Vascular and Interventional Radiology Original Research

Transcription

1 Vascular and Interventional Radiology Original Research Wacker-Gußmann et al. Cystatin C to Predict Contrast-Induced Nephropathy Vascular and Interventional Radiology Original Research Annette Wacker-Gußmann 1 Katharina Bühren 2 Caroline Schultheiss 2 Siegmund Lorenz Braun 3 Sharon Page 4 Bernd Saugel 2 Sebastian Schmid 5 Sebastian Mair 2 Albert Schoemig 1 Roland M. Schmid 2 Wolfgang Huber 2 Wacker-Gußmann A, Bühren K, Schultheiss C, et al. Keywords: contrast-induced nephropathy (CIN), coronary angiography, creatinine, cystatin C DOI: /AJR Received December 30, 2012; accepted after revision April 17, I. Medizinische Klinik, Klinikum rechts der Isar der Technischen Universität München, München, Germany. 2 II. Medizinische Klinik, Klinikum rechts der Isar der Technischen Universität München, Ismaningerstr. 22, D München, Germany. Address correspondence to W. Huber (wolfgang.huber@lrz.tu-muenchen.de). 3 Deutsches Herzzentrum, Institut für Laboratoriumsmedizin, München, Germany. 4 Institut für Klinische Chemie und Pathobiochemie, Klinikum rechts der Isar der Technischen Universität München, München, Germany. 5 Klinik für Anaesthesiologie, Klinikum rechts der Isar der Technischen Universität München, München, Germany. AJR 2014; 202: X/14/ American Roentgen Ray Society Prediction of Contrast-Induced Nephropathy in Patients With Serum Creatinine Levels in the Upper Normal Range by Cystatin C: A Prospective Study in 374 Patients OBJECTIVE. Preexisting renal impairment is a risk factor for contrast-induced nephropathy (CIN). In patients with creatinine in the upper normal level, cystatin C might be a more sensitive predictor of CIN than creatinine. Therefore, in this study, we investigated the usefulness of cystatin C to predict CIN. SUBJECTS AND METHODS. In 400 consecutive patients with creatinine baseline levels between 0.8 and 1.3 mg/dl undergoing coronary angiography (n = 200) or CT (n = 200), baseline values of cystatin C, creatinine, blood urea nitrogen (BUN) and risk factors of CIN were determined. Creatinine was also assessed 24 and 48 hours after contrast administration. RESULTS. Creatinine significantly () increased after 24 hours and 48 hours compared with baseline (1.06 ± 0.28 and 1.07 ± 0.28 vs 0.99 ± 0.18 mg/dl). Fifty-three of 373 evaluable patients (14.2%) had an increase in creatinine of 25% or 0.5 mg/dl within 48 hours. CIN according to this definition was significantly more frequent after intraarterial contrast administration (38/190, 20%) compared with IV contrast administration (15/183, 8.2%; p = 01). CIN was predicted by baseline cystatin C (area under the receiver operating characteristic [ROC] curve [AUC], 0.715; ), whereas creatinine, creatinine clearance, and BUN were not predictive. The best predictive capabilities were provided by cystatin C/ creatinine-ratio (AUC, 0.826; ). Multivariate regression analysis showed that intraarterial contrast administration (p = 02) and higher baseline cystatin C () combined with low creatinine (p = 44) were independently associated with higher increases in creatinine within 48 hours after contrast administration. CONCLUSION. CIN in patients with creatinine within the upper normal range is significantly more frequent after intraarterial than after IV contrast administration. In these patients, renal impairment after contrast administration is independently predicted by cystatin C and cystatin C/creatinine-ratio, whereas BUN and creatinine were not predictive. A cute kidney injury is a frequent illness that is clearly associated with hospital stay and mortality. Early recognition of acute kidney injury and immediate corrective measures may improve the outcome [1]. Within the past decade, a number of urinary and serum biomarkers have been suggested for early detection of acute renal impairment including cystatin C, neutrophil gelatinase associated lipocalin (NGAL), interleukin-18, liver-type fatty acid binding proteins (L-FABP), and kidney injury molecule 1 (KIM-1). Despite numerous hints for earlier detection of acute kidney injury by these markers compared with creatinine, their clinical use is restricted and the ideal biomarker remains to be discovered [1]. Serum creatinine is still widely used to assess glomerular filtration rate (GFR) and to detect acute kidney injury. However, creatinine may lack sensitivity regarding slightly impaired GFR due to decreased muscle mass [2], and it can be lowered by tubular secretion. Compared with creatinine, cystatin C has been reported to provide higher sensitivity with equivalent specificity in detecting abnormal GFR [3 5]. One advantage of cystatin C might be its production at a constant rate by nucleated cells irrespective of nutrition and muscle mass. Low molecular weight and positive charge at physiologic ph facilitate glomerular filtration. Cystatin C is reabsorbed and almost completely catabolized in the proximal renal tube [5 8]. This results in effective clearance from circulation 452 AJR:202, February 2014

2 Cystatin C to Predict Contrast-Induced Nephropathy in case of normal renal function and circulating plasma concentrations reflecting GFR in case of renal impairment. These properties may result in an earlier increase in cystatin C compared with creatinine in patients with slight renal impairment. Interestingly, a recent study showed that GFR-estimation equations, including both cystatin C and creatinine, provided better accuracy and precision than equations including only one of these parameters [9]. Therefore, cystatin C might be useful in early detection of patients at risk of contrastinduced nephropathy (CIN). CIN is the third leading cause of hospital-acquired acute renal failure, and it is associated with mortality and prolonged hospitalization [10 14]. According to a consensus panel, CIN is defined as an increase of serum creatinine of at least 25% or 0.5 mg/dl within 48 hours of contrast administration [13, 14]. Several risk factors for CIN have been reported, including preexisting renal dysfunction, high amounts of contrast material, diabetes, and concurrent use of nephrotoxic drugs [14]. CIN incidence is low in the absence of risk factors. However, in high-risk patients, CIN occurs in up to 50% of cases [14 20]. Assuming a low incidence of CIN in patients with normal creatinine, most of the previous trials on prevention of CIN have included patients with moderate and severe increases in creatinine [12, 13, 16 20]. Consequently, little data exist on the incidence of CIN in patients with borderline creatinine within the upper normal range. However, because of the nonlinear relationship between creatinine and GFR, creatinine values in the upper normal range can be associated with a significantly reduced GFR ( creatinine-blind area ) [21]. In these patients, cystatin C might be a more sensitive parameter to screen for patients with normal creatinine but already impaired GFR and increased risk of CIN [2 4, 6 8, 20, 21]. Therefore, the aims of our study were to investigate incidence, risk factors, and predictors of CIN, including baseline cystatin C in 400 patients with borderline serum creatinine ( mg/dl) who underwent either IV or intraarterial contrast administration. Subjects and Methods We analyzed data from a prospectively maintained database on the incidence of CIN, and 200 patients undergoing intraarterial contrast administration due to coronary angiography and 200 patients undergoing CT with IV contrast administration were included. No additional medical prophylaxis was applied except to advise a fluid intake of at least 2 L during the 24 hours before the procedure. Inclusion and Exclusion Criteria Patients were eligible for the study if the screening value of serum creatinine within 2 days before inclusion was between 0.8 and 1.3 mg/dl and the expected amount of contrast material was at least 100 ml. Patients were excluded in case of screening creatinine values outside the stated range, renal replacement therapy, and pregnancy. TABLE 1: Baseline Characteristics of Patients Undergoing Contrast-Enhanced Procedure Study Protocol The study was approved by the institutional review board of our hospital (Technical University of Munich, project number 5628/12). Baseline characteristics were documented including risk factors of CIN, such as diabetes, hypertension, and potentially nephrotoxic medications (NSAID use, including acetylic salicylic acid, aminoglycosides, vancomycin, diuretics, and amphotericin B). Within 1 hour before contrast administration, baseline values of serum levels of creatinine, creatinine clearance [22], cystatin C, and BUN were determined. Follow-up serum values of creatinine and BUN were determined 24 hours and 48 hours after contrast administration. The primary endpoint was comparison of serum creatinine levels with baseline values after 24 and 48 hours. Secondary endpoints were comparison of the incidence of CIN, defined as increase in serum creatinine of 0.5 mg/dl or 25% within 48 hours; increase in serum creatinine of 0.5 mg/dl in patients undergoing intraarterial or IV contrast administration; univariate and multivariate association increases in serum creatinine within 48 hours after contrast administration with risk factors and biochemical markers at baseline including creatinine, BUN, and cystatin C; and incidence of acute kidney injury according to the Acute Kidney Injury Network (AKIN) definition [23]. Statistical Analysis Values are expressed as the mean ± SD. Percentages were calculated on the basis of the measurements with valid data. For comparison of followup to baseline values the Wilcoxon test for paired comparisons was used. For group comparisons of continuous variables between patients undergoing intraarterial and IV contrast administration, the Mann-Whitney U test was performed. The propor- Characteristic All Patients (n = 373) IV Contrast Material (n = 183) Intraarterial Contrast Material (n = 190) p Age (y) 65.3 ± ± ± Sex (%) Women Men Weight (kg) ± ± Serum creatinine (mg/dl) 0.99 ± ± ± 0.15 < 01 Serum BUN (mg/dl) 28.3 ± ± ± 11.1 < 01 Serum cystatin C (mg/l) 1.09 ± ± ± 0.18 < 01 Creatinine clearance (MDRD) (ml/min 1.73 m 2 ) 73.8 ± ± ± 14.4 < 01 Amount of contrast material (ml) ± ± ± < 01 Cigarroa quotient (mg/dl ml kg 1 ) 3.17 ± ± ± 2.63 < 01 Cystatin C quotient (mg/l ml kg 1 ) 3.48 ± ± ± 3.03 < 01 Hypertension (%) < 01 Diabetes (%) Note Where indicated, data are ± SD. Significant values are in bold. BUN = blood urea nitrogen test, MDRD = Modification of Diet in Renal Disease study. AJR:202, February

3 tion of preexisting diabetes and hypertension and the incidence of CIN were compared between the groups undergoing IV and intraarterial contrast administration using the chi-square test. Similar analyses were performed regarding the incidence of acute renal impairment according to RIFLE (Risk, Injury, Failure, Loss, End-Stage) and AKIN criteria. RIFLE and AKIN define and quantify acute renal impairment irrespective of its cause. Both classifications grade renal impairment from slight changes in serum creatinine (R in RIFLE: increase in creatinine of 50%; stage I of AKIN: increase in creatinine 0.3 mg/dl or 50%) up to persistent requirement of renal replacement therapy. Univariate analysis regarding association of risk factors to increases in creatinine after contrast administration was performed using Spearman correlation. To characterize risk factors for higher maximum increases in serum creatinine, multiple regression analysis (backward selection) was performed. Furthermore, we performed receiver operating characteristic (ROC) analysis to evaluate the discriminative power of baseline characteristics and combined predictors regarding CIN. Statistical analyses were performed using IBM SPSS Statistics 20 software. Results Patients Characteristics Among 200 patients undergoing coronary angiography and 200 patients undergoing IV contrast administration, 190 and 183 patients were included in the final analysis, respectively. The remaining 27 of 400 patients (6.8%) were not included because of discharge less than 48 hours after contrast administration. Table 1 shows patient characteristics and baseline data for the totality of patients and the two subgroups. Despite the same inclusion and exclusion criteria, except the route of contrast administration, patients undergoing intraarterial contrast administration had significantly older age; higher weight; higher prevalence of hypertension; higher baseline values of serum creatinine, BUN, and cystatin C; lower creatinine clearance; and were administered about twice the amount of contrast material (304.4 ± ml vs ± 75.6 ml, ) compared with patients with IV contrast administration. Time Course of Retention Parameters For all patients (n = 373), creatinine levels after 24 hours (1.06 ± 0.28 mg/dl, p < 001) and 48 hours (1.07 ± 0.28 mg/dl, p < 001) were significantly higher compared with the baseline values (0.99 ± 0.18 mg/dl) (Fig. 1). Similar findings were observed in both the subgroup of patients undergoing intraarterial contrast administration and the subgroup undergoing IV contrast administration. In patients undergoing intraarterial contrast ad- Fig. 1 Graph shows Wacker-Gußmann et 1.4al. time course of serum creatinine after 1.3 administration of contrast material 1.2 (mean ± SD). Double asterisks indicate p < versus mean baseline creatinine. 1.0 Single asterisk indicates p < 5 versus mean 0.9 baseline creatinine. Serum Creatinine Before and After Contrast (mg/dl) ** ** ** ** 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h All Patients Intraarterial Contrast IV Contrast TABLE 2: Relation of Baseline Characteristics to Increase in Serum Creatinine Characteristic Measure Creatinine Change 48 Hours vs 0 Hours 24 Hours vs 0 Hours Absolute Relative Absolute Relative Intraarterial vs IV contrast administration r a a a a Serum creatinine 0 hours r p Creatinine clearance (MDRD) 0 hours r p Serum BUN 0 hours r a a a a Serum cystatin C 0 hours r a a a a Amount of contrast material r a a a a p Weight r p Cigarroa quotient r a a b 97 p Height r b b p Age r a a p Hypertension r a a a a Diabetes r p Cystatin C/creatinine ratio 0 hours r a a a a Cystatin C quotient r a a a a Note Cigarroa quotient defined as amount of contrast material (ml) baseline creatinine (mg/dl) / bodyweight (kg). Cystatin C quotient defined as amount of contrast material (ml) baseline cystatin C (mg/l) / bodyweight (kg). Significant values are in bold. BUN = blood urea nitrogen test, MDRD = Modification of Diet in Renal Disease study. a versus mean baseline creatinine. b p < 5 versus mean baseline creatinine. * * 454 AJR:202, February 2014

4 Cystatin C to Predict Contrast-Induced Nephropathy ministration, mean creatinine levels were significantly higher after 24 hours (1.16 ± 0.20 mg/dl, ) and 48 hours (1.16 ± 0.18 mg/dl, ) compared with baseline (1.04 ± 0.15 mg/dl) (Fig. 1). Similarly, creatinine levels increased from baseline (0.94 ± 0.19 mg/dl) to 24 hours (0.95 ± 0.31 mg/dl, p = 1) and 48 hours (0.98 ± 0.32 mg/dl, p = 15) in patients undergoing IV contrast administration. Incidence of Contrast-Induced Nephropathy In total, 53 of 373 patients (14.2%) had an increase in creatinine of 25% or 0.5 mg/ dl within 48 hours. CIN according to this definition was significantly more frequent in patients undergoing intraarterial contrast administration (38/190, 20%) compared with patients undergoing IV contrast administration (15/183, 8.2%) (p = 01) (Fig. 2). A total of 11 of 373 patients (2.9%) developed an increase in creatinine of 0.5 mg/ dl. Contrast-induced renal impairment according to this definition was found in seven of 190 (3.7%) patients with intraarterial contrast administration and four of 183 (2.2%) (p = 0.393) of patients undergoing IV contrast administration. Incidence of AKIN and RIFLE Forty-one of 373 (11.0%) patients fulfilled the criteria of acute renal impairment according to AKIN stage 1, whereas RIFLE risk was found in one patient. AKIN stage 1 was significantly more frequent after intraarterial (31/190, 16.3%) compared with IV contrast administration (10/183, 5.5%) (p = 01). Risk Factors and Predictors of CIN Univariate analysis To select a limited number of variables for multivariate analysis, univariate relation of creatinine increase after 48 hours to baseline data of biometry, renal function, medical history, and specific data of the procedure were analyzed (Table 2). Among biometric data, older age and smaller height were at best weakly associated with contrast-induced increases in serum creatinine. Although diabetes was not a significant risk factor, preexisting arterial hypertension was modestly associated with relative and absolute increases in creatinine ( for all analyses) after contrast administration. Compared with these baseline data, characteristics of the contrast administration procedure provided closer association to postprocedure renal impairment: increases in creatinine were significantly higher for intraarterial compared Fig. 2 Graph shows incidence of contrastinduced nephropathy (CIN), defined as increase in serum creatinine of 25% or 0.5 mg/dl within 48 hours. Percentage of Patients with IV contrast administration (, r = for absolute creatinine increase after 48 hours). Furthermore, the amount of contrast material was significantly associated with the increase in creatinine (, r = for absolute creatinine increase after 48 hours). Among the baseline parameters of renal function, neither creatinine nor creatinine clearance was associated with relative or absolute increases in creatinine levels after 24 hours and 48 hours. By contrast, baseline serum levels of BUN and cystatin C were significantly ( for all analyses) associated with relative and absolute increases in creatinine. The best predictive capabilities were provided by baseline cystatin C regarding absolute changes in creatinine after 48 hours (r = 0.358, ). Multivariate analysis On the basis of these data, multivariate association of older age; height; hypertension; diabetes; amount and route of contrast administration; and baseline values of cystatin C, BUN, and creatinine with an increase in serum creatinine within 48 hours was analyzed by multiple regression analysis. Among these variables, intraarterial contrast administration (p = 02) and high values of cystatin C () combined with low values of creatinine (p = 44) were independently associated with higher increases in serum creatinine after 48 hours (r = 0.345). With these data suggesting that patients with higher levels of cystatin C combined with low values of creatinine are at the highest risk, we evaluated the capabilities of the cystatin C/creatinine baseline ratio to predictive renal impairment after contrast administration. As shown in Table 2, this ratio provides the highest predictive capabilities All Patients 20 Intraarterial Contrast CIN p = IV Contrast 11 All Patients 16.3 Intraarterial Contrast AKIN Stage IV Contrast regarding absolute (r = 0.467, ) and relative (r = 0.486, ) increases in serum creatinine among all parameters investigated by univariate analysis. Receiver operating characteristic analyses regarding contrast-induced nephropathy after 48 hours Regarding the high clinical relevance of straightforward predictors of clinically relevant renal impairment, we performed ROC analysis regarding CIN, defined as increase in serum creatinine of at least 0.5 mg/dl or 25% 48 hours after contrast administration. Among different parameters qualified after univariate and multivariate analyses, the baseline ratio of serum cystatin C/creatinine (area under the ROC [AUC], 0.826; 95% CI, ; ) and baseline serum cystatin C (AUC, 0.715; 95% CI, ; ) were the only significant predictors of CIN after 48 hours (Fig. 3). Regarding practical implications, we determined the cutoff with a maximum sum of sensitivity and specificity: A cystatin C/creatinine ratio 1.23 at baseline provided sensitivity of 81%, specificity of 85%, and a Youden index of 0.66 regarding CIN. By contrast, neither baseline creatinine (AUC, 0.483; 95% CI, ; p = 0.751), creatinine clearance (AUC, 0.523; 95% CI, ; p = 0.671), nor baseline BUN (AUC, 0.589; 95% CI, ; p = 95) were predictive. ROC analysis regarding AKIN Similarly, acute renal impairment according to the AKIN classification within 48 hours was best predicted by the baseline ratio of serum cystatin C/creatinine (AUC, 0.867; 95% CI, ; ) and baseline serum cystatin C (AUC, 0.856; 95% CI, ; ) AJR:202, February

5 Wacker-Gußmann et al. (Fig. 4). A cystatin C/creatinine ratio greater than the threshold of 1.23 provided sensitivity of 83% and specificity of 83% regarding AKIN classification (Youden index, 0.66). Furthermore, baseline BUN (AUC, 0.636; 95% CI, ; p = 30) was a significant predictor of AKIN classification. By contrast, the amount of contrast material administered (AUC, 0.590; 95% CI, ; p = 0.149), intraarterial administration of contrast material (AUC, 0.615; 95% CI, ; p = 66), baseline creatinine (AUC, 0.592; 95% CI, ; p = 0.139), creatinine clearance (AUC, 0.474; 95% CI, ; p = 0.674), and creatinine-related Cigarroa quotient (defined as amount of contrast material [ml] baseline creatinine [mg/ dl] / bodyweight [kg]) (AUC, 0.567; 95% CI, ; p = 0.284) were not predictive. By contrast, cystatin C quotient derived from replacement of serum creatinine by serum cystatin C in the Cigarroa quotient resulted in significant prediction of AKIN classification (AUC, 0.652; 95% CI, ; p = 15). Discussion This study investigated the incidence and predictors of CIN, defined as an increase in creatinine of 25% or 0.5 mg/dl within 48 hours [13 16] in patients with baseline creatinine in the upper normal range ( mg/dl). Sensitivity Specificity Cystatin creatinine Cystatin C BUN Creatinine clearence Creatinine AUC p p = 95 p = p = Fig. 3 Graph shows area under the receiver operating characteristic curve (AUC) prediction of contrast-induced nephropathy (CIN), defined as increase in serum creatinine of 25% or 0.5 mg/dl within 48 hours. BUN = blood urea nitrogen test. The incidence of an increase in creatinine of 25% from baseline values in all 373 evaluable patients neglecting the route of contrast administration was 14.2%. An increase in serum creatinine of 25% was significantly more frequent in patients undergoing intraarterial compared with IV contrast administration (20% vs 8.2%, ). Regarding the high sensitivity of this definition in patients with creatinine 1.3 mg/dl, we also included an increase in creatinine of 0.5 mg/dl as a secondary endpoint, which was found in 3.7% of the intraarterial and in 2.2% of the IV contrast procedures. Although baseline creatinine was not predictive for renal impairment after contrast administration, baseline values of cystatin C and, to a lower degree, of BUN were significantly predictive. Multivariate analysis showed independent association of intraarterial contrast administration and higher baseline cystatin C levels with CIN. Interestingly, in combination with these parameters, low values of baseline creatinine were also independently associated with renal impairment after contrast administration. To make these findings of a prognostic interaction of baseline cystatin C and creatinine more readily available in clinical routine, we simplified multiple regression by introducing a cystatin C/creatinine ratio that provided AUCs of and to predict CIN and AKIN classification, respectively. Sensitivity Specificity Cystatin creatinine Cystatin C BUN Cigarroa cystatin C Intraarterial contrast administration AUC p p = 30 p = 15 p = 66 Fig. 4 Graph shows area under the receiver operating characteristic curve (AUC) prediction of kidney injury according to the Acute Kidney Injury definition. BUN = blood urea nitrogen test. Our results show that even patients with creatinine in the upper normal range are at substantial risk for developing CIN, particularly after intraarterial contrast administration. These data might be of clinical importance because patients with creatinine in the upper normal range usually do not receive any pharmacologic prophylaxis (e.g., theophylline or acetylcysteine [18, 19]) and do not undergo routine follow-up including laboratory control of renal parameters after contrast administration. Acute deterioration in renal function caused by radiographic agents can be mild and transient but also can result in permanent renal dysfunction leading to renal replacement therapy [10]. Although several predictive markers and risk scores for CIN have been established for patients with impaired renal function [20, 24], prediction of CIN in patients with creatinine in the upper normal range is difficult. The scarce data are mainly restricted to ICU populations [25 28]. Among these studies, the setting of the study by Polena and colleagues [27] seems to be most comparable to our results. In this study including 75 ICU patients without preexisting renal impairment, the incidence of an increase in creatinine of 25% was 18%, which is in accordance with our data. By including an otherwise comparable control group that did not undergo contrast administration and did not show an increase in creatinine, this study nicely showed that renal impairment in the study group was most likely due to contrast administration. Among the parameters and scores investigated in patients with elevated baseline creatinine, the Cigarroa quotient (amount of contrast material [ml] serum creatinine [mg/dl] body weight [kg] 1 ) [20] has been suggested as valuable predictor of CIN. Although a Cigarroa quotient > 5 ml mg/dl kg 1 has been associated with increased risk for CIN in several studies including patients with elevated values of creatinine, the Cigarroa quotient was a poor predictor of CIN in our study [18, 20]. On the basis of higher predictive capabilities of cystatin C compared with creatinine, we replaced creatinine with cystatin C in this formula, resulting in a cystatin C quotient (amount of contrast material [ml] serum cystatin C [mg/dl] body weight [kg] 1, which was found to be a significant predictor of AKIN stage 1 in our study. However, similar to CIN, AKIN stage was better predicted by the cystatin C/creatinine ratio at baseline, showing that patients with higher cystatin C 456 AJR:202, February 2014

6 Cystatin C to Predict Contrast-Induced Nephropathy combined with lower levels of creatinine are at particular risk of renal impairment after contrast administration. At first glance, the association of lower creatinine levels with increased risk in CIN might be surprising [29]. However, this increased risk has to be interpreted regarding the interaction of lower values of creatinine with already elevated cystatin C levels. Several explanations for this finding should be discussed. Regarding highly significant predictive capabilities of cystatin C alone, even improved prediction by cystatin C/creatinine suggests that the combination of elevated cystatin C with lower creatinine indicates a particular risk. Low creatinine in relation to cystatin C might be related to comorbidities with reduced muscle mass. Another explanation in this group with creatinine levels in the upper normal range might be that patients with an already decreased GFR but still able to lower creatinine by tubular secretion might be particularly prone to further renal impairment by nephrotoxic agents such as contrast material. Interestingly, a recent study [9] showed the usefulness of combining cystatin C and creatinine in a formula to optimize prediction of GFR. Regarding practical applications of these findings, a cutoff of 1.23 mg/l/mg/dl provided sensitivities and specificities of more than 80% for both CIN and AKIN. Furthermore, with regard to practical applicability, cystatin C still is more readily available in clinical routine compared with newer biomarkers of acute renal impairment, such as NGAL and KIM-1 [1, 5, 9]. By contrast, a cutoff point in creatinine beyond which the risk for CIN increases is difficult to define because the relationship of serum creatinine and GFR is not linear [20]. A normal serum creatinine can be associated with a markedly reduced GFR, which makes creatinine an unreliable indicator of acute changes in kidney function [2, 30]. Compared with creatinine, cystatin C offers greater sensitivity in detecting an abnormal GFR with similar specificity [3 5]. For instance, both for IgA nephropathy and for diabetic nephropathy, serum cystatin C was associated with the stages of nephropathy earlier than creatinine [1, 31]. Similar considerations seem to apply for the role of cystatin C in predicting CIN: Because cystatin C increases earlier (maximum 24 hours) compared with creatinine (still increases after 48 hours) in acute renal damage [32], cystatin C is likely to indicate the development of CIN earlier than creatinine. Regarding postinterventional surveillance, the absence of an increase in cystatin C after 24 hours might better predict that renal impairment after contrast administration is not likely. Although the predictive capability of cystatin C in this particular cohort is supported by several endpoints, the marked impact of intraarterial contrast administration has to be interpreted with caution: Despite being an independent predictor in a multivariate analysis with a limited number of variables, the impact of intraarterial contrast administration per se might be overestimated: It has been repeatedly shown that intraarterial contrast administration is likely to be associated with a number of well-known risk factors for CIN, such as a high amount of contrast material hypertension, older age, and more severe preexisting renal impairment. This also applies for our patients, particularly regarding the amount of contrast material, which was twice as much for intraarterial compared with IV contrast administration (304 vs 151 ml, ). The strengths of the study include the analysis of two prospectively acquired databases, providing appropriate power to compare several markers of renal function and their association with CIN after IV and intraarterial contrast administration. Although the patients included with borderline renal function according to serum creatinine and creatinine-based GFR estimation equations are of particular interest, incidence and predictors of CIN in these patients is not well evaluated. Most of the studies on CIN prophylaxis have investigated patients at high risk with already increased serum creatinine [33, 34]. Including an appropriate number of patients, this study showed significant risk of CIN in a substantial number of patients, particularly after intraarterial contrast administration. Furthermore, the database provided appropriate power to characterize risk factors of CIN in multivariate analysis. Because the current definition of CIN might be particularly sensitive in this group of patients, we also included the AKIN definition for acute renal impairment, which is rarely used in this context. Analyses regarding AKIN stage confirmed the main findings of the study. Finally, the investigated cohort is of high sociomedical interest because a similar cohort with creatinine-based estimates of GFR of ml/min 1.73 m 2 represents about 10.9% of the U.S. population [9, 35]. Despite an acceptable number of patients included in two hospitals, it has to be clearly stated that our findings, particularly the cystatin C/creatinine ratio, need to be evaluated in further multicentric studies also including other novel biomarkers of early renal impairment, such as NGAL, KIM-1, and L-FABP. Furthermore, the usefulness of cystatin C and the cystatin C/creatinine ratio might be restricted to the group of patients within the upper normal range of serum creatinine. In general, it is difficult to provide definite conclusions regarding incidence and prediction of renal impairment defined by changes of surrogate markers of renal function in a population with creatinine levels in the upper normal range and GFR values between 40 and 140 ml/min according to the Modification of Diet in Renal Disease formula. It has been repeatedly shown that surrogate markers of GFR, including these formulas derived from cohorts with chronic renal impairment, might be false-positive in nonselected populations. This might also apply for cystatin C and the cystatin C/creatinine ratio. Therefore, the usefulness of our findings should be validated in larger population-based cohorts also including other intraarterial procedures in addition to coronary angiography. This might provide more generalizable data. Finally, the question remains whether pharmacologic prophylaxis beyond a certain cutoff point of the cystatin C/creatinine ratio is associated with a lower risk for developing CIN. Conclusion CIN occurs in a substantial number of patients with creatinine within the upper normal range. CIN is significantly more frequent after intraarterial than after IV contrast administration. The cystatin C/creatinine ratio provided the highest predictive capability regarding renal impairment after contrast administration. Regarding widespread availability, cystatin C and the cystatin C/creatinine-ratio are straightforward and may be useful in screening for patients at particular risk for CIN. Acknowledgment We thank Juliane Preiss for the data collection. References 1. Gonzalez F, Vincent F. Biomarkers for acute kidney injury in critically ill patients. Minerva Anestesiol 2012; 78: Shimizu A, Horikoshi S, Rinno H, Kobata M, Saito K, Tomino Y. Serum cystatin C may predict the early prognostic stages of patients with type 2 diabetic nephropathy. J Clin Lab Anal 2003; 17: Herget-Rosenthal S, Marggraf G, Hüsing J, et al. Early detection of acute renal failure by serum cystatin C. Kidney Int 2004; 66: AJR:202, February

7 Wacker-Gußmann et al. 4. Newman DJ, Thakkar H, Edward RG, et al. Serum cystatin C measured by automated immunoassay: a more sensitive marker of changes in GFR than serum creatinine. Kidney Int 1995; 47: Zhang Z, Lu B, Sheng X, Jin N. Cystatin C in prediction of acute kidney injury: a systemic review and meta-analysis. Am J Kidney Dis 2011; 58: Keller T, Messow CM, Lubos E, et al. Cystatin C and cardiovascular mortality in patients with coronary artery disease and normal or mildly reduced kidney function: results from the Atherogene Study. Eur Heart J 2009; 30: Tenstad O, Roald AB, Grubb A, Aukland K. Renal handling of radio labelled human cystatin C in the rat. Scand J Clin Lab Invest 1996; 56: Abrahamson M, Olafsson I, Palsdottir A, et al. Structure and expression of the human cystatin C gene. Biochem J 1990; 268: Inker LA, Schmid CH, Tighiouart H, et al.; CKD- EPI Investigators. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 2012; 367: Hou SH, Bushinsky DA, Wish JB, Cohen JJ, Harrington JJ. Hospital-acquired renal insufficiency: a prospective study. Am J Med 1983; 74: Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis 2002; 39: Levy EM, Viscoli CM, Horwitz RI. High mortality in contrast-associated renal failure is not explained by co-morbidity. (abstract) J Am Soc Nephrol 1996; 5(suppl): Harjai KJ, Raizada A, Shenoy C, et al. A comparison of contemporary definitions of contrast nephropathy in patients undergoing percutaneous coronary intervention and a proposal for a novel nephropathy grading system. Am J Cardiol 2008; 101: McCullough PA, Wolyn R, Rocher LL, Levin RN, O Neil WW. Acute renal failure after coronary intervention: incidence, risk-factors, and relationship to mortality. Am J Med 1997; 103: McCullough PA, Stacul F, Becker CR, et al. Contrast-Induced Nephropathy (CIN) Consensus Working Panel: executive summary. Rev Cardiovasc Med 2006; 7: Barrett BJ, Parfrey PS. Prevention of nephrotoxicity induced by radiocontrast agents. N Engl J Med 1994; 331: Manske CL, Sprafka JM, Strony JT, Wang Y. Contrast nephropathy in azotemic diabetic patients undergoing coronary angiography. Am J Med 1990; 89: Huber W, Schipek C, Ilgmann K, et al. Effectiveness of theophylline prophylaxis of renal impairment after coronary angiography in patients with chronic renal insufficiency. Am J Cardiol 2003; 91: Tepel M, van der Giet M, Schwarzfeld C, Laufer U, Liermann D, Zidek W. Prevention of radiographic-contrast-agent-induced reductions in renal function by acetylcysteine. N Engl J Med 2000; 343: Cigarroa RG, Lange RA, Williams RH, Hillis RD. Dosing of contrast material to prevent contrast nephropathy in patients with renal disease. Am J Med 1989; 86: Shemesh O, Golbetz H, Kriss JP, Myers BD. Limitations of creatinine as a filtration marker in glomerulopathic patients. Kidney Int 1985; 28: Levey AS, Coresh J, Greene T, et al. Expressing the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 2007; 53: Mehta RL, Kellum JA, Shah SV, et al.; Acute Kidney Injury Network. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007; 11:R Mehran R, Aymong ED, Nikolsky E, et al. A simple risk score for prediction of contrast-induced nephropathy after percutaneous coronary intervention: development and initial validation. J Am Coll Cardiol 2004; 44: Huber W, Jeschke B, Page M, et al. Reduced incidence of radiocontrast-induced nephropathy in ICU patients under theophylline prophylaxis: a prospective comparison to series of patients at similar risk. Intensive Care Med 2001; 27: Chew DP, Astley C, Molloy D, Vaile J, De Pasquale CG, Alyward P. Morbidity, mortality and economic burden of renal impairment in cardiac intensive care. Intern Med J 2006; 36: Polena S, Yang S, Alam R, et al. Nephropathy in critically ill patients without preexisting renal disease. Proc West Pharmacol Soc 2005; 48: Rashid AH, Brieva JL, Stokes B. Incidence of contrast-induced nephropathy in intensive care patients undergoing computerised tomography and prevalence of risk factors. Anaesth Intensive Care 2009; 37: Fujisawa N, Yashiro M, Segawa H, Kadoya Y, Kamata T. Discrepancy between cystatin C and creatinine points leading to a diagnosis of postrenal acute kidney injury and its reversibility: three case reports. Clin Exp Nephrol 2010; 14: Bellomo R, Kellum JA, Ronco C. Defining acute renal failure: physiological principles. Intensive Care Med 2004; 30: Tomino Y, Suzuki S, Gohda T, et al. Serum cystatin C may predict the prognostic stages of patients with IgA nephropathy prior to renal biopsy. J Clin Lab Anal 2001; 15: Rickli H, Benou K, Ammann P. Time course of serial cystatin C levels in comparison with serum creatinine after application of radiocontrast media. Clin Nephrol 2004; 61: Shaker OG, El-Shehaby A, El-Khatib M. Early diagnostic markers for contrast nephropathy in patients undergoing coronary angiography. Angiology 2010; 61: Rihal CS, Textor SC, Grill DE, et al. Incidence and prognostic importance of acute renal failure after percutaneous coronary intervention. Circulation 2002; 105: Stevens LA, Padala S, Levey AS. Advances in glomerular filtration rate-estimating equations. Curr Opin Nephrol Hypertens 2010; 19: AJR:202, February 2014