Serum Cystatin C as a Reliable Marker of Changes in Glomerular Filtration Rate in Children With Urinary Tract Malformations

Serum Cystatin C as a Reliable Marker of Changes in Glomerular Filtration Rate in Children With Urinary Tract Malformations A. M. Corrao, G. Lisi, G. Di Pasqua, M. Guizzardi, N. Marino, E. Ballone and P. Lelli Chiesa* From the Pediatric Surgery Unit (AMC, NM, PLC), University “G. d’Annunzio” of Chieti-Pescara (GL, PLC), Department of Laboratory Medicine, “Spirito Santo” Hospital of Pescara (GDP) and Department of Biomedical Science, Laboratory of Biostatistics, University “G. d’Annunzio” of Chieti and Ce.S.I., Clinical Research Centre of “G. d’Annunzio” University Foundation (MG, EB), Pescara, Italy

Purpose: Cystatin C has been suggested as a simple method of estimating GFR more accurately than creatinine in children. We compared the diagnostic accuracy of cystatin C with serum creatinine and the Schwartz formula for estimating GFR in patients with UTMs. Materials and Methods: We prospectively compared 72 patients with UTMs (20 days to 36 months old, 58 males and 14 females) with a group of 72 healthy controls (10 days to 48 months old, 53 males and 19 females). All patients underwent nuclear medicine clearance investigations with 99mTc DTPA. Results: Serum concentration of cystatin C revealed a higher correlation with 99mTc DTPA (r ⫽ 0.62, p ⬍0.001) than serum concentration of creatinine (r ⫽ 0.30, p ⬍0.01) or Schwartz formula (r ⫽ 0.51, p ⬍0.001). These results were more evident in patients with uropathy (19) with mild renal impairment. Agreement between methods was assessed using Bland Altman analysis. Mean differences between GFR calculated with 99mTc DTPA and cystatin C based GFR estimation or Schwartz formula were ⫺2.6% ⫾ 46.7% and ⫺73.4% ⫾ 53.6%, respectively. Diagnostic accuracy in identifying decreased GFR measured as AUC was always highest for cystatin C but hardly sufficient for the 3 variables. Cystatin C performed better in the 0 to 6-month-olds (0.70 ⫾ 0.08 for cystatin C, 0.58 ⫾ 0.07 for Schwartz estimate) and patients older than 12 months (0.82 ⫾ 0.09 for cystatin C, 0.65 ⫾ 0.11 for Schwartz estimate). Conclusions: Cystatin C proved to be a superior marker rate over serum creatinine in estimating glomerular filtration in children younger than 3 years with UTMs and mild renal impairment, thus, offering a more specific and practical measure for monitoring GFR. Key Words: post gamma-globulins, glomerular filtration rate, urogenital abnormalities, pediatrics

TMs are responsible for the majority of cases of kidney impairment in children. The results of the ItalKid Project (national population based registry of chronic renal failure with 1,197 registered patients) demonstrated that hypodysplasia associated with UTM is the leading cause of chronic renal failure and end stage renal disease in children.1 Measurement of GFR is the most commonly used test of renal function in this setting. The gold standard method of measuring GFR uses an exogenous clearance marker, such as inulin or radiolabeled compounds.2,3 Because of the complexity of this investigation, the most popular tests are endogenous markers such as serum creatinine.4 Serum creatinine in neonates and infants requires adjustment for height, sex, muscle mass, diet, etc,5 while the timed collection of urine samples features a whole other series of difficulties, making the determination of creatinine clearance inaccurate.6 The use of low molecular weight plasma proteins (2microglobulin and ␣1-microglobulin) as markers of decreased GFR has been suggested.6 However, these markers do not reflect GFR accurately. Subsequently, cystatin C has

been proposed as an alternative low molecular weight protein marker of GFR.2,6 The aims of this study were to compare serum cystatin C to serum creatinine levels and creatinine clearance as markers of GFR, and to analyze their modifications in a large group of children with UTMs. The gold standard for the determination of renal function was 99m Tc DTPA clearance value.

U

PATIENTS AND METHODS Patients. We studied 72 children (58 males and 14 females) with a mean age of 6.7 ⫾ 7.9 months (range 20 days to 36 months, mean height 63.86 ⫾ 12.27 cm, range 37 to 98.5) with UTMs (table 1). No preterm neonates were included in the study group. Most of the patients (61) were younger than 12 months. Patients were recruited consecutively between September 2001 and January 2004. Patients were attendees of the Pediatric Surgery Department at the Civic Hospital in Pescara, Italy. Parental informed consent was obtained for each patient.

Submitted for publication July 27, 2004. Study received institutional review board approval. * Correspondence: Cattedra di Chirurgia Pediatrica c/o Unità Operativa di Chirurgia Pediatrica, Ospedale Civile “Spirito Santo,” Via Fonte Romana, 8, 65100 Pescara, Italy (e-mail: p.lellichiesa@ chirped.unich.it).

0022-5347/06/1751-0303/0 THE JOURNAL OF UROLOGY® Copyright © 2006 by AMERICAN UROLOGICAL ASSOCIATION

Controls. Reference values were obtained from 72 healthy volunteers (53 males and 19 females, mean age 9.4 ⫾ 12 months, range 10 days to 48 months, mean height 66.27 ⫾ 19.25 cm, range 42 to 142) who were age, sex and height

303

Vol. 175, 303-309, January 2006 Printed in U.S.A. DOI:10.1016/S0022-5347(05)00015-7

304

CYSTATIN C AS MARKER OF CHANGES IN GLOMERULAR FILTRATION RATE TABLE 1. Diagnoses in 72 patients with urinary tract malformation Diagnoses

No. Pts

Primary megaureter: Obstructive Obstructive-refluxing Ureteropelvic junction obstruction: Grade I–II (ultrasonography) Grade III Grade IV Vesicoureteral reflux: Grade II–III (managed by continuous urodynamic monitoring) Grade IV–V Posterior urethral valves Miscellaneous: Multicystic dysplastic kidney Congenital giant bladder diverticulum Renal agenesis

22 4 2 18 4 3 11 3 3 1 1

matched to the patients with uropathy. Patients and controls were tested for thyroid function at birth, which was within normal range of hormonal production in all individuals. Neither group was tested further for hypothyroidism or hyperthyroidism when enrolled in the study. Data on the use of glucocorticoids in both groups at the time of the study were not available. Laboratory measurements. Glomerular filtration rate was determined using a 99mTc DTPA single injection technique (99mTc DTPA GFR).7 After rapid intravenous injection of an appropriate amount of tracer a direct external scintigraphic determination of fractional radionuclide accumulation within each kidney, occurring at a specified time interval after radionuclide administration (6 minutes), was performed using a gamma camera without blood samples. Clearance values of 40 to 60 ml per minute⫺1 ⫻ 1.73 m⫺2 body surface area in patients up to age 6 months, 60 to 80 ml per minute⫺1⫻ 1.73 m⫺2 in those up to 1 year and 80 to 120 ml per minute⫺1 ⫻ 1.73 m⫺2 in those older than 1 year were defined as normal GFR on the basis of scintigraphic studies performed in a pediatric population to standardize the gamma camera uptake method (unpublished data). Serum creatinine was measured with an enzymatic assay. Creatinine clearance was estimated according to the Schwartz formula [height (cm) ⫻ constant/serum creatinine (mg/dl)].8 The constant was 0.55 for children older than 12 months8 and 0.45 for term neonates and infants.9 Determination of cystatin C was performed using the N Latex Cystatin C™ test kit on a BN™II analyzer. Average interassay and intra-assay coefficients of variation for creatinine and cystatin C measurements were less than 5%.

Statistical analysis. Data were expressed as mean ⫾ SD or median, 25th and 75th percentile, and range and charted in a box plot. Student’s t test for unpaired data, MannWhitney U test or Kruskal-Wallis H test was used to evaluate differences in means or medians, where appropriate. Pearson’s correlation coefficient, Spearman’s rank correlation coefficient and standardized partial regression coefficients were used to evaluate correlations and concordances between parameters. Agreement between the gold standard GFR (99mTc DTPA) and either GFRcys or Schwartz estimate was assessed using the Bland Altman method. This statistical analysis is performed plotting the differences between 2 measurement techniques against the average. Results are reported as mean difference ⫾ 1.96 SD. If the results do not have differences in ranges, then the 2 methods can be considered interchangeable. Sensitivity (95% CI), specificity (95% CI) and ROC analysis were used to evaluate, at various cutoffs, the diagnostic accuracy of cystatin C, creatinine and Schwartz formula compared to GFR in 144 children. Cutoff values were determined using 2 different criteria, either the one used by the software that maximized the sum of SE and SP, or the value with 90% SE. To illustrate the diagnostic performance of each GFR marker, ROC curves were reported and the AUC was calculated. All statistical analysis was performed using SPSS Advanced Statistical™ 10.0 software and MedCalc® 7.3.0.1 software. RESULTS Means ⫾ SD for serum creatinine, Schwartz formula, cystatin C levels and 99mTc DTPA assessment of GFR values in controls and patients with UTMs are listed in table 2. In the group of children with UTMs only cystatin C levels were significantly different compared to controls (p ⫽ 0.005), while the values at different ages were statistically significant at younger than 6 months and older than 1 year (p ⬍0.01 and p ⬍0.05, respectively). The difference in cystatin C levels between patients with UTMs and controls, observed up to and after age 1 year, were statistically significant (p ⬍0.05, H test, fig. 1). The relationship between serum concentration of cystatin C and 99mTc DTPA GFR for all 72 patients with UTMs and for the subset of 61 patients 0 to 12 months old exhibited greater correlation (r ⫽ 0.62, p ⬍0.001 and r ⫽ 0.66, p ⬍0.001, respectively). A plot of serum concentration of cystatin C vs 99mTc DTPA GFR illustrated a nonlinear correlation, and a log/log transformation was required to obtain a straight line. From this transformation the following func-

TABLE 2. Glomerular filtration rate markers in 72 patients with urinary tract malformation and 72 controls Variables

UTM Pts

Matched Controls

p Value*

Serum creatinine (mg/dl) Creatinine clearance (Schwartz formula) Cystatin C (mg/l): 6 Mos or younger Older than 6 but younger than 12 mos 12 Mos or older 99m Tc DTPA GFR (ml/min/1.73 m2)

0.29 ⫾ 0.12 129.18 ⫾ 45.13 1.28 ⫾ 0.53 1.50 ⫾ 0.57 0.91 ⫾ 0.16 1.04 ⫾ 0.28 53.22 ⫾ 15.54

0.31 ⫾ 0.10 124.48 ⫾ 42.18 1.09 ⫾ 0.26 1.22 ⫾ 0.25† 0.93 ⫾ 0.09 0.84 ⫾ 0.11‡ —

Not significant Not significant 0.005

* Student’s t test for unpaired data. † Mann-Whitney U test u p ⬍ 0.01. ‡ Mann-Whitney U test u p ⬍ 0.05.

—

CYSTATIN C AS MARKER OF CHANGES IN GLOMERULAR FILTRATION RATE

305

serum concentration of cystatin C and Schwartz formula vs Tc DTPA GFR was particularly significant (p ⬍0.001). The ␳ for creatinine was less significant (p ⬍0.05). We then performed multiple regression analysis using 99m Tc DTPA GFR as a dependent variable and GFR biochemical markers as independent variables (table 3). The results indicated that serum concentration of cystatin C was strongly associated with GFR (p ⬍0.001), while the Schwartz formula was less significantly associated (p ⬍0.05). The result for serum concentration of creatinine was not statistically significant. The results of Bland Altman analysis displayed a mean difference between the GFRcys and the gold standard method of ⫺2.6% ⫾ 46.7%, whereas the mean difference between the Schwartz estimate and 99mTc DTPA GFR was ⫺73.4% ⫾ 53.6% (fig. 2). For all children the AUCs for cystatin C, creatinine and Schwartz formula were 0.67 ⫾ 0.05, 0.55 ⫾ 0.05 and 0.54 ⫾ 0.05, respectively (data not shown). With a large range of normal values during the first 12 months for each of the 3 variables and 99mTc DTPA clearance values a single cutoff for these parameters for infants and children was considered not applicable. Data were then analyzed dividing patients into 3 age groups, ie 6 months or younger (group I), older than 6 but younger than 12 months (group II) and 12 months or older (group III, table 4 and fig. 3). The clinical SE and SP were calculated at a selected decision point to obtain either a cutoff with the optimum sum of SE ⫹ SP, or a cutoff with a sensitivity of at least 90%. In group I the AUCs for cystatin C, creatinine and Schwartz formula were 0.70 ⫾ 0.08, 0.61 ⫾ 0.08 and 0.58 ⫾ 0.07, respectively (table 4). With regard to cystatin C, cutoff values of 1.31 mg/l (SE ⫽ 76%, SP ⫽ 62%) and 1.21 mg/dl (SE ⫽ 88.2%, SP ⫽ 42.3%) were calculated. In group II the AUCs for cystatin C, creatinine and Schwartz formula were 0.63 ⫾ 0.06, 0.61 ⫾ 0.06 and 0.54 ⫾ 0.05, respectively. The cystatin C best diagnostic performance was calculated at a cutoff value of 1.31 mg/l (SE ⫽ 41%, SP ⫽ 75%), while a sensitivity of at least 81% was obtained at a cutoff value of 0.92 mg/l. In group III the AUCs for cystatin C, creatinine and Schwartz formula were 0.82 ⫾ 0.09, 0.53 ⫾ 0.12 and 0.65 ⫾ 0.11, respectively (fig. 3, C). The 2 cutoff values for cystatin C were 1.03 mg/dl (SE ⫽ 60%, SP ⫽ 100%) and 0.92 mg/l (SE ⫽ 80%, SP ⫽ 77%). In the group of 19 patients with UTMs and slightly decreased renal function the corresponding AUCs were 0.65 ⫾ 0.08, 0.54 ⫾ 0.07 and 0.53 ⫾ 0.08 (data not shown). 99m

FIG. 1. Box plot distribution demonstrates serum cystatin C concentrations at different ages (up to and after 12 months) in 72 children with urinary tract malformation (UTM) and 72 controls. Box plot extends from 25th to 75th percentile. Horizontal line represents median, and whiskers represent range of data (p ⬍0.05, KruskalWallis H test).

tion for estimating GFRcys was obtained: log(GFR) ⫽ 4.03 ⫹ [0.53 ⫻ log(serum concentration of cysC)]. Slightly lower correlation coefficients were obtained for serum concentration of creatinine adjusted according to the Schwartz formula vs 99mTc DTPA GFR for all 72 patients with UTMs and for the 61 patients 0 to 12 months old (r ⫽ 0.51, p ⬍0.001 and r ⫽ 0.56, p ⬍0.001, respectively).The correlation between serum concentration of creatinine and 99mTc DTPA GFR was even lower (r ⫽ 0.30, p ⬍0.01). Only the differences between the latter correlation coefficient and the other coefficients were statistically significant (p ⬍0.05, data not shown). In the group of patients with UTMs (19 patients 20 days to 18 months old, 17 younger than 12 months) with 99mTc DTPA GFR values 10% to 20% lower than the age predicted value 99mTc DTPA GFR and serum concentration of cystatin C exhibited the greatest correlation (r ⫽ 0.70, p ⫽ 0.001). Lower correlation was obtained between 99mTc DTPA GFR and the Schwartz estimate (r ⫽ 0.62, p ⬍0.01), whereas GFR and serum concentration of creatinine were not significantly related (r ⫽ 0.28, p ⫽ 0.24, data not shown). Comparison of the concordance indexes (␳) revealed statistically significant agreement between each of the 3 variables and 99mTc DTPA GFR, confirming Pearson’s correlation coefficients (table 3). The concordance index for

DISCUSSION Urologists must monitor renal function, particularly in children with abnormalities of the urinary tract and progressive

TABLE 3. Spearman’s rank correlation coefficient and standardized partial regression coefficients of investigated variables on GFR in 72 patients with UTM Variables

␳*

p Value

␤

t

p Value

Serum cystatin C Serum creatinine Schwartz formula

0.60 0.26 0.49

⬍0.001 ⬍0.05 ⬍0.001

0.43 ⫺0.20 0.43

3.43 ⫺1.05 1.97

⬍0.001 Not significant ⬍0.05

Patient age ranged from 0 to 36 months. * Used such as concordance index.

306

CYSTATIN C AS MARKER OF CHANGES IN GLOMERULAR FILTRATION RATE

FIG. 2. Bland Altman plot for agreement between glomerular filtration rate (GFR) and GFR predicted from either serum cystatin C (GFRcys, A) or Schwartz formula (B). Cystatin C based estimation of GFR was calculated according to newly derived formula: log(GFR) ⫽ 4.03 ⫹ [0.53 ⫻ log(serum concentration of cysC)]. Regression analysis of values obtained from Bland Altman analysis revealed, for both estimates, slope that was significantly nonzero (p ⫽ 0.01 for cystatin C, p ⫽ 0.001 for Schwartz). Horizontal lines represent mean ⫾ 1.96 SD.

renal disease. In clinical practice measurement of serum creatinine is the most commonly accepted endogenous filtration marker for GFR estimation in children and adults. However, its limitations are well established.4 – 6 Measurement of mildly impaired GFR, the so-called “creatinine blind area,” has been a challenge up to now. This is especially true in pediatric populations because of the low muscle mass in children, which leads to low creatinine values, where increased assay imprecision is present. Therefore, it can be difficult to detect accurately small changes in GFR with serum creatinine in children younger than 4 years, whose normal creatinine values are only 2.0 to 4.0 mg/l.6 Low molecular weight plasma proteins such as ␤2-microglobulin and ␣1-microglobulin, eliminated mainly by glomerular filtration, could be markers of renal impairment but do not reflect GFR accurately.6 Thus, established methods for measurement of GFR include the clearance of exogenous substances such as 99mTc DTPA and 51 chrome ethylenediaminetetraacetic acid. These methods are commonly used as “gold standards” but they are time consuming, labor intensive and expensive, and require the administration of substances that make them incompatible with routine monitoring. Cystatin C, a low molecular weight protein, has been proposed as a marker of GFR.4,6 This protein is produced by all nucleated cells, and, therefore, exhibits a stable production rate even in the presence of acute inflammatory response.10 It is freely filtered by the glomerulus, and almost completely reabsorbed and catabolized in the tubules.11 Unlike creatinine, cystatin levels are not influenced by dietary or constitutional factors. More recently, an impact of thyroid function on cystatin C levels was suggested.12 This molecule can be measured rapidly and precisely by single blood measurement, and it does not require any correction for age or height as does creatinine, so it seems well suited as a marker of GFR.13,14 Automated homogeneous immunoassays have been developed using latex particles coated with cystatin C specific antibodies to determine the concentration of cystatin C in serum. These assays are more precise than the previous methods, and reference intervals seem more consistent than

TABLE 4. Diagnostic accuracy, sensitivity, specificity and areas under ROC curves of cystatin-C, creatinine and Schwartz formula in 144 children in 3 age groups Variables Younger than 6 mos: Cystatin-C (mg/l) Creatinine (mg/dl) Schwartz (ml/min) Older than 6 but younger than 12 mos: Cystatin-C (mg/l) Creatinine (mg/dl) Schwartz (ml/min) 12 Mos or older: Cystatin-C (mg/l) Creatinine (mg/dl) Schwartz (ml/min)

SE (95% CI)

SP (95% CI)

AUC* ⫾ SE

1.31* 1.21† 0.33* 0.22† 138.81*,†

76.5 (50.1–93.0) 88.2 (63.5–98.2) 41.2 (18.5–67.3) 82.4 (56.6–96.0) 94.1 (71.2–99.0)

62.0 (49.7–73.2) 42.3 (30.6–54.6) 80.3 (69.1–88.8) 21.1 (12.3–32.4) 28.2 (18.1–40.1)

0.70 ⫾ 0.08

1.31* 0.92† 0.28* 0.23† 138.81*

41.7 (25.5–59.2) 80.6 (64.0–91.8) 61.1 (43.5–76.8) 83.3 (67.2–96.6) 72.2 (54.8–85.8)

75.0 (66.7–82.8) 23.1 (15.6–32.2) 58.3 (48.5–67.7) 27.8 (19.6–37.2) 41.7 (32.3–51.4)

1.03* 0.92† 0.40* 0.28† 158.65*,†

60.0 (26.4–87.6) 80.0 (44.4–96.9) 40.0 (12.4–73.6) 80.0 (44.4–96.9) 90.0 (55.5–98.3)

100.0 (75.1–100.0) 76.9 (46.2–94.7) 73.3 (44.9–92.0) 26.7 (8.0–55.1) 46.7 (21.3–73.4)

Cutoffs

* Cutoff chosen by software to maximize sum of SE and SP. † Value corresponding with sensitivity of at least 90%.

0.61 ⫾ 0.08 0.58 ⫾ 0.07 0.63 ⫾ 0.06 0.61 ⫾ 0.06 0.54 ⫾ 0.05 0.82 ⫾ 0.09 0.53 ⫾ 0.12 0.65 ⫾ 0.11

CYSTATIN C AS MARKER OF CHANGES IN GLOMERULAR FILTRATION RATE

FIG. 3. ROC plots for diagnostic accuracy of serum cystatin C (CYS_C), creatinine (CREA_S) and Schwartz formula (S) in 144 children in age groups 6 months or younger (A), older than 6 but younger than 12 months (B) and 12 months or older (C).

those reported using previous assays.6 Reference ranges for plasma cystatin C measurements, using the same latex particle enhanced immunoassay we used, have been similar to those reported in our study in neonates and children.15,16

307

Finney et al, who measured cystatin C levels in 291 children 1 day to 17 years old, reported the highest concentrations in premature and full-term neonates, with a gradual decline until about age 1 year (mean levels in children younger than 1 year were 1.20 mg/l, range 0.59 to 1.97), when adult values were reached (mean concentration in children 1 to 17 years old was 0.80 mg/l, range 0.50 to 1.27).15 Randers et al, who evaluated 137 children 7 days to 14.1 year old, without kidney diseases, confirmed the decreasing trend of cystatin C measurements during the first year of life (mean values ⫾ SD were 1.63 ⫾ 0.26 mg/l within the first month of life, and 0.95 ⫾ 0.22 mg/l in patients 1 to 12 months old) and observed a subsequent constant level (0.72 ⫾ 0.12 mg/l).16 Cystatin C has been proposed as a biochemical marker endowed with a greater degree of diagnostic sensitivity than creatinine in situations in which there is only a moderate decrease in GFR. Thus, cystatin C might prove highly advantageous in situations of “blind creatinine period” and initial renal impairment.17 In recent years cystatin C has proved to be a marker that reflects GFR accurately in adults as well as children and newborns.6,13,14,18 A recent meta-analysis of 54 studies comparing the accuracy of cystatin C and creatinine in relation to a reference standard of GFR demonstrated that cystatin C is clearly superior to serum creatinine as a marker of GFR.19 In children with various renal diseases serum cystatin C exhibits superior diagnostic accuracy for decreased GFR compared to serum creatinine, although its performance will unlikely supplant 51 chrome ethylenediaminetetraacetic acid.20,21 To date, the available data in children indicate that cystatin C is at least as useful as serum creatinine in assessing GFR, with several studies indicating that cystatin C may perform better.22,23 In 27 patients with spina bifida (mean age approximately 12 years) only cystatin C served as a suitable marker to estimate GFR.22 Thus, cystatin C proved to be a superior marker in this setting. However, there have been no studies specifically designed to evaluate the reliability of cystatin C in pediatric patients with uropathy, and it is impossible to obtain these data from the available literature. Still, few children younger than 4 years (particularly with regard to the problem of the “blind creatinine period”), for whom it is hypothesized that cystatin C may be most effective, have been included in published clinical studies evaluating diagnostic performance.6 We evaluated the reliability of cystatin C as a marker of GFR and its diagnostic performance in 72 children younger than 3 years old with uropathy. Our results indicate that cystatin C correlates with 99mTc DTPA GFR better than the Schwartz formula or serum creatinine, with correlation coefficients at the lower end of reported data in children.21,24 Nevertheless, correlation analysis is insufficient to evaluate the agreement between diagnostic investigations. We then performed a Bland Altman analysis, obtaining an appreciable result when using the cystatin C based estimation of GFR (average bias ⫺2.6% ⫾ 46.7%). This result is in line with published data regarding the pediatric population. Filler and Lepage23 reported a result of 0.3% ⫾ 43.5%, while Bokenkamp et al25 found an average bias of 2.37% ⫾ 44%. Both studies referred to a population of children and adolescents. For the Schwartz estimate we found an unacceptable larger bias (average ⫺73.4% ⫾ 53.6%). The same studies

308

CYSTATIN C AS MARKER OF CHANGES IN GLOMERULAR FILTRATION RATE

previously cited23,25 demonstrated different results, 10.8% ⫾ 47.55% and ⫺4.6% ⫾ 44%, respectively. Our result is not surprisingly poorer, considering the already known large overestimation of GFR when using the Schwartz formula for infants and children with immature or severely decreased renal function.26 Analyzing these data, we should then conclude that cystatin C allows a largely better estimation of GFR than the Schwartz formula. However, we need to indicate that both estimations resulted in an error of nearly 50%, which is unacceptable when an accurate determination of GFR is necessary for clinical and therapeutic decisions. We believe that in this situation a full renal scan cannot be replaced by a GFR estimate. The results of ROC analysis revealed barely sufficient diagnostic performance of GFR markers, particularly in patients younger than 12 months. These results are poorer than those for published pediatric reports but are not entirely comparable, considering that our study population was mainly made up of infants. Data in this restricted age group are unavailable at the moment. In our series cystatin C performed better than the Schwartz formula, particularly in patients younger than 6 months and those older than 12 months. Both groups presented with cystatin C levels that differed significantly between patients with UTMs and controls. To our knowledge the data published in the literature refer to a pediatric population with a wide age range and a large predominance of children older than 1 year. Our study population included patients up to 3 years, most of whom were younger than 12 months. Whether our results are due to the high percentage of infants requires further investigation. We can speculate that other variables, such as the possible inaccuracy of the scintigraphic determination of GFR and the wide variability of renal function, cystatin C and creatinine levels during the first year of life, can negatively influence the results.

CONCLUSIONS Our study adds further evidence to the usefulness of cystatin C in a larger pediatric urological setting. Our results confirm the trend of normal values of cystatin C up to and after the first 12 months of life reported in literature, and show that cystatin C correlates with GFR better than the Schwartz formula in an entire group of patients with uropathy and in a subgroup of patients with mildly impaired renal function. Moreover, to our knowledge this is the first such study in a large group of patients younger than 3 years, most of whom were younger than 12 months. Bland Altman analysis demonstrated that cystatin C is the best method for estimating GFR, and ROC analysis revealed that serum cystatin C is more sensitive and specific than serum creatinine, and as sensitive as but more specific than the Schwartz clearance estimate. Thus, we have shown the superiority of cystatin C over serum creatinine in patients younger than 3 years with urinary tract malformations and mild renal impairment, offering a more specific and practical measurement for monitoring GFR.

Abbreviations and Acronyms AUC ⫽ area under the receiver operating characteristic plot GFR ⫽ glomerular filtration rate GFRcys ⫽ cystatin C based estimate ␳ ⫽ Spearman’s rank correlation coefficient r ⫽ Pearson’s correlation coefficient ROC ⫽ receiver operating characteristic SD ⫽ standard deviation SE ⫽ sensitivity SP ⫽ specificity 99m Tc DTPA ⫽ 99mtechnetium diethylenetetramine pentaacetic acid 99m Tc DTPA GFR ⫽ glomerular filtration rate determined using 99m technetium diethylenetetramine pentaacetic acid UTM ⫽ urinary tract malformation

REFERENCES 1. Ardissino, G., Dacco, V., Testa, S., Bonaudo, R., Claris-Appiani, A., Taioli, E. et al: Epidemiology of chronic renal failure in children: data from the ItalKid project. Pediatrics, 111: e382, 2003 2. Campens, D. and Buntinx, F.: Selecting the best renal function tests. A meta-analysis of diagnostic studies. Int J Technol Assess Health Care, 13: 343, 1997 3. Price, C. P. and Finney, H.: Developments in the assessment of glomerular filtration rate. Clin Chim Acta, 297: 55, 2000 4. Perrone, R. D., Masias, N. E. and Levey, A. S.: Serum creatinine as an index of renal function new insights into old concepts. Clin Chem, 38: 1933, 1992 5. Schwartz, G. J., Brion, L. P. and Spitzer, A.: The use of plasma creatinine concentration for estimating glomerular filtration rate in infants, children, and adolescents. Pediatr Clin North Am, 34: 571, 1987 6. Laterza, O. F., Price, C. P. and Scott, M. G.: Cystatin C: an improved estimator of glomerular filtration rate? Clin Chem, 48: 699, 2002 7. Gates, G. F.: Glomerular filtration rate: estimation from fractional renal accumulation of 99mTc-DTPA (stannous). AJR Am J Roentgenol, 138: 565, 1982 8. Schwartz, G. J., Haycock, G. B., Edelmann, C. M., Jr. and Spitzer, A.: A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics, 58: 259, 1976 9. Schwartz, G. J., Feld, L. G. and Langford, D. J.: A simple estimate of glomerular filtration rate in full-term infants during the first year of life. J Pediatr, 104: 849, 1984 10. Abrahamson, M., Olafsson, I., Palsdottir, A., Ulvsback, M., Lundwall, A., Jensson, O. et al: Structure and expression of the human cystatin C gene. Biochem J, 268: 287, 1990 11. Tenstad, O., Roald, A. B., Grubb, A. and Aukland, K.: Renal handling of radiolabelled human cystatin C in the rat. Scand J Clin Lab Invest, 56: 409, 1996 12. Fricker, M., Wiesli, P., Brandle, M., Schwegler, B. and Schmid, C.: Impact of thyroid dysfunction on serum cystatin C. Kidney Int, 63: 1944, 2003 13. Herget-Rosenthal, S., Trabold, S., Pietruck, F., Holtmann, M., Philipp, T. and Kribben, A.: Cystatin C: efficacy as screening test for reduced glomerular filtration rate. Am J Nephrol, 20: 97, 2000

CYSTATIN C AS MARKER OF CHANGES IN GLOMERULAR FILTRATION RATE 14. Galteau, M. M., Guyon, M., Gueguen, R. and Siest, G.: Determination of serum cystatin C biological variation reference values. Clin Chem Lab Med, 39: 850, 2001 15. Finney, H., Newman, D. J., Thakkar, H., Fell, J. M. and Price, C. P.: Reference ranges for plasma cystatin C and creatinine measurements in premature infants, neonates, and older children. Arch Dis Child, 82: 71, 2000 16. Randers, E., Krue, S., Erlandsen, E. J., Danielsen, H. and Hansen, L. G.: Reference interval for serum cystatin C in children. Clin Chem, 45: 1856, 1999 17. Coll, E., Botey, A., Alvarez, L., Poch, E., Quinto, L., Saurina, A. et al: Serum cystatin C as a new marker for noninvasive estimation of glomerular filtration rate and as a marker for early renal impairment. Am J Kidney Dis, 36: 29, 2000 18. Cataldi, L., Mussap, M., Bertelli, L., Ruzzante, N., Fanos, V. and Plebani, M.: Cystatin C in healthy women at term pregnancy and their infant newborns: relationship between maternal and neonatal serum levels and reference values. Am J Perinatol, 16: 287, 1999 19. Dharnidharka, V. R., Kwon, C. and Stevens, G.: Serum cystatin C is superior to serum creatinine as a marker of kidney function a meta-analysis. Am J Kidney Dis, 40: 221, 2002

309

20. Kilpatrick, E. S., Keevil, B. G. and Addison, G. M.: Does adjustment of GFR to extracellular fluid volume improve the clinical utility of cystatin C? Arch Dis Child, 82: 499, 2000 21. Ylinen, E. A., Ala-Houhala, M., Harmoinen, A. P. and Knip, M.: Cystatin C as a marker for glomerular filtration rate in pediatric patients. Pediatr Nephrol, 13: 506, 1999 22. Pham-Huy, A., Leonard, M., Lepage, N., Halton, J. and Filler, G.: Measuring glomerular filtration rate with cystatin and ␤-trace protein in children with spina bifida. J Urol, 169: 2312, 2003 23. Filler, G. and Lepage, N.: Should the Schwartz formula for estimation of GFR be replaced by cystatin C formula? Pediatr Nephrol, 18: 981, 2003 24. Willems, H. L., Hilbrands, L. B., van de Calseyde, J. F., Monnens, L. A. H. and Swinkels, D. W.: Is serum cystatin C the marker of choice to predict glomerular filtration rate in pediatric patients? Ann Clin Biochem, 40: 60, 2003 25. Bokenkamp, A., Domanetzki, M., Zinck, R., Schumann, G., Byrd, D. and Brodehl, J.: Cystatin C—a new marker of glomerular filtration rate in children independent of age and height. Pediatrics, 101: 875, 1998 26. Seikaly, M. G., Browne, R., Bajaj, G. and Arant, B. S., Jr.: Limitations to body length/serum creatinine ratio as an estimation of glomerular filtration in children. Pediatr Nephrol, 10: 709, 1996