Plasma cystatin C for estimating residual GFR (rGFR) in dialysis patients

Nephrology Dialysis Transplantation, Mar 2008

Martin Tidman, Per Sjöström, Ian Jones

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

Plasma cystatin C for estimating residual GFR (rGFR) in dialysis patients

Conflict of interest statement. None declared. Mohammad R. Ardalan 1 Mohammadali M. Shoja 0 0 Department of Clinical Chemistry University Hospital of O rebro S-701 85 O rebro , Sweden 1 Department of Medicine Plasma cystatin C for estimating residual GFR (rGFR) in dialysis patients - Sir, Hoek et al. present a new formula for the estimation of residual GFR (rGFR) in dialysis patients using the plasma (P) cystatin C level [1]. The formula is constructed by regression from a group of haemodialysis (HD) and peritoneal dialysis (PD) patients with a very narrow GFR interval, 16.5 ml/min. The slope factor of the formula is very low, resulting in poor discrimination, i.e. P-cystatin C values of 4 and 7.6 mg/l (CI limits of the HD validation group) give rGFR of 4.4 and 2.0 ml/min/1.73 m2 respectively. The superior bias, accuracy and precision of the formula compared to the MDRD formula can be explained by the patient inclusion criteria and the narrow result interval of the formula. In addition, we would like to ask what cystatin C concentration do anuric dialysis patients have? Putting rGFR = 0 in the formula described gives a P-cystatin C concentration of 31.4 mg/l, which is a very high and very unlikely value considering the confidence interval of the study group. Figure 1 compares three different formulae for eGFR. The authors hypothesis for the low-slope factor of their formula is a greatly decreased production rate of cystatin C in dialysis patients, 25% of that of patients with mild renal insufficiency (from 0.08 to 0.02 mg/min) according to their Fig. 1. Calculated eGFR from S-cystatin C using our formula (100/ S-Cys 14; ), Hoeks formula for mild renal impairment (80.35/ S-Cys 4.32; ) and for HD patients (21/S-Cys 0.77; ). own previous study [2]. The production rate of cystatin C has been shown to be constant in other situations and we have found that all CKD stages fit into our formula (100/ S-cysC 14 using Gentians turbidimetric method) [3] as well as anuric dialysis patients (Tidman et al., unpublished data). The authors support for the hypothesis of reduced production rate is that the new steady-state concentration of P-cystatin C is reached much earlier than that for creatinine after acute renal failure. Our explanation is that cystatin C has a lower distribution volume (1/3) and a higher extra-renal clearance than creatinine [4]. It may take 2 weeks to reach a new steady state for creatinine, when GFR is <5 ml/min [5]. Thus, the MDRD formula cannot be used to calculate rGFR in HD patients, as it implies a steady-state concentration of creatinine and will give rise to falsely high eGFR values until this steady state is reached. Cystatin C would be more suitable in this respect, but we think that the biological variation in the production rate and extra-renal clearance is too high to make its use for measuring rGFR acceptable. An error of say 4 ml/min in GFR is insignificant at normal renal function but makes the result meaningless in a dialysis patient. We use plasma clearance of iohexol to measure rGFR in our HD patients [6]. Conflict of interest statement. None declared. Advance Access publication 8 December 2007 Reply Sir, We appreciate the interest of Tidman et al. in our study [1]. Their first point of concern is the narrow GFR interval of the patients investigated in this study. The residual GFR (rGFR) interval between 0.3 and 6.2 ml/min/1.73 m2 is the reflection of the criteria for starting haemodialysis or CAPD. Inclusion criteria for the study population were rGFR below 6.5 ml/min/1.73 m2 and a daily urine production of at least 200 ml. Then they ask what cystatin C concentration anuric patients have. However, the formula presented by us for the estimation of rGFR in dialysis patients is an empirical formula, which of course cannot be used without reservations in a population different from the one for which it was developed. Furthermore, the determination of an rGFR in anuric patients, who do not have any residual function and no urine production, is meaningless and irrelevant. In addition, the plasma cystatin C concentration observed in anuric patients does not give unequivocal information about the synthesis rate of cystatin C or its extra-renal clearance. As long as the ratio of synthesis rate and extra-renal clearance remains the same, the plasma cystatin C concentration will not change. We conclude from the constants in the formula which gives the relationship between cystatin C and rGFR that the production rate of cystatin C is reduced in dialysis patients. The fact that after acute renal failure a new steady state is reached much earlier for cystatin C than for creatinine was not used as a supportive argument by us. B okenkamp et al. [2] showed that after bilateral nephrectomy in rats, the rise in cystatin C was much smaller than for creatinine. They concluded that these and additional human data strongly suggested a change in cystatin C production or extra-renal clearance. Their conclusion is strongly supported by our data. In their letter, Tidman et al. draw attention to the new formula they have derived using the Gentian turbidimetric cystatin C assay (Tidman et al., Nephrol Dial Transplant, accepted for publication). They claim that it fits all CKD stages, with a GFR range of the study population from 12 to 125 ml/min/1.73 m2; however, the new formula has not been validated in dialysis patients. The linear formula presented in this new study is derived with the assumption that cystatin C production and extra-renal clearance are constant, just like the Sj ostr om formula published previously by the same authors [3]. The main difference between the new formula and the previous one is the standardization of the assay and a different patient cohort. When they used the DAKO assay in this new patient cohort to derive a formula, it differed only slightly from the previous Sj ostr om one. Even though it is derived in a completely different patient cohort with the use of a different gold standard, the general formula for GFR estimation published by us [4] performed at least as well as this new Tidman formula. Nevertheless, our general formula cannot be used in dialysis patients. Nor is the formula of Tidman et al. useful for dialysis patients, because it results in a negative rGFR in 2030% of the patients, and gives for a cystatin C concentration below 4.16 mg/L abnormally high rGFR values above 10 mL/min/1.73 m2. Only longitudinal follow-up of patients with our new formula for estimation of rGFR [1] can show if the estimates adequately reflect the course of rGFR decline. This is independent of the issue as to whether the constants in our formula do or do not represent the production rate and extra-renal clearance of cystatin C. 1Department of Clinical Chemistry Academic Medical Center University of Amsterdam Amsterdam, The Netherlands 2Department of Clinical Epidemiology and Biostatistics Academic Medical Center University of Amsterdam Amsterdam, The Netherlands 3Department of Clinical Epidemiology, Leiden University Medical Centre, Leiden 4Hans Mak Institute, Naarden 5Department of Nephrology Academic Medical Center University of Amsterdam Amsterdam, The Netherlands E-mail:

This is a preview of a remote PDF:

Martin Tidman, Per Sjöström, Ian Jones. Plasma cystatin C for estimating residual GFR (rGFR) in dialysis patients, Nephrology Dialysis Transplantation, 2008, 1072-1073, DOI: 10.1093/ndt/gfm771