring JE, Ridker PM. The Effect of Including C-Reactive Protein in Cardiovascular Risk Prediction Models for Women. Ann Intern Med (2006) 145:21–29.[Abstract/Free Full Text]
  11. Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med (2002) 347:1557–1565.[Abstract/Free Full Text]
  12. U.S. Renal Data System, USRDS 2005. Annual Data Report: Atlas of End-Stage Renal Disease in the United States, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases. (2005) Bethesda, MD.
  13. Zoccali C, Mallamaci F, Tripepi G. Traditional and emerging cardiovascular risk factors in end-stage renal disease. Kidney Int Suppl (2003) 85:S105–S110.[Medline]
  14. Arici M, Walls J. End-stage renal disease, atherosclerosis, and cardiovascular mortality: is C-reactive protein the missing link? Kidney Int (2001) 59:407–414.[CrossRef][Web of Science][Medline]
  15. den Elzen WP, van Manen JG, Boeschoten EW, Krediet RT, Dekker FW. The effect of single and repeatedly high concentrations of C-reactive protein on cardiovascular and non-cardiovascular mortality in patients starting with dialysis. Nephrol Dial Transplant (2006) 21:1588–1595.[Abstract/Free Full Text]
  16. Ducloux D, Bresson-Vautrin C, Kribs M, Abdelfatah A, Chalopin JM. C-reactive protein and cardiovascular disease in peritoneal dialysis patients. Kidney Int (2002) 62:1417–1422.[CrossRef][Web of Science][Medline]
  17. Mallamaci F, Tripepi G, Cutrupi S, Malatino LS, Zoccali C. Prognostic value of combined use of biomarkers of inflammation, endothelial dysfunction, and myocardiopathy in patients with ESRD. Kidney Int (2005) 67:2330–2337.[CrossRef][Web of Science][Medline]
  18. Tripepi G, Mallamaci F, Zoccali C. Inflammation markers, adhesion molecules, and all-cause and cardiovascular mortality in patients with ESRD: searching for the best risk marker by multivariate modeling. J Am Soc Nephrol (2005) 16 [Suppl 1]:S83–S88.[CrossRef]
  19. Stenvinkel P, Lindholm B. C-reactive protein in end-stage renal disease: are there reasons to measure it? Blood Purif (2005) 23:72–78.[Web of Science][Medline]
  20. Khan IH, Catto GR, Edward N, Fleming LW, Henderson IS, MacLeod AM. Influence of coexisting disease on survival on renal-replacement therapy. Lancet (1993) 341:415–418.[CrossRef][Web of Science][Medline]
  21. Hermans MM, Brandenburg V, Ketteler M, et al. Study on the relationship of serum fetuin-A concentration with aortic stiffness in patients on dialysis. Nephrol Dial Transplant (2006) 21:1293–1299.[Abstract/Free Full Text]
  22. Tsirpanlis G. The pattern of inflammation and a potential new clinical meaning and usefulness of C-reactive protein in end-stage renal failure patients. Kidney Blood Press Res (2005) 28:55–61.[CrossRef][Web of Science][Medline]
  23. Woloshin S, Schwartz LM. Distribution of C-reactive protein values in the United States. N Engl J Med (2005) 352:1611–1613.[Free Full Text]
  24. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet (1986) 1:307–310.[CrossRef][Web of Science][Medline]
  25. McIntyre C, Harper I, Macdougall IC, Raine AE, Williams A, Baker LR. Serum C-reactive protein as a marker for infection and inflammation in regular dialysis patients. Clin Nephrol (1997) 48:371–374.[Web of Science][Medline]
  26. Park CW, Shin YS, Kim CM, et al. Increased C-reactive protein following hemodialysis predicts cardiac hypertrophy in chronic hemodialysis patients. Am J Kidney Dis (2002) 40:1230–1239.[CrossRef][Web of Science][Medline]
  27. Spittle MA, Hoenich NA, Handelman GJ, Adhikarla R, Homel P, Levin NW. Oxidative stress and inflammation in hemodialysis patients. Am J Kidney Dis (2001) 38:1408–1413.[Web of Science][Medline]
  28. Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med (1996) 15:361–387.[CrossRef][Web of Science][Medline]
  29. Clarke JL, Anderson JL, Carlquist JF, et al. Comparison of differing C-reactive protein assay methods and their impact on cardiovascular risk assessment. Am J Cardiol (2005) 95:155–158.[CrossRef][Web of Science][Medline]
  30. Kistorp C, Raymond I, Pedersen F, Gustafsson F, Faber J, Hildebrandt P. N-terminal pro-brain natriuretic peptide, C-reactive protein, and urinary albumin levels as predictors of mortality and cardiovascular events in older adults. JAMA (2005) 293:1609–1616.[Abstract/Free Full Text]
  31. Yeun JY, Levine RA, Mantadilok V, Kaysen GA. C-Reactive protein predicts all-cause and cardiovascular mortality in hemodialysis patients. Am J Kidney Dis (2000) 35:469–476.[Web of Science][Medline]
  32. Fleiss JL, Shrout PE. The effects of measurement errors on some multivariate procedures. Am J Public Health (1977) 67:1188–1191.[Abstract/Free Full Text]
  33. Hutchinson WL, Koenig W, Frohlich M, Sund M, Lowe GD, Pepys MB. Immunoradiometric assay of circulating C-reactive protein: age-related values in the adult general population. Clin Chem (2000) 46:934–938.[Abstract/Free Full Text]
Received for publication: 20.10.06
Accepted in revised form: 22. 5.07


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Eur Heart JHome page
R. Koos, V. Brandenburg, A. H. Mahnken, G. Muhlenbruch, S. Stanzel, R. W. Gunther, J. Floege, W. Jahnen-Dechent, M. Kelm, and H. P. Kuhl
Association of fetuin-A levels with the progression of aortic valve calcification in non-dialyzed patients
Eur. Heart J., August 2, 2009; 30(16): 2054 - 2061.
[Abstract] [Full Text] [PDF]

Home page
 J. Am. Soc. Nephrol.Home page
F. L.H. Muntinghe, M. Verduijn, M. W. Zuurman, D. C. Grootendorst, J. J. Carrero, A. R. Qureshi, K. Luttropp, L. Nordfors, B. Lindholm, V. Brandenburg, et al.
CCR5 Deletion Protects Against Inflammation-Associated Mortality in Dialysis Patients
J. Am. Soc. Nephrol., July 1, 2009; 20(7): 1641 - 1649.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
22/11/3277    most recent
gfm381v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Grootendorst, D. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Grootendorst, D. C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?