| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Clinical and Experimental Medicine, Unit of Internal Medicine, and
2 Institute of Clinical Chemistry, University of Verona, 37134 Verona, Italy;
aaddress correspondence to this author at: Department of Clinical and Experimental Medicine, Unit of Internal Medicine, Policlinico G.B. Rossi, P.le A. Scuro 10, 37134 Verona, Italy; fax 39-45-580111, e-mail oliviero.olivieri{at}univr.it
Renovascular disease (RVD) is emerging as a common and frequently unsuspected cause of chronic renal failure and end-stage renal disease (1); moreover, it is frequently associated with renovascular hypertension, which generally carries a worse cardiovascular prognosis than other hypertensive conditions (2)(3). RVD and its associated hypertension are often not readily distinguishable from essential hypertension with renal damage, and the currently accepted diagnostic workup strategy is largely based on early clinical suspicion of the disease (4).
Among the main elements contributing to an increased clinical suspicion of RVD is a decreased glomerular filtration rate (GFR), often detected by an increase in serum creatinine concentration. The creatinine concentration is indeed convenient and inexpensive to determine, but it is not ideal because of its sensitivity to changes in muscle mass, dietary protein intake, tubular secretion, and extrarenal metabolism. In addition, it is a relatively poor index of the function of individual kidneys and renal mass because unilateral renal artery stenosis >70% is often associated with decreased GFR in one kidney but no change in the creatinine concentration until 50% of the total renal mass is lost ("creatinine blinded area") (1)(5).
Recent evidence suggests that serum cystatin C is a reliable marker of GFR (5)(6)(7)(8). Moreover, cystatin C may more readily detect subtle decrements of GFR than does serum creatinine (5)(9)(10). We therefore investigated whether cystatin C can serve as a more sensitive indicator than creatinine in a rule-out strategy for the diagnostic workup for RVD.
We studied 128 free-living individuals (64 males/64 females) who were normotensive and free of any drugs. No renal or other significant diseases were recorded in the history. The participant selection strategy was also aimed at having five age classes (<30, 31–40, 41–50, 50–60, and >61 years; age range, 22–78 years) for each gender composed of 11–13 participants each.
We then studied 150 consecutive hypertensive atherosclerotic patients without overt renal failure who had an intermediate to high index of clinical suspicion of RVD, as defined according to the Mann–Pickering criteria (11). All patients underwent clinical examination, routine biochemical analysis (including serum cystatin C and creatinine), and renal angiographic evaluation [for details, see Ref. (12)]. According to the angiographic results, patients were classified as RVD or RVD-free (at least 70% lumen reduction). We also recruited 12 RVD patients undergoing percutaneous renal angioplasty and stenting and measured their serum cystatin C and creatinine within 48 h (short-term) and 90 days (long-term) after the procedure to evaluate whether angioplasty-induced renal blood flow changes were associated with parallel variations in the GFR markers.
Creatinine was assayed by the Jaffe method on a DAX 96 multichannel analyzer, according to the manufacturer’s procedures (Bayer Diagnostic); the intraassay and interassay CVs were <3% and <5%, respectively. Cystatin C was measured by commercially available particle-enhanced nephelometric immunoassay with antiserum, calibrator, and a BNII nephelometer from Dade Behring (13). The intraassay CV was calculated on 10 control replicates and interassay CV on duplicates over 10 days. Imprecision was within the manufacturer’s specifications, i.e., the intraassay CV was <2.1% and the interassay CV was <2.5%. Details concerning the statistical analysis are reported as data supplements accompanying the online version of this Technical Brief athttp://www.clinchem.org/content/vol48/issue12/.
The mean (SD) cystatin C and creatinine concentrations in normotensive controls were 0.78 (0.09) mg/L (range, 0.57–1 mg/L) and 74.5 (13.5) µmol/L (range, 44.2–114.9 µmol/L), respectively. The percentile distributions of serum cystatin C and creatinine in healthy normotensive population are reported in Table A in the online data supplements.
A total of 65 patients had angiographically significant RVD, whereas the remaining 85 patients with no lesions and no significant stenosis (<70%) were considered the RVD-free group. In the RVD group, 12 patients also had a contralateral stenosis <50%, and 10 patients had bilaterally significant lesions. The two groups were well matched for most features, but RVD patients were older and had higher serum cystatin C and creatinine (Table B in the online data supplements).
The ROC curves for cystatin C and creatinine were nearly superimposable [areas under the curves, 0.73 (95% confidence interval, 0.64–0.80) and 0.74 (95% confidence interval, 0.66–0.82) for cystatin C and creatinine, respectively]. However, in agreement with the fact that the value corresponding to the 90th (or 95th) percentile of the distribution of the normotensive population was equivalent to the 25th (or 30th) percentile for cystatin C and to the 45th (or 55th) percentile for creatinine of the distribution for the hypertensive population, marked differences in the diagnostic performance were observed when either the 90th or the 95th percentile of the normotensive population was used as the decision threshold. The performance of both analytes according to different decision thresholds is reported in Table 1
.
|
The likelihood ratio of a positive test (LR+) and the likelihood ratio of a negative test (LR-) were also calculated. As expected, they were virtually identical at the ROC-selected cutoffs, but for cystatin C, LR+ was 1.55 and 1.57 and LR- was 0.19 and 0.28 at the 90th and 95th percentiles, respectively; for creatinine, LR+ was 1.90 and 2.70 and LR- was 0.45 and 0.39 at the 90th and 95th percentiles, respectively. The value of LR- obtained for cystatin C (0.19 for the 90th percentile) was consistent with an odds ratio of 4.26 (i.e., probability against the occurrence of true RVD diagnosis if cystatin C was <0.90 mg/L), suggesting a possible clinical usefulness of the test.
The changes in cystatin C and creatinine concentrations after renal angioplasty were evaluated over both the short and long term (Fig. 1
). Angioplasty decreased cystatin C (Fig. 1A
), but did not appear to affect creatinine values over either the short or long term (Fig. 1B
).
|
We also compared the sensitivities of creatinine and cystatin C in reflecting differences in arterial luminal narrowing and the coexistence of significant bilateral stenosis. Artery luminal stenosis of 65–70% was generally associated with a cystatin C concentration >0.90 mg/L, the decision cutoff corresponding to the 90th percentile of the normotensive individuals (Fig. 2 in the online data supplements). In the case of creatinine, however, 80% (or more) lumen reduction seemed necessary to cause a steady increase in concentrations to >92.9 µmol/L, the 90th percentile of the normotensive individuals (Fig. 3 in the online data supplements).
Patients with bilateral lesions had statistically higher cystatin C (1.42 ± 0.43 vs 1.17 ± 0.33 mg/L; P <0.05) but not significantly higher creatinine (114 ± 28 vs 111 ± 28.6 µmol/L) than patients with unilateral stenosis (Fig. 4 in the online data supplements).
The results of our study suggest that cystatin C may be a possible "discovery test" for RVD and a more sensitive marker of ischemic GFR impairment than creatinine. The difference in diagnostic sensitivity between cystatin C and creatinine was not directly observed by comparing operationally equivalent points on the ROC curve, but hypertensive atherosclerotic patients with cystatin C values within the reference interval had a much higher probability of being free of disease than did patients with creatinine values within the reference interval. The probability against the occurrence of true RVD diagnosis in an atherosclerotic hypertensive patient with cystatin C <0.90 mg/L was 4.26:1, which appears a reasonable odds ratio for screening purpose. Conversely, the same probability in a patient with creatinine values in the reference interval was 1.22:1.
Compensatory hyperfunction of the unaffected kidney generally minimizes the consequences of unilateral ischemic damage, and this is indeed what currently happens in the so-called creatinine blinded area, in which creatinine concentrations exceed the upper limit of the reference interval only when GFR is reduced to <75 mL/min (5). In our study, such an increase in cystatin C reflected an earlier stage of renal damage because it increased progressively above the 90th percentile of the normotensive range as the lumen of renal artery was reduced by 65–70%, whereas a corresponding increase in creatinine was recognizable when at least 80% lumen stenosis was present. These data agree with the notion that serum cystatin C increases when GFR is 
88 mL/min (5). Additional evidence was also provided by our analysis of RVD patients treated with angioplasty and stenting of the stenosed vessels. Improved renal blood flow induced by successful angioplasty was followed by changes in cystatin C but not creatinine (Fig. 1
), suggesting that relatively subtle GFR variations were adequately reflected by cystatin C but not by creatinine concentrations. This finding highlights the possibility that improvement of renal function may be underestimated or not recognized by creatinine measurements in some ischemic patients treated with angioplasty. It is possible that this reduced sensitivity is partially attributable to the aspecific chromogenic interferences inherent in the Jaffe method and that our conclusions cannot be applied to the enzymatic method, which does not imply such interferences. Although the Jaffe method is less expensive and more widely used, a comparison between enzymatically measured creatinine and cystatin could be an interesting issue for future research in the field. Taking into account that the value of the present results must be restricted to the specific clinical setting of atherosclerotic hypertensive patients, cystatin C assay may be of clinical utility in a rule-out diagnostic strategy for RVD.
Acknowledgments
This study was supported by grants from the Ministry of the University and Scientific and Technological Research, the Veneto Region Department of Health, and the Fondazione Cariverona-Progetto Sanità 1996–97.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
