The inaccuracy of cystatin C and creatinine-based equations in predicting GFR in orthotopic liver transplant recipients

Transcription

1 NDT Advance Access published June 8, 2009 Nephrol Dial Transplant (2009) 1 5 doi: /ndt/gfp255 Original Article The inaccuracy of cystatin C and creatinine-based equations in predicting GFR in orthotopic liver transplant recipients Neil Boudville 1,2, Muna Salama 3, Gary P. Jeffrey 1,3 and Paolo Ferrari 1,4 1 School of Medicine and Pharmacology, University of Western Australia, 2 Department of Renal Medicine, Sir Charles Gairdner Hospital, 3 Department of Gastroenterology, Sir Charles Gairdner Hospital and 4 Department of Renal Medicine, Fremantle Hospital, Perth, Australia Correspondence and offprint requests to: Neil Boudville; neil.boudville@uwa.edu.au Abstract Background. As survival with an orthotopic liver transplant (OLT) improves, the incidence of chronic kidney disease in OLT recipients increases. Measurement of kidney function using creatinine-based estimates is often inaccurate, while cystatin C may overcome the biases that effect creatinine. The aim of this study was to assess the accuracy of creatinine- and cystatin C-based equations to estimate kidney function in long-term OLT recipients. Methods. This was a cross-sectional study performed on OLT recipients within a single liver transplant centre where creatinine (n = 41) and cystatin C (n = 30) were measured and glomerular filtration rate (GFR) estimated using the Modification of Diet and Renal Disease (MDRD), Cockcroft Gault (CG), Hoek, Larsson, Filler and Le Bricon equations. Comparison was made with the nuclear GFR (ngfr) (n = 41) measured through 51-Cr EDTA clearance. Results. The mean age of recipients was 56 ± 13 years, and they were 6.5 ± 4.7 years post-olt. Fifty-six percent of recipients had a ngfr 60 ml/min/1.73 m 2. ngfr correlated significantly with all predictive equations (P < 0.001). The MDRD, CG and Le Bricon equations had the smallestdegreeofbias( 7.6, 7.3 and 3.4 ml/min/1.73 m 2, respectively), with 22%, 22% and 27% of estimates, respectively, being within 10% of ngfr measurements. In OLT recipients with ngfr 60 ml/min/1.73 m 2, the degree of bias of both the creatinine-base MDRD and cystatinbased Hoek equations was within 2 ml/min/1.73 m 2 difference between the measured and estimated GFR, but 41% and 36% of estimates were within 10% of the ngfr measurement. Conclusions. Therefore, the degree of inaccuracy in cystatin C- and creatinine-based predictive equations brings into question their clinical utility in OLT recipients. We have no evidence that cystatin C is superior to creatinine in this population. Keywords: chronic kidney failure; glomerular filtration rate; kidney function tests; liver transplantation; predictive equations Introduction Chronic kidney disease (CKD) following orthotopic liver transplantation (OLT) is becoming an increasing problem as patient and graft survival following OLT improves. The 1-, 5- and 8-year graft survival following OLT is currently estimated at 85%, 70% and 62%, respectively [1,2]. Recent data indicate that the incidence of CKD following OLT is 18% after 5 years and 26% after 10 years [3,4]. End-stage kidney disease (ESKD) requiring dialysis or kidney transplantation occurs in 9.5% of patients 13 years following OLT, and these patients had a 28% 13-year survival compared to 55% in the control group [3,4]. The aetiology of CKD is multifactorial in nature with nephrotoxicity secondary to calcineurin inhibitors probably playing a major role [3 5]. CKD leads to significant complications, with an increased incidence of morbidity and mortality [3,4]. ESKD results in considerable cost and an annual mortality of up to 20% [6]. Kidney function is most commonly estimated using serum creatinine-based equations. In the context of chronic liver disease, however, creatinine has been shown to overestimate kidney function due to poor muscle bulk, impaired creatinine biosynthesis and hyperbilirubinaemia [7 9]. Hence, equations that utilize serum creatinine to estimate glomerular filtration rate (GFR) in chronic liver disease patients are prone to error [7 9]. The gold standard method of estimation of kidney function is with inulin clearance or nuclear medicine tests [10]. These, however, are not practical for usual clinical practice due to their complexity and cost. Creatinine clearance using 24-h urine collection is also prone to measurement error due to difficulties with collection [11]. Cystatin C, a low-molecular-weight ( 13.3 kd) protease inhibitor, is produced at a constant rate in all nucleated cells [12,13]. It is freely filtered in the glomerulus where it is then fully reabsorbed in the proximal tubule and catabolized. Its level is unaffected by muscle bulk, age and gender. For these reasons, recent evidence suggests that it may provide a more accurate estimate of kidney function, especially C The Author [2009]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 2 N. Boudville et al. Table 1. Predictive equations for calculating glomerular filtration rate (GFR) based upon creatinine and cystatin C MDRD GFR = 175 (creatinine ) age (1 for white male, for females, 1.2 for African Americans) CG Creatinine clearance = (1 age) weight/creatinine (1.23 if male or 1.04 if female) Hoek GFR = ( /cystatin C) Filler Log (GFR) = [1.123 log(1/cystatin C)] Larsson GFR = (cystatin C ) Le Bricon GFR = [78 (1/cystatin C)] + 4 MDRD, Modification of Diet in Renal Disease; CG, Cockcroft Gault. through the use of predictive equations [13 16]. Our aim was to investigate further the accuracy of creatinine- and cystatin C-based predictive equations for estimating GFR in long-term OLT patients. Subjects and methods This was a cross-sectional study performed on 41 OLT patients between July 2005 and January A total of 133 patients attend regular followup visits at the Sir Charles Gairdner Hospital Liver Transplant Unit with the first 41 eligible participants included into this study. This liver transplant unit services 2 million people in the state of Western Australia, Australia. The study was approved by the local Human Research Ethics Committee. Written informed consent was obtained from each patient. Patients that were recipients of multi-organ transplants were not recruited. All patients were over the age of 18 years and had an OLT for at least 3 months. Patient information was collected including demographic data, indication for transplantation, transplant date, immunosuppression regimen and hepatitis B and C status. Blood tests were performed on the same day as nuclear GFR (ngfr) was measured. Creatinine was measured using a kinetic colourimetric assay (Jaffe method) analysed on the Roche Hitachi 917 analyser (Roche Diagnostic GmbH, Mannheim, Germany). Total coefficients of variation (cv) for this assay are 6.6% at 70.5 μmol/l and 4.1% at 485 μmol/l. The assay used was calibrated with an IDMS reference laboratory. Cystatin C was measured on the same day as the ngfr was performed, using particleenhanced immunonephelometry run on the BNII analyser (Dade Boehring GmbH, Marbug, Germany). The total cv for this assay are 4.0% at 0.44 mg/l and 5.3% at 0.76 mg/l. ngfr was estimated through the injection of 3.0 megabequerels of radioactive 51 Cr ethylenediaminetetraacetic acid (EDTA). Blood samples were obtained 1, 2 and 3 h later for a total of three blood samples. 51 Cr EDTA clearance was then calculated and corrected for body surface area. GFR was estimated from serum creatinine utilizing the Modification of Diet in Renal Disease (MDRD) [17] study and Cockcroft Gault equations (Table 1) [18,19]. In addition, GFR was estimated from cystatin C using the equations validated by Hoek [19], Larsson [20], Filler [21] and Le Bricon [22]. Because MDRD-GFR has been shown to have a poor correlation with ngfr at values >60 ml/min/1.73 m 2, a separate analysis was done for correlations between creatinine- and cystatin-based equations and ngfr for values below and above 60 ml/min/1.73 m 2 by EDTA clearance [23]. Statistical analysis Results are expressed as mean ± standard deviation, unless otherwise stated. Linear regression was utilized to compare the inverse of creatinine, inverse of cystatin C and GFR predictive equations with ngfr. Spearman s rank correlation was used instead of linear regression if the assumptions of normality of residuals and homoscedasticity were not met. To estimate the reliability of the GFR equations, the difference in ngfr indicated the bias, while the standard deviation of the difference indicated precision [24]. Accuracy in this article is used to quantify the combination of bias and precision. Participants were divided into two groups based upon ngfr <60 ml/min/1.73 m 2 and 60 ml/min/1.73 m 2. STATA 8.2 (Stata Corporation, College Station, TX, USA) was used for the analysis. Table 2. Baseline data in 41 orthotopic liver transplant recipient by nuclear GFR (ngfr) category (mean ±SD or percent) ngfr <60 ml/ ngfr 60 ml/ min/1.73 m 2 min/1.73 m 2 N Age (years) 56.3 ± ± ± 12.0 Time since 6.5 ± ± ± 4.0 transplantation (years) Weight (kg) 79.9 ± ± ± 17.7 Nuclear GFR 58.0 ± ± ± 13.5 (ml/min/1.73 m 2 ) Serum creatinine ± ± ± 24.6 (μmol/l) Serum albumin 43.4 ± ± ± 3.5 (g/l) Cystatin C (mg/l) a 1.6 (1.3; 1.8) 1.8 (1.6; 2.0) 1.4 (1.2; 1.7) Male (%) Race Caucasian (%) Asian (%) Hepatitis B (%) 7.1% Hepatitis C (%) 21.4% a Cystatin C is not normally distributed and so results are expressed as median (25th percentile; 75th percentile). Results Forty-one OLT recipients participated in this study with measurements of ngfr and serum creatinine. Thirty subjects also had serum cystatin C measured. The mean time since transplantation was 6.5 ± 4.7 years (Table 2). The cause of liver failure included autoimmune liver disease in 44%, viral hepatitis in 29%, alcohol in 7% and other causes in 20% of subjects. The standard immunosuppression regimens at the time of transplantation were prednisolone, azathioprine and calcineurin inhibitors (cyclosporine or tacrolimus). Current immunosuppression included 68% on dual therapy with 61% on cyclosporine and 27% on tacrolimus. The mean spot urine protein:creatinine was 14.3 ± 9.5 mg/mmol (range mg/mmol). On ngfr, the mean was 58.0 ± 19.8 ml/min/1.73 m 2 (range ml/min/1.73 m 2 ) and 56% of participants had a ngfr 60/mL/min/1.73 m 2. All creatinine-based and cystatin-based equations correlated significantly with ngfr (Table 3). The inverse of creatinine (R 2 = 0.44) and cystatin (R 2 = 0.64) also correlated significantly with ngfr, P In the entire population, the MDRD equation correlated best with ngfr on linear regression analysis (R 2 = 0.62, P < ). The cystatin-based equations all correlated closely with ngfr, with the Filler

3 Inaccuracy of cystatin C equations in OLT recipients 3 Table 3. Correlation between estimates of renal function and nuclear GFR Overall GFR <60 ml/min/1.73 m 2 GFR 60 ml/min/1.73 m 2 R 2 P-value R 2 P-value R 2 P-value 1/creatinine 0.44 < Cystatin 0.47 < /cystatin a MDRD 0.62 < < Cockcroft Gault 0.47 < Hoek a Larsson a Filler Le Bricon a 1/cystatin; Hoek and Larsson estimated GFR were analyzed using Spearman s rank correlation. Table 4. Bias and precision of predictive equations compared to nuclear GFR Within Within Bias Precision 10% 30% Overall Hoek % 73% Filler a % 23% Larsson % 60% Le Bricon % 80% Cockcroft Gault % 63% MDRD % 80% GFR<60 ml/min/1.73 m 2 Hoek % 93% Filler a % 21% Larsson % 71% Le Bricon % 79% Cockcroft Gault % 55% MDRD % 86% GFR 60 ml/min/1.73 m 2 Hoek % 56% Filler a % 25% Larsson % 50% Le Bricon % 81% Cockcroft Gault % 74% MDRD % 68% Bias = mean difference between measured and estimated GFR. Precision = SD of the difference between measured and estimated GFR. a Filler equation estimated GFR was not normally distributed and so Bias is the median difference between measured and estimated GFR and precision can not be estimated. equation performing the most poorly (R 2 = 0.32, P = 0.001). The MDRD, Cockcroft Gault and Le Bricon equations featured the smallest degree of bias, 7.6, 7.3 and 3.4 ml/min/1.73 m 2, respectively (Table 4). In addition, the MDRD and Le Bricon equations resulted in 80% of estimated GFRs in OLT recipients being within 30% of the ngfr. The cystatin C-based Filler equation performed the most poorly, underestimating the ngfr by a median of 23.5 ml/min/1.73 m 2. This formula also only managed to estimate the GFR within 30% of the ngfr in 23% of patients. Almost all predictive equations performed better in OLT recipients with a ngfr <60 ml/min/1.73 m 2 (Table 4). The equations with the smallest degree of bias in those patients with a GFR <60 ml/min/1.73 m 2 were the MDRD ( 1.7 ml/min/1.73 m 2 ) and Hoek equations (0.98 ml/min/ 1.73 m 2 ) (Figure 1). Discussion For a multitude of reasons, including improved long-term survival, OLT recipients are developing problems with the increase of CKD [1 4]. CKD alone leads to increased morbidity and mortality that is magnified in its severest manifestation of ESKD [6,25]. There are interventions available that may allow a degree of protection against progressive kidney failure, including the avoidance of nephrotoxic calcineurin inhibitors [5]. Thus, early detection and ongoing monitoring of those patients with a progressive loss of GFR would allow OLT recipients to benefit from interventions. One of the major issues in the management of CKD in OLT recipients is the overestimation of creatinine-based predictive equations for GFR [7 9]. Cystatin C is not influenced by many of the factors that usually affect creatinine concentrations in OLT recipients. As such cystatin C-based predictive equations for GFR have been suggested to be more accurate in this patient population [26 29]. The results of our study demonstrate that, in our cohort, both creatinine- and cystatin C-based estimations of kidney function correlated significantly with ngfr. The creatinine-based MDRD equation and the cystatin C-based Le Bricon equation displayed the least bias and greatest precision of all the formulas examined, underestimating the true GFR by 4 8 ml/min/1.73 m 2 on average. The accuracy even with these formulae was, however, poor, with only 22% and 27%, respectively, estimating the GFR within 10% of the ngfr measurements. The accuracy of these equations was improved when it was restricted to OLT recipients with a ngfr <60 ml/min/1.73 m 2.Eveninthis subgroup, the clinical utility of these equations, in view of the degree of inaccuracy, is questionable. A limitation of this study is the small sample size, which reduces the ability to detect differences between equations. The only other published study to date comparing creatinine and cystatin C predictive equations with ngfr in long-term OLT recipients examined 59 adults with a mean of 5.1 ± 3.0 years post-transplant [30]. They used the Hoek,

4 4 N. Boudville et al. A C B D Fig. 1. Bland Altman plots of nuclear GFR with MDRD and Hoek equations. (A) Nuclear GFR and MDRD equation (all participants). (B) Nuclear GFR and MDRD equation (only for nuclear GFR<60 ml/min/1.73 m 2 ). (C) Nuclear GFR and Hoek equation (all participants). (D) Nuclear GFR and Hoek equation (only for nuclear GFR<60 ml/min/1.73 m 2 ). Filler and Larson cystatin C equations and the MDRD equation to estimate GFR and compared it to ngfr, measured by 99 mtc-diethylene triamine pentaacetic acid clearance. The correlation was similar between formulae and similar to our results. The bias was lowest with the Hoek and Larsson formulae, underestimating the ngfr by 0.1 and 2.3 ml/min/1.73 m 2 (P = not significant). The Hoek equation had the greatest accuracy with 76.3% of estimations being within 30% of the ngfr, though this was not significantly different to the MDRD equation (64.4%). These findings are consistent with ours, except that the Le Bricon equation was not examined. Schuk et al. examined 58 adult OLT recipients with a mean of 14 ± 10 months post-transplant and measured GFR with inulin clearance [29]. They concluded that cystatin C levels <1.2 mg/l excluded mild reductions in GFR, with cystatin C having significantly greater sensitivity than creatinine. The only other study to use ngfr and compare it with cystatin C and creatinine concentration examined 62 paediatric OLT recipients with a median of 3.1 years post-transplant had similar findings [28]. Neither study examined cystatin C predictive equations. Other studies that have compared cystatin C with creatinine in long-term OLT recipients had similar inconsistent findings with two demonstrating superiority of cystatin C and the other one showing equivalence [26,27,31]. These studies did not use ngfr or creatinine clearance as a comparator. A study in kidney transplant recipients demonstrated a similar accuracy to our study of creatinine-based formula, with 53 80% of recipients estimating within 30% of the ngfr [24]. However, they found that the Filler and Le Bricon equations were superior with a bias of 1.7 and 3.8 ml/min/1.73 m 2, precision of 11.4 and 11.8 ml/min/1.73 m 2 and with 87 and 89% of estimates within 30% of the ngfr measurements [24]. On the other hand, similar to our findings, Poge et al. did not demonstrate any difference between the MDRD equation and cystatin C compared to ngfr in kidney transplant recipients [32]. Accurate estimation of kidney function is of increasing importance in OLT recipients as survival improves. We have shown that creatinine- and cystatin C-based equations vary in their bias and precision in estimating GFR. Even though these equations are improved in OLT recipients with a GFR <60 ml/min/1.73 m 2, the degree of inaccuracy places doubt over the utility of these equations in clinical practice. There is no evidence in our data that cystatin C is superior to MDRD, and in view of the added expense its use cannot be recommended at this stage. Acknowledgement. This study was wholly funded by the Department of Renal Medicine at Sir Charles Gairdner Hospital. Conflict of interest statement. None declared.

5 Inaccuracy of cystatin C equations in OLT recipients 5 References 1. Futagawa Y, Terasaki PI, Waki K et al. No improvement in long-term liver transplant graft survival in the last decade: an analysis of the UNOS data. Am J Transplant 2006; 6: Roberts MS, Angus DC, BryceCL et al. Survival after liver transplantation in the United States: a disease-specific analysis of the UNOS database. Liver Transpl 2004; 10: Gonwa TA, Mai ML, Melton LB et al. End-stage renal disease (ESRD) after orthotopic liver transplantation (OLTX) using calcineurin-based immunotherapy: risk of development and treatment. Transplantation 2001; 72: Ojo AO, Held PJ, Port FK et al. Chronic renal failure after transplantation of a nonrenal organ. NEnglJMed2003; 349: Creput C, Blandin F, Deroure B et al. Long-term effects of calcineurin inhibitor conversion to mycophenolate mofetil on renal function after liver transplantation. Liver Transpl 2007; 13: Goodkin DA, Bragg-Gresham JL, Koenig KG et al. Association of comorbid conditions and mortality in hemodialysis patients in Europe, Japan, and the United States: the Dialysis Outcomes and Practice Patterns Study (DOPPS). J Am Soc Nephrol 2003; 14: Caregaro L, Menon F, Angeli P et al. Limitations of serum creatinine level and creatinine clearance as filtration markers in cirrhosis. Arch Intern Med 1994; 154: Gerard SK, Khayam-Bashi H. Characterization of creatinine error in ketotic patients. A prospective comparison of alkaline picrate methods with an enzymatic method. Am J Clin Pathol 1985; 84: Sherman DS, Fish DN, Teitelbaum I. Assessing renal function in cirrhotic patients: problems and pitfalls. Am J Kidney Dis 2003; 41: Brochner-Mortensen J, Giese J, Rossing N. Renal inulin clearance versus total plasma clearance of 51Cr-EDTA. Scand J Clin Lab Invest 1969; 23: Toto RD. Conventional measurement of renal function utilizing serum creatinine, creatinine clearance, inulin and para-aminohippuric acid clearance. Curr Opin Nephrol Hypertens 1995; 4: ; discussion Barrett AJ, Davies ME, Grubb A. The place of human gamma-trace (cystatin C) amongst the cysteine proteinase inhibitors. Biochem Biophys Res Commun 1984; 120: Grubb A, Lofberg H. Human gamma-trace, a basic microprotein: amino acid sequence and presence in the adenohypophysis. Proc Natl Acad Sci USA 1982; 79: Jung K, Jung M. Cystatin C: a promising marker of glomerular filtration rate to replace creatinine. Nephron 1995; 70: Newman DJ, Thakkar H, Edwards RG et al. Serum cystatin C: a replacement for creatinine as a biochemical marker of GFR. Kidney Int Suppl 1994; 47: S20 S Nilsson-Ehle P, Grubb A. New markers for the determination of GFR: iohexol clearance and cystatin C serum concentration. Kidney Int Suppl 1994; 47: S17 S Levey AS, Bosch JP, Lewis JB et al. Modification of Diet in Renal Disease Study Group. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med 1999; 130: Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: Hoek FJ, Kemperman FA, Krediet RT. A comparison between cystatin C, plasma creatinine and the Cockcroft and Gault formula for the estimation of glomerular filtration rate. Nephrol Dial Transplant 2003; 18: Larsson A, Malm J, Grubb A et al. Calculation of glomerular filtration rate expressed in ml/min from plasma cystatin C values in mg/l. Scand J Clin Lab Invest 2004; 64: Filler G, Lepage N. Should the Schwartz formula for estimation of GFR be replaced by cystatin C formula? Pediatr Nephrol 2003; 18: Le Bricon T, Thervet E, Froissart M et al. Plasma cystatin C is superior to 24-h creatinine clearance and plasma creatinine for estimation of glomerular filtration rate 3 months after kidney transplantation. Clin Chem 2000; 46(8 Pt 1): Froissart M, Rossert J, Jacquot C et al. Predictive performance of the modification of diet in renal disease and Cockcroft Gault equations for estimating renal function. J Am Soc Nephrol 2005; 16: White C, Akbari A, Hussain N et al. Estimating glomerular filtration rate in kidney transplantation: a comparison between serum creatinine and cystatin C-based methods. J Am Soc Nephrol 2005; 16: Anavekar NS, McMurray JJ, Velazquez EJ et al. Relation between renal dysfunction and cardiovascular outcomes after myocardial infarction. N Engl J Med 2004; 351: Hermida J, Romero R, Tutor JC. Relationship between serum cystatin C and creatinine in kidney and liver transplant patients. Clin Chim Acta 2002; 316: Hermida J, Tutor JC. Comparison of estimated glomerular filtration rates from serum creatinine and cystatin C in patients with impaired creatinine production. Clin Lab 2006; 52: Samyn M, Cheeseman P, Bevis L et al. Cystatin C, an easy and reliable marker for assessment of renal dysfunction in children with liver disease and after liver transplantation. Liver Transpl 2005; 11: Schuck O, Gottfriedova H, Maly J et al. Glomerular filtration rate assessment in individuals after orthotopic liver transplantation based on serum cystatin C levels. Liver Transpl 2002; 8: Gerhardt T, Poge U, Stoffel-Wagner B et al. Estimation of glomerular filtration rates after orthotopic liver transplantation: evaluation of cystatin C-based equations. Liver Transpl 2006; 12: Podracka L, Feber J, Lepage N et al. Intra-individual variation of cystatin C and creatinine in pediatric solid organ transplant recipients. Pediatr Transplant 2005; 9: Poge U, Gerhardt T, Stoffel-Wagner B et al. Cystatin C-based calculation of glomerular filtration rate in kidney transplant recipients. Kidney Int 2006; 70: Received for publication: ; Accepted in revised form: