American Journal of Kidney Diseases
Volume 54, Issue 6 , Pages 1025-1033, December 2009

Serum Cystatin C for Prediction of Dialysis Requirement or Death in Acute Kidney Injury: A Comparative Study

  • Mary C. Perianayagam, PhD

      Affiliations

    • Department of Medicine, Division of Nephrology, Kidney and Dialysis Research Laboratory, St Elizabeth's Medical Center, Boston, MA
    • Department of Medicine, Tufts University School of Medicine, Boston, MA
    • ,
  • Victor F. Seabra, MD

      Affiliations

    • Department of Medicine, Division of Nephrology, Kidney and Dialysis Research Laboratory, St Elizabeth's Medical Center, Boston, MA
    • Department of Medicine, Tufts University School of Medicine, Boston, MA
    • ,
  • Hocine Tighiouart, MS

      Affiliations

    • Department of Medicine, Tufts University School of Medicine, Boston, MA
    • Biostatistics Research Center, Tufts Medical Center, Boston, MA
    • ,
  • Orfeas Liangos, MD

      Affiliations

    • Department of Medicine, Division of Nephrology, Kidney and Dialysis Research Laboratory, St Elizabeth's Medical Center, Boston, MA
    • Department of Medicine, Tufts University School of Medicine, Boston, MA
    • ,
  • Bertrand L. Jaber, MD, MS

      Affiliations

    • Department of Medicine, Division of Nephrology, Kidney and Dialysis Research Laboratory, St Elizabeth's Medical Center, Boston, MA
    • Department of Medicine, Tufts University School of Medicine, Boston, MA
    • Corresponding Author InformationAddress correspondence to Bertrand L. Jaber, MD, MS, St Elizabeth's Medical Center, 736 Cambridge St, Boston, MA 02135

Received 4 April 2009; accepted 9 June 2009. published online 07 August 2009.

Article Outline

Background

Serum cystatin C has emerged as a new and potentially more reliable marker of kidney function. However, its utility and performance in patients with acute kidney injury (AKI), particularly for the prediction of dialysis requirement, is not well known.

Study Design

Prospective cohort study.

Settings & Participants

Adult patients with AKI enrolled at 2 academic medical centers, at time of nephrology consultation.

Predictors

Serum cystatin C (primary predictor), serum creatinine, and serum urea nitrogen levels and 24-hour urine output measured at enrollment.

Outcomes

The composite of dialysis requirement or in-hospital death.

Covariates

Acute Physiology and Chronic Health Evaluation II (APACHE II) score, liver disease, sepsis, and mechanical ventilation.

Results

200 participants were enrolled for this analysis. Mean age was 65 years, 55% were men, and mean APACHE II score was 20. In unadjusted analyses, increases in serum cystatin C (odds ratio [OR], 1.87; 95% confidence interval [CI], 1.36 to 2.59), serum creatinine (OR, 1.53; 95% CI, 1.12 to 2.09), and serum urea nitrogen levels (OR, 1.84; 95% CI, 1.34 to 2.54) were associated with a higher odds (per 1-SD increase) for the composite outcome, whereas greater urine output (OR, 0.56; 95% CI, 0.39 to 0.80) was associated with lower odds. These associations persisted after adjustment for APACHE II score. The addition of serum cystatin C, serum creatinine, and serum urea nitrogen levels or urine output to a basic model entailing APACHE II score, liver disease, sepsis, and assisted mechanical ventilation improved its prediction, evidenced by increases in areas under a receiver operator characteristic curve from 0.816 to 0.829, 0.826, 0.837, and 0.836, respectively. However, there was no significant difference between each of these models.

Limitations

Observational study, single serum cystatin C measurement.

Conclusion

In patients with AKI, serum cystatin C level performs similarly to serum creatinine level, serum urea nitrogen level, and urine output for predicting dialysis requirement or in-hospital death. Larger studies are needed to confirm these findings.

Index Words: Acute kidney injury, acute renal failure, biomarker, serum, cystatin C, creatinine, serum urea nitrogen, urine output, epidemiology, prognosis, dialysis, in-hospital death

 

Acute kidney injury (AKI) is a commonly observed complication in hospitalized patients1, 2, 3, 4 and is associated with increased mortality, hospital length of stay, costs, and post–acute care resource utilization.4, 5 The currently accepted marker for diagnosing and staging AKI is serum creatinine level,6, 7 which is not an ideal glomerular filtration marker in the acute setting because of several factors, including extracellular fluid volume expansion, muscle wasting, and malnutrition.8, 9, 10, 11 These factors can slow down the rate of increase in serum creatinine level despite significant kidney injury and delay diagnosis, which in turn might hamper timely nephrology consultation and provision of supportive care.

Cystatin C is a low-molecular-weight (13.4 kDa) cysteine proteinase inhibitor synthesized by all nucleated cells. It is filtered freely by glomeruli and completely metabolized by proximal tubules. Cystatin C can be assayed easily, and routine laboratory measurement increasingly is becoming available to clinicians. In the future, this marker is expected to complement or replace serum creatinine level to more accurately detect a decrease in kidney function, particularly in the setting of chronic kidney disease (CKD).12, 13, 14 In recent years, cystatin C has emerged as a potential new early detection marker for AKI.15, 16, 17, 18, 19, 20 However, few small studies have assessed its prognostic value for predicting adverse outcomes in patients with AKI.21, 22

In the present study, we explored the potential usefulness of serum cystatin C level for predicting adverse clinical outcomes in a large prospective cohort of hospitalized patients with an established diagnosis of AKI at the time of nephrology consultation and compared its performance characteristics with more traditional measures of kidney function, mainly serum creatinine level, serum urea nitrogen (SUN) level, and timed urine output. We hypothesized that serum cystatin C level would outperform these other indices in predicting adverse outcomes.

Back to Article Outline

Methods 

Study Design and Setting 

This is an ancillary study to an ongoing prospective cohort study of hospitalized patients with AKI being conducted at 2 tertiary acute-care hospitals located in Boston, MA. The parent study is examining the association of polymorphism of candidate genes with adverse clinical outcomes in patients with AKI.

Study Participants 

All consecutive hospitalized adult patients with AKI for whom nephrology consultation was requested were eligible for study enrollment. AKI was defined as an incremental increase in serum creatinine level by 0.5, 1.0, or 1.5 mg/dL from a baseline of 1.9 or less, 2.0 to 4.9, and 5.0 or greater mg/dL, respectively.1 This definition was adopted before the development of the new consensus definition of AKI.7

Exclusion criteria were age younger than 18 years, pregnancy, long-term maintenance dialysis therapy, an organ transplant within the previous year, and acute obstructive uropathy. Patients receiving palliative or terminal care also were excluded. Written informed consent was obtained from study participants or next of kin. The institutional review board of each participating center approved the study protocol.

Data Source 

Clinical Data 

Medical records of study participants were reviewed prospectively to retrieve hospitalization data, including baseline demographic characteristics, and coexisting conditions. At study enrollment, kidney-related variables were recorded, including serum creatinine level, SUN level, and 24-hour urine output, before enrollment. At enrollment, severity of illness was measured by calculating the Acute Physiology and Chronic Health Evaluation II (APACHE II) score.23 The presence of sepsis was ascertained by using systemic inflammatory response syndrome criteria.24 Preexisting CKD was defined on the basis of baseline estimated glomerular filtration rate less than 60 mL/min/1.73 m2, calculated by using the 4-variable Modification of Diet in Renal Disease (MDRD) Study equation (not isotope dilution mass spectrometry traceable).25

Blood Sampling and Measurements 

At enrollment, venous blood was drawn. Serum was separated within 30 minutes after collection, and aliquots were stored at −80°C. Cystatin C was measured by means of immunonephelometry using the Behring Nephelometer II System (Dade Behring, Deerfield, IL). The cystatin C lower limit of detection was 0.05 mg/L, and average intra-assay and interassay coefficients of variation were less than 2.5% and less than 3.5%, respectively. Creatinine was measured by means of a modified Jaffé method, and SUN, by means of an enzymatic rate method, using a Beckman DXC 800 analyzer (Beckman Coulter, Brea, CA).

Predictor Variables 

The predictive variable of interest was serum cystatin C level, which was compared with 3 other measures of kidney function in patients with AKI: serum creatinine level, SUN level, and 24-hour urine output.

Outcome Measures 

The primary outcome of interest was the composite of dialysis therapy requirement (initiated on the day of enrollment or thereafter) or in-hospital death. This composite end point was chosen because it takes into consideration survival bias for dialysis requirement.

Statistical Analysis 

Continuous variables were described as mean ± SD, and categorical variables, as percentages. Comparisons between serum cystatin C quartiles were made by using analysis of variance test (with a global P value) for continuous variables and χ2 test for categorical variables. Figure box plots of serum cystatin C values stratified by APACHE II score quartiles were constructed using SPSS software (version 12.0.1; SPSS, Chicago, IL) and analyzed by using analysis of variance test.

The functional form of serum cystatin C level, serum creatinine level, SUN level, and 24-hour urine output with the composite outcome also was examined and graphically displayed by using restricted cubic splines functions with 4 knots.26 Restricted cubic spline functions are piecewise cubic polynomials used in curve fitting to approximate a wide variety of functions. Plots of the restricted splines were constructed using the Design package (version 2.1.2)27 for R software (version 2.8.1; R Foundation for Statistical Computing, Vienna, Austria).28 Ninety-five percent confidence intervals (CIs) for areas under the curve were calculated.

Separate logistic regression analyses were used to examine the association of serum cystatin C level, serum creatinine level, SUN level, or 24-hour urine output with the composite outcome. The models included the following baseline covariates: APACHE II score, presence of sepsis, preexisting liver disease, and requirement for assisted mechanical ventilation. These covariates were chosen based on their association with the composite outcome on univariate analysis. We accepted P ≤ 0.05 as the criterion to introduce covariates in the multivariable model. We ran additional logistic regression analyses by using a modified APACHE II score in which the points attributed to serum creatinine level were removed. We also formally tested for interactions between serum cystatin C level and several covariates for the outcome of interest. Results of logistic regression analyses are shown as odds ratios (ORs; per 1-SD increase) with 95% CIs.

The diagnostic performance of serum cystatin C level for predicting the composite outcome was described by using the area under a receiver operator characteristic curve (AUC) and compared with the performance of serum creatinine level, SUN level, and 24-hour urine output. For the latter variable, we used the negative value because of its inverse association with adverse outcomes. Serum creatinine level at study enrollment was chosen to serve as a standard for comparison. To test formally whether various models (ie, differences in AUCs between the various models) are statistically more or less discriminating, the nonparametric method of DeLong et al29 was used.

All statistical analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC). Differences were considered statistically significant at P < 0.05.

Back to Article Outline

Results 

Cohort Characteristics Stratified by Serum Cystatin C Quartiles 

A total of 235 participants were enrolled in the parent cohort study between November 2003 and May 2007. The present study includes 200 participants who underwent measurement of serum cystatin C and, if indicated, were initiated on dialysis therapy after study enrollment. Excluded were 33 participants enrolled 1 day or more after dialysis therapy initiation, 1 participant without cystatin C measurement, and 1 participant with multifactorial causes of AKI that included obstructive uropathy. Mean age was 65 years, 55% were men, and 90% were of white ethnicity. Mean APACHE II score was 20, 73% were in the intensive care unit, and 25% required assisted mechanical ventilation. At enrollment, mean serum creatinine level was 3.4 mg/dL, mean SUN level was 63 mg/dl, 24-hour urine output was 1.3 L/d, and serum cystatin C level was 3.0 mg/L. The in-hospital mortality rate was 25%, 32% required dialysis therapy, and 42% experienced the composite outcome of dialysis therapy requirement or in-hospital death.

Table 1 lists enrollment characteristics of the cohort according to serum cystatin C quartiles. In brief, there were no significant differences in characteristics of the study participants among cystatin C quartiles, except for a greater prevalence of preexisting CKD, higher APACHE II score, lower hemoglobin level, and greater white blood cell count in the higher quartiles. There was a nonsignificant trend toward a greater prevalence of diabetes mellitus in the higher cystatin C quartiles. Similarly, at enrollment, serum creatinine and SUN levels were greater and urine output was lower in the higher cystatin C quartiles. Serum cystatin C levels increased in tandem with APACHE II score quartiles (Fig 1; P = 0.004). This trend persisted when we used the modified APACHE II score in which points attributed to serum creatinine level were subtracted (P = 0.03). Of note, serum creatinine levels did not increase across APACHE II quartiles (data not shown). Finally, greater serum cystatin C quartiles were associated with greater dialysis requirement and the composite of dialysis requirement or in-hospital death (Table 1).

Table 1. Characteristics and Outcomes of the Acute Kidney Injury Cohort According to Serum Cystatin C Quartiles
Serum Cystatin C (mg/L)P
1.7 ± 0.4; Quartile 1 (n = 51)2.5 ± 0.2; Quartile 2 (n = 49)3.2 ± 0.2; Quartile 3 (n = 50)4.6 ± 0.9; Quartile 4 (n = 50)
Enrollment variables
Age (y)61±1668±1368±1364±170.06
Men (%)595156520.9
White ethnicity (%)868494960.1
APACHE II score18±720±719±522±60.03
Assisted mechanical ventilation (%)252918280.6
Sepsis (%)454936460.6
Preexisting chronic kidney disease (%)⁎44687880<0.001
Congestive heart failure (%)141614140.9
Liver disease (%)481480.3
Diabetes mellitus (%)313754500.07
Intensive care unit setting (%)657372820.3
Serum albumin (g/dL)2.8±12.5±0.72.5±0.82.5±0.80.3
White blood cells (103/μL)11.0±6.110.8±4.711.1±6.114.1±7.90.03
Hemoglobin (g/dL)10.9±2.210.7±2.99.9±1.610.0±1.50.03
Serum creatinine (mg/dL)2.5±1.13.0±1.03.7±1.94.5±1.4<0.001
Serum urea nitrogen (mg/dL)42±2054±1966±3089±33<0.001
Urine output (L/d)†1.8±1.51.4±1.01.2±0.80.9±0.90.001
Outcome variables
Dialysis requirement (%)16223256<0.001
In-hospital death (%)183518280.1
Dialysis requirement or in-hospital death (%)273938620.004

Note: Continuous variables are presented as mean ± SD, and categorical variables, as percentages. Conversion factors for units: serum cystatin C in mg/L to nmol/L, ×74.9; serum creatinine in mg/dL to μmol/L, ×88.4; serum urea nitrogen in mg/dL to mmol/L, ×0.357; serum albumin in g/dL to g/L, ×10; hemoglobin in g/dL to g/L, ×10. No conversion necessary for white blood cells in 103/μL and 109/L.

Abbreviation: APACHE, Acute Physiology and Chronic Health Evaluation.

⁎Data for preexisting chronic kidney disease were available for 196 patients (50 in quartile 1, 47 in quartile 2, 49 in quartile 3, and 50 in quartile 4).

†Data for urine output were available for 197 patients (51 in quartile 1, 48 in quartile 2, 50 in quartile 3, and 48 in quartile 4).

  • View full-size image.
  • Figure 1. 

    Serum cystatin C level stratified by Acute Physiology and Chronic Health Evaluation II (APACHE II) score quartiles. P = 0.004 by means of analysis of variance test. Box plots show 25th, 50th (median), and 75th percentiles. Conversion factor for serum cystatin C in mg/L to nmol/L, ×74.9.

Serum Cystatin C and Prediction of Dialysis Requirement or In-hospital Death 

We constructed a restrictive cubic spline to ascertain the crude relationship between serum cystatin C level and dialysis requirement or in-hospital death. Increasing serum cystatin C level was associated with the composite outcome (P < 0.001; Fig 2A). A similar association was observed for serum creatinine (P = 0.04; Fig 2B) and SUN levels (P = 0.01; Fig 2C). In contrast, increasing urine output was associated with a progressively lower probability of dialysis therapy requirement or in-hospital death (P < 0.001; Fig 2D).

  • View full-size image.
  • Figure 2. 

    Restrictive cubic spline shows the unadjusted relationship (with 95% confidence bands) of (A) serum cystatin C (P = 0.004), (B) serum creatinine (P = 0.04), and (C) serum urea nitrogen levels (P = 0.005), and (D) urine output (P = 0.002) at enrollment with predicted probability of dialysis therapy requirement or in-hospital death. For each spline, P for linearity test is greater than 0.10, suggesting no deviation from linearity. Vertical lines show the spike-histogram of the variable shown on the x-axis. Conversion factors for units: serum cystatin C in mg/L to nmol/L, ×74.9; serum creatinine in mg/dL to μmol/L, ×88.4; serum urea nitrogen in mg/dL to mmol/L, ×0.357.

Results of logistic regression analyses are listed in Table 2. In unadjusted analyses, increases in serum cystatin C, serum creatinine, and SUN levels were associated with a greater odds (per 1-SD increase) for the composite outcome of dialysis therapy requirement or in-hospital death, whereas greater urine output was associated with lower odds, with a magnitude of 1.87-, 1.53-, 1.84-, and 0.56-fold, respectively (Table 2). Other variables associated with the composite outcome were APACHE II score (OR, 4.21 per 1-SD increase; 95% CI, 2.71 to 6.55; P < 0.001), sepsis (OR, 2.88; 95% CI, 1.61 to 5.16; P < 0.001), presence of liver disease (OR, 3.79; 95% CI, 1.28 to 11.20; P = 0.02), and requirement for assisted mechanical ventilation (OR, 6.43; 95% CI, 3.13 to 13.23; P < 0.001). The association of serum cystatin C, serum creatinine, and SUN levels and urine output with the composite outcome persisted after adjustment for either APACHE II score (model 1, Table 2) or APACHE II score, presence of liver disease, sepsis, and assisted mechanical ventilation (model 2, Table 2), with the exception of serum creatinine level, which showed a trend toward significance after adjustment for APACHE II score alone (P = 0.08). In separate models, use of the modified APACHE II score in which points attributed to serum creatinine level were subtracted did not affect these point estimates (data not shown). There was no interaction between serum cystatin C level and either preexisting diabetes mellitus, CKD, APACHE II score, or white blood cell count for the outcome of interest (data not shown). Finally, despite colinearity among the 4 predictor variables, when entered together into 1 model, only urine output remained independently associated with the composite outcome (OR, 0.63; 95% CI, 0.43 to 0.92; P = 0.02). This association persisted after adding the 4 covariates to the model: APACHE II score, liver disease, sepsis, and mechanical ventilation (OR, 0.65; 95% CI, 0.43 to 0.97; P = 0.03).

Table 2. Association of Serum Cystatin C and Other Indices of Acute Kidney Injury With the Composite Outcome of Dialysis Requirement or In-hospital Death
Serum Cystatin CSerum CreatinineSerum Urea NitrogenUrine Output
Unadjusted analysis1.87(1.36-2.59)1.53(1.12-2.09)1.84(1.34-2.54)0.56(0.39-0.80)
P < 0.001P = 0.008P < 0.001P = 0.002
Adjusted analysis
Model 11.62(1.11-2.36)1.40(0.97-2.04)1.75(1.19-2.57)0.61(0.43-0.88)
P = 0.01P = 0.08P = 0.004P = 0.009
Model 21.71(1.15-2.54)1.61(1.08-2.41)1.84(1.24-2.73)0.59(0.41-0.87)
P = 0.008P = 0.02P = 0.002P = 0.007

Note: Values shown as odds ratio per 1-SD increase (95% confidence interval). Model 1 includes Acute Physiology and Chronic Health Evaluation (APACHE) II score. Model 2 includes APACHE II score, liver disease, sepsis, and assisted mechanical ventilation.

Comparative Prognostic Performance of Serum Cystatin C With Traditional Indices of AKI 

Table 3 lists AUCs for serum cystatin C, serum creatinine, SUN, and 24-hour urine output for prediction of the composite outcome. All 4 variables performed similarly, evidenced by AUCs of 0.654, 0.602, 0.644, and 0.665, respectively.

Table 3. Diagnostic Accuracy of Serum Cystatin C and Other Indices of Acute Kidney Injury for Prediction of Dialysis Requirement or In-hospital Death
Prediction ModelAUC (95% confidence interval)
Model 10.813(0.753-0.872)⁎
Model 20.816(0.756-0.876)⁎
Serum cystatin C0.654(0.577-0.731)⁎
Serum cystatin C + model 10.821(0.763-0.880)⁎
Serum cystatin C + model 20.829(0.771-0.886)⁎
Serum creatinine0.602(0.521-0.683)†
Serum creatinine + model 10.820(0.761-0.878)⁎
Serum creatinine + model 20.826(0.767-0.884)⁎
SUN0.644(0.565-0.722)⁎
SUN + model 10.829(0.773-0.886)⁎
SUN + model 20.837(0.782-0.892)⁎
Urine output0.665(0.589-0.742)⁎
Urine output + model 10.829(0.772-0.886)⁎
Urine output + model 20.836(0.780-0.892)⁎

Note: Model 1 includes APACHE II score. Model 2 includes APACHE II score, liver disease, sepsis, and assisted mechanical ventilation.

Abbreviations: APACHE, Acute Physiology and Chronic Health Evaluation; AUC, area under the receiver operator characteristic curve; SUN, serum urea nitrogen.

⁎P < 0.001 (2-sided P testing difference with AUC of 0.50).

†P = 0.01 (2-sided P testing difference with AUC of 0.50).

We next explored the potential added value of each variable to a basic prediction model entailing APACHE II score, presence of liver disease, sepsis, and assisted mechanical ventilation, which yielded an already impressive AUC of 0.816. The addition of each of the 4 markers, serum cystatin C level, serum creatinine level, SUN level, or urine output, to the model improved its prediction, evidenced by increases in AUCs from 0.816 to 0.829, 0.826, 0.837, and 0.836, respectively. However, tested formally, there was no significant difference between each of these models (Table 3). Finally, use of the modified APACHE II score in these models did not affect AUCs compared with models using the standard APACHE II score (data not shown).

Back to Article Outline

Discussion 

Serum creatinine level is an established marker for diagnosing and monitoring the course of AKI. In recent years, several studies have suggested that serum cystatin C level might be superior to serum creatinine level for the early detection of AKI.15, 16, 17, 18, 19, 20 In the present study, we evaluate the prognostic utility of serum cystatin C level, obtained at the time of nephrology consultation, in patients with an established diagnosis of AKI for predicting the need for dialysis therapy. To overcome the potential for survival bias, we chose a composite end point of dialysis therapy requirement or in-hospital death. Serum cystatin C level performed similarly to the other 3 indices of AKI, serum creatinine level, SUN level, and 24-hour urine output, but did not surpass them. We also explored the potential added value of each of these predictor variables to a basic prediction model comprising APACHE II score, sepsis, liver disease, and assisted mechanical ventilation. We found that the addition of serum cystatin C level, serum creatinine level, SUN level, or urine output improves the diagnostic performance of the model for predicting the composite outcome.

Our study objectives are of particular relevance to practicing nephrologists, who often encounter patients with an established diagnosis of AKI, and for whom improved ability to predict such adverse clinical outcomes as dialysis therapy requirement, for example, might potentially facilitate its timely initiation. To our knowledge, this is the first large observational study testing the hypothesis of how serum cystatin C level performs compared with serum creatinine level, SUN level, and urine output for the prediction of adverse outcomes in patients with AKI. A previously published smaller study of 73 patients with nonoliguric AKI found urine cystatin C level to be a promising marker for predicting dialysis requirement (AUC = 0.860).22 In another study, serum cystatin C measured serially in 202 critically ill patients during the first 3 days of admission to the ICU failed to show good performance in predicting mortality (AUC = 0.624), but was slightly better than serum creatinine level (AUC = 0.598).21 In a large observational study of 618 critically ill patients with AKI, greater SUN level and lower serum creatinine level were associated with increased risk of in-hospital mortality.30

There are several limitations to our study that should be noted. We used a single measurement of serum cystatin C, which took place at the time of nephrology consultation and introduces potential selection bias in terms of timing of the measurement. Incorporation of day-to-day changes in serum cystatin C levels clearly would have enhanced the predictive power of our models. An important limitation of serum cystatin C level is that although it is less influenced by age, sex, and muscle mass compared with serum creatinine level,31, 32, 33 it still can be affected by these patient variables.34 Moreover, serum cystatin C levels can be affected by levels of glucocorticoids,35 thyroid hormones,36, 37, 38 and insulin,39 as well as such markers of inflammation as white blood cell count and C-reactive protein level.40 The potential influence of these factors is of particular importance in critically ill patients with AKI and might confound our results. Although no circulating hormonal levels were measured in the present study, we found that greater cystatin C levels were associated with higher APACHE II scores, greater white blood cell counts, and lower hemoglobin levels, representing direct and indirect measures of inflammation and acute-phase response. Increased serum cystatin C level might reflect solely the underlying inflammatory state. Conversely, cystatin C level might be greater because patients with worse inflammation have worse kidney function. Our data do not allow us to decipher this conundrum, and future studies will have to address this question.

There is emerging literature of the potential value of novel biomarkers for the early detection of AKI before there is a noticeable increase in serum creatinine level.41, 42 Notwithstanding these advances, there is limited knowledge about the prognostic utility of such biomarkers in patients with established AKI for prediction of adverse clinical outcomes,43 and additional research is needed to clarify their potential role for predicting dialysis therapy requirement.

In conclusion, in a hospital-based cohort of patients with AKI, serum cystatin C measured at nephrology consultation performed similarly to other traditional indices of kidney injury in predicting dialysis therapy requirement or in-hospital death. However, each of the 4 markers, serum cystatin C level, serum creatinine level, SUN level, and urine output, improved the diagnostic performance of a predicting model that included APACHE II score for the prediction of that composite outcome. The potential prognostic value of serum cystatin C level needs to be validated in a larger sample size and across a broad spectrum of patients with AKI.

Back to Article Outline

Acknowledgements 

The authors thank the patients and their next of kin for participating in this study.

Support: This study was funded in part by grants from the National Institutes of Health (DK065102 and DK077751 to Dr Jaber) and was made possible in part by an International Society of Nephrology Commission for the Global Advancement of Nephrology (COMGAN) Fellowship awarded to Dr Seabra. Dr Liangos is supported in part by a grant from the American Heart Association.

Financial Disclosure: None.

Back to Article Outline

References 

  1. Hou SH, Bushinsky DA, Wish JB, Cohen JJ, Harrington JT. Hospital-acquired renal insufficiency: A prospective study. Am J Med. 1983;74:243–248
  2. Levy EM, Viscoli CM, Horwitz RI. The effect of acute renal failure on mortality (A cohort analysis). JAMA. 1996;275:1489–1494
  3. Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis. 2002;39:930–936
  4. Liangos O, Wald R, O'Bell JW, Price L, Pereira BJ, Jaber BL. Epidemiology and outcomes of acute renal failure in hospitalized patients: A national survey. Clin J Am Soc Nephrol. 2006;1:43–51
  5. Chertow GM, Burdick E, Honour M, Bonventre JV, Bates DW. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol. 2005;16:3365–3370
  6. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute renal failure—Definition, outcome measures, animal models, fluid therapy and information technology needs: The Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care. 2004;8:R204–R212
  7. Mehta RL, Kellum JA, Shah SV, et al. Acute Kidney Injury Network: Report of an initiative to improve outcomes in acute kidney injury. Crit Care. 2007;11:R31
  8. Moran SM, Myers BD. Course of acute renal failure studied by a model of creatinine kinetics. Kidney Int. 1985;27:928–937
  9. Star RA. Treatment of acute renal failure. Kidney Int. 1998;54:1817–1831
  10. Bonventre JV, Weinberg JM. Recent advances in the pathophysiology of ischemic acute renal failure. J Am Soc Nephrol. 2003;14:2199–2210
  11. Rosner MH, Okusa MD. Acute kidney injury associated with cardiac surgery. Clin J Am Soc Nephrol. 2006;1:19–32
  12. Kottgen A, Selvin E, Stevens LA, Levey AS, Van Lente F, Coresh J. Serum cystatin C in the United States: The Third National Health and Nutrition Examination Survey (NHANES III). Am J Kidney Dis. 2008;51:385–394
  13. Premaratne E, MacIsaac RJ, Finch S, Panagiotopoulos S, Ekinci E, Jerums G. Serial measurements of cystatin C are more accurate than creatinine-based methods in detecting declining renal function in type 1 diabetes. Diabetes Care. 2008;31:971–973
  14. Stevens LA, Coresh J, Schmid CH, et al. Estimating GFR using serum cystatin C alone and in combination with serum creatinine: A pooled analysis of 3,418 individuals with CKD. Am J Kidney Dis. 2008;51:395–406
  15. Herget-Rosenthal S, Marggraf G, Husing J, et al. Early detection of acute renal failure by serum cystatin C. Kidney Int. 2004;66:1115–1122
  16. Abu-Omar Y, Mussa S, Naik MJ, MacCarthy N, Standing S, Taggart DP. Evaluation of cystatin C as a marker of renal injury following on-pump and off-pump coronary surgery. Eur J Cardiothorac Surg. 2005;27:893–898
  17. Biancofiore G, Pucci L, Cerutti E, et al. Cystatin C as a marker of renal function immediately after liver transplantation. Liver Transpl. 2006;12:285–291
  18. Villa P, Jimenez M, Soriano MC, Manzanares J, Casasnovas P. Serum cystatin C concentration as a marker of acute renal dysfunction in critically ill patients. Crit Care. 2005;9:R139–R143
  19. Zhu J, Yin R, Wu H, et al. Cystatin C as a reliable marker of renal function following heart valve replacement surgery with cardiopulmonary bypass. Clin Chim Acta. 2006;374:116–121
  20. Koyner JL, Bennett MR, Worcester EM, et al. Urinary cystatin C as an early biomarker of acute kidney injury following adult cardiothoracic surgery. Kidney Int. 2008;74:1059–1069
  21. Ahlstrom A, Tallgren M, Peltonen S, Pettila V. Evolution and predictive power of serum cystatin C in acute renal failure. Clin Nephrol. 2004;62:344–350
  22. Herget-Rosenthal S, Poppen D, Husing J, et al. Prognostic value of tubular proteinuria and enzymuria in nonoliguric acute tubular necrosis. Clin Chem. 2004;50:552–558
  23. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of disease classification system. Crit Care Med. 1985;13:818–829
  24. Bone RC. Sir Isaac Newton, sepsis, SIRS, and CARS. Crit Care Med. 1996;24:1125–1128
  25. Levey AS, Greene T, Kusek JW, Beck GJ. A simplified equation to predict glomerular filtration rate from serum creatinine. J Am Soc Nephrol. 2000;11:0828A;(abstr)
  26. Stone C, Koo C. Additive splines in statistics. Proc Stat Comput Section Am Stat Assoc. 1985;45–48
  27. Harrell FE. Design: Design Package. http://biostat.mc.vanderbilt.edu/s/DesignAccessed April 2, 2009
  28. R Development Core Team. R: A Language and Environment for Statistical Computing. http://www.R-project.orgAccessed April 2, 2009
  29. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics. 1988;44:837–845
  30. Chertow GM, Soroko SH, Paganini EP, et al. Mortality after acute renal failure: Models for prognostic stratification and risk adjustment. Kidney Int. 2006;70:1120–1126
  31. Vinge E, Lindergard B, Nilsson-Ehle P, Grubb A. Relationships among serum cystatin C, serum creatinine, lean tissue mass and glomerular filtration rate in healthy adults. Scand J Clin Lab Invest. 1999;59:587–592
  32. Finney H, Newman DJ, Price CP. Adult reference ranges for serum cystatin C, creatinine and predicted creatinine clearance. Ann Clin Biochem. 2000;37:49–59
  33. Bokenkamp A, Domanetzki M, Zinck R, Schumann G, Byrd D, Brodehl J. Cystatin C—A new marker of glomerular filtration rate in children independent of age and height. Pediatrics. 1998;101:875–881
  34. Groesbeck D, Kottgen A, Parekh R, et al. Age, gender, and race effects on cystatin C levels in US adolescents. Clin J Am Soc Nephrol. 2008;3:1777–1785
  35. Bokenkamp A, Laarman CA, Braam KI, et al. Effect of corticosteroid therapy on low-molecular weight protein markers of kidney function. Clin Chem. 2007;53:2219–2221
  36. Fricker M, Wiesli P, Brandle M, Schwegler B, Schmid C. Impact of thyroid dysfunction on serum cystatin C. Kidney Int. 2003;63:1944–1947
  37. Manetti L, Pardini E, Genovesi M, et al. Thyroid function differently affects serum cystatin C and creatinine concentrations. J Endocrinol Invest. 2005;28:346–349
  38. Jayagopal V, Keevil BG, Atkin SL, Jennings PE, Kilpatrick ES. Paradoxical changes in cystatin C and serum creatinine in patients with hypo- and hyperthyroidism. Clin Chem. 2003;49:680–681
  39. Yokoyama H, Inoue T, Node K. Effect of insulin-unstimulated diabetic therapy with miglitol on serum cystatin C level and its clinical significance. Diabetes Res Clin Pract. 2009;83:77–82
  40. Stevens LA, Schmid CH, Greene T, et al. Factors other than glomerular filtration rate affect serum cystatin C levels. Kidney Int. 2009;75:652–660
  41. Wagener G, Gubitosa G, Wang S, Borregaard N, Kim M, Lee HT. Urinary neutrophil gelatinase-associated lipocalin and acute kidney injury after cardiac surgery. Am J Kidney Dis. 2008;52:425–433
  42. Han WK, Waikar SS, Johnson A, et al. Urinary biomarkers in the early diagnosis of acute kidney injury. Kidney Int. 2008;73:863–869
  43. Liangos O, Perianayagam MC, Vaidya VS, et al. Urinary N-acetyl-beta-(D)-glucosaminidase activity and kidney injury molecule-1 level are associated with adverse outcomes in acute renal failure. J Am Soc Nephrol. 2007;18:904–912

 Originally published online as doi:10.1053/j.ajkd.2009.05.022 on August 7, 2009.

 Drs Perianayagam and Seabra contributed equally to this article.

 Because an author of this manuscript is an editor for AJKD, the peer-review and decision-making processes were handled entirely by an Associate Editor (Russell Chesney, MD, University of Tennessee Health Science Center) who served as Acting Editor-in-Chief. Details of the journal's procedures for potential editor conflicts are given in the Editorial Policies section of the AJKD website.

PII: S0272-6386(09)00926-3

doi:10.1053/j.ajkd.2009.05.022

American Journal of Kidney Diseases
Volume 54, Issue 6 , Pages 1025-1033, December 2009

Article Tools

* Image Usage & Resolution