Urinary N-Acetyl-β-(D)-Glucosaminidase Activity and Kidney Injury Molecule-1 Level Are Associated with Adverse Outcomes in Acute Renal Failure

Study Design

This was a prospective cohort study of hospitalized patients with ARF, which was conducted between November 2003 and January 2006 at two tertiary care hospitals located in Boston, MA. All consecutive hospitalized adult patients who had ARF and for whom nephrology consultation was requested were eligible for enrollment. ARF was defined as an incremental increase in serum creatinine by 0.5, 1.0, or 1.5 mg/dl from a baseline level of ≤1.9, 2.0 to 4.9, and ≥5.0 mg/dl, respectively, as described previously (1). These criteria are graded to require a larger increase in serum creatinine according to the baseline value. Exclusion criteria were age <18 yr, current pregnancy, chronic dialysis therapy, receipt of an organ transplant within the previous year, and presence of acute obstructive uropathy. Patients who were receiving comfort measures only and those for whom therapeutic interventions were considered futile by the caregivers also were excluded. Written informed consent was obtained from all participants or their next of kin. The institutional review board of each participating center approved the study protocol.

Data Collection

Medical records of study participants were reviewed prospectively to retrieve hospitalization data, including baseline demographic characteristics; coexisting conditions; and renal variables including serial serum creatinine values, presence of oliguria (as defined by urine output <400 ml/d), fractional excretion of sodium, and urinary sediment findings. Contributing causes of ARF were determined according to the notes of the nephrology consult service. A blinded investigator (O.L. or R.W.M.) who was unaware of the patient’s clinical characteristics examined the urinary sediment. At the time of enrollment, two severity-of-illness scores were calculated: The Acute Physiology and Chronic Health Evaluation II (APACHE II) score (11) and the Multiple Organ Failure (MOF) score (12). The presence of sepsis was ascertained using the systemic inflammatory response syndrome criteria (13). Outcomes of interest were dialysis requirement, hospital death, and the composite outcome of dialysis requirement or hospital death.

Processing and Storage of Urinary Samples

Fresh urinary samples were obtained at the time of enrollment and were centrifuged immediately to remove insoluble elements after routine test-strip urinalysis. The urine sediment was examined by light microscopy and evaluated for the presence or absence of granular casts. The supernatant was treated with a protease inhibitor cocktail tablet (Complete, Mini; Roche Diagnostics, Mannheim, Germany) and stored at −80°C until assayed. This cocktail tablet inhibits a broad spectrum of serine, cysteine, and metalloproteases as well as calpains that are present in mammalian tissues.

Measurement of Urinary NAG Activity

NAG activity was measured in the urine by a colorimetric assay (Boehringer Mannheim, Mannheim, Germany). In brief, this method uses the substrate 3-cresolsulfonphthaleinyl-N-acetyl-β-D-glucosaminidine-sodium, which is hydrolyzed by NAG when present in the urinary sample. This reaction releases 3-cresolsulfonphthalein-sodium, which is measured by spectrophotometry. According to the manufacturer’s instructions, 1 ml of the substrate solution was incubated for 5 min at 37°C. A 50-μl aliquot of the urinary sample then was added to the substrate solution, mixed, and incubated for 15 min at 37°C. After incubation, 2 ml of the stop reagent solution that contained sodium carbonate was added to the sample mixture and allowed to stand for 10 min at room temperature. The absorbance then was measured by a spectrophotometer (Beckman Coulter, Fullerton, CA) set at 580 nm. A single measurement was performed per sample. The inter- and intra-assay coefficients of variation were 4.3 and 6.0%, respectively. Results were normalized to urinary creatinine values and expressed in mU/mg creatinine.

Measurement of Urinary KIM-1 Level

Urinary KIM-1 measurements were performed using microsphere-based Luminex xMAP technology with polyclonal antibodies raised against the human KIM-1 ectodomain. This technique is an adaptation of the sandwich ELISA assay, as described previously (7,14,15). For measurements, 30 μl of urine samples were analyzed in duplicate.

The lowest limit of detection for this assay is 0.125 ng/ml. The inter- and intra-assay variability was <10%. The urinary KIM-1 level was expressed in absolute terms and also normalized to the urinary creatinine concentration. Three blinded investigators (M.C.P, W.K.H., and V.S.V.) who were unaware of the patients’ clinical characteristics performed all of the urinary biomarker measurements.

Statistical Analyses

Comparisons between urinary NAG and KIM-1 quartiles were made by the Kruskal-Wallis or ANOVA tests, as appropriate, for continuous variables and by χ² test for categorical variables. Logistic regression analysis was used to examine the association of urinary NAG activity and KIM-1 level with the composite outcome of dialysis requirement or hospital death. This composite end point was chosen because it takes into consideration hospital survival bias for dialysis requirement. The models were adjusted for either the APACHE II score or the MOF score. These baseline covariates were chosen because the APACHE II score represents a composite illness-severity score that takes into consideration several demographic, physiologic, and laboratory variables and the presence of sepsis, whereas the MOF score represents a composite organ failure score that is characterized by physiologic and laboratory criteria that indicate severe vital organ dysfunction. In a separate model, the analysis was adjusted for the four covariates: Sepsis, oliguria, cirrhosis, and mechanical ventilation. These four variables were selected on the basis of the strength of their univariate associations with the composite outcome. In this model, a backward selection approach was adopted with a P < 0.2 as the criterion to retain covariates in the final model.

Restricted cubic splines functions with four knots (i.e., piece-wise cubic polynomials) were used to explore the functional form of the urinary NAG activity and KIM-1 level and graphically display their univariate relationship with the composite outcome of dialysis requirement or hospital death (16). Ninety-five percent confidence intervals for the area under the curves were calculated.

The performance characteristics of urinary NAG activity and KIM-1 level in predicting the composite outcome were described using the area under a receiver operator characteristic (ROC) curve and compared with the performance of more traditional clinical severity indices of kidney injury, including serum creatinine, urine output (using the negative value because of its inverse association with adverse outcomes), and APACHE II score. The serum creatinine at enrollment was chosen to serve as a standard for comparison to test formally whether various models (i.e., differences in the areas under the models’ ROC curves) are statistically more or less discriminating, using the nonparametric method of DeLong et al. (17).

Results are expressed as means (with SD) or percentages. Figures are graphically displayed as box and whisker plots. Differences were considered statistically significant at P < 0.05. All statistical procedures were performed using SAS version 9.1 (SAS Institute, Cary, NC), except for Figures 1 and 2, which were generated using the SPSS version 12.0 (Chicago, IL).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Urinary N-acetyl-β-(D)-glucosaminidase (NAG) activity (A) and kidney injury molecule-1 (KIM-1) level (B) stratified by the Acute Physiology and Chronic Health Evaluation (APACHE) II score quartiles. P < 0.001 for NAG activity and P = 0.052 for KIM-1 level, by the Kruskal-Wallis test. The box and whisker plots show the 10th, 25th, 50th (median), 75th, and 90th percentile values.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Urinary NAG activity (A) and KIM-1 level (B) stratified by the Multiple Organ Failure (MOF) score. P = 0.003 for NAG activity and P = 0.001 for KIM-1 level, by the Kruskal-Wallis test. The box and whisker plots show the 10th, 25th, 50th (median), 75th, and 90th percentile values.

Overall Characteristics of the Cohort

A total of 201 patients were enrolled. Mean age was 65 yr, 91% were white, and 45% were women. Mean APACHE II score was 16, 73% were in the intensive care unit, 43% had sepsis, and 24% required mechanical ventilation. ARF often was the result of multifactorial causes, and 18% of patients were oliguric. Thirty-nine percent required dialysis, and hospital mortality was 24%. Among the survivors, 13% were dialysis dependent at the time of hospital discharge.

Characteristics of the Cohort Stratified by Urinary NAG Activity and KIM-1 Level

Table 1 displays the characteristics of the cohort according to the urinary NAG activity quartiles. At baseline, patients in the highest urinary NAG quartile had a significantly higher prevalence of sepsis, higher APACHE II score, higher MOF score, a higher likelihood of mechanical ventilation, higher prevalence of oliguria, and fractional excretion of sodium >1% as compared with those with lower urinary NAG. Urinary NAG activity increased in tandem with APACHE II score quartiles (P < 0.001; Figure 1A) and MOF scores (P < 0.003; Figure 2A). Dialysis requirement, hospital death, and the composite outcome of dialysis requirement or hospital death also were significantly higher in the urinary NAG fourth quartile patient group as compared with the other quartile groups (Table 1).

Table 2 displays the characteristics of the cohort according to the urinary KIM-1 level quartiles. At baseline, the urinary KIM-1 fourth quartile patient group had significantly higher MOF scores and a higher prevalence of oliguria as compared with the other quartile groups. Urinary KIM-1 level also increased in tandem with APACHE II score quartiles (P = 0.052; Figure 1B) and MOF scores (P = 0.001; Figure 2B). The composite outcome of dialysis requirement or hospital death also was significantly higher in the urinary KIM-1 fourth quartile patient group as compared with the other quartile groups.

Association of Urinary NAG Activity and KIM-1 Level with the Composite Outcome of Dialysis Requirement or Hospital Death

Results of the logistic regression analyses are shown in Table 3. Compared with the NAG first quartile group, the second, third, and fourth quartile groups were associated with 3.0-, 3.7-, and 9.1-fold higher odds for dialysis requirement or hospital death on univariate analysis, respectively. This graded association persisted after adjustment for the APACHE II score; the MOF score; or the combined covariates cirrhosis, sepsis, oliguria, and mechanical ventilation.

Compared with the urinary KIM-1 first quartile group, the second, third, and fourth quartile groups were associated with 1.4-, 1.4-, and 3.2-fold higher odds for dialysis requirement or hospital death on univariate analysis, respectively. This association was attenuated after adjustment for the APACHE II score; the MOF score; or the combined covariates cirrhosis, sepsis, oliguria, and mechanical ventilation.

The results of the restricted cubic splines functions that explored the univariate relationship between urinary NAG activity and KIM-1 level, respectively, with the composite outcome of dialysis requirement or hospital death are displayed in Figure 3. Increasing levels of both NAG and KIM-1 were associated with a heightened probability for the composite outcome but differed in their curve characteristics. Increasing urinary NAG activity at levels <50 mU/mg was associated with steeper increases in probability for the outcome as compared with levels >50 mU/mg, where the direct relationship persisted but in an attenuated manner (Figure 3A). In comparison, KIM-1 levels were associated with the outcome in the pattern of an S-shaped curve with increasing slope below 10 ng/mg, then decreasing slope between 10 and 15 ng/mg, and asymptotic flattening of the slope above 15 ng/mg (Figure 3B). Accordingly, urinary KIM-1 levels >15 ng/mg were not associated with large increases in the predicted probability of the outcome.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Univariate relationship of urinary NAG activity (A) and KIM-1 level (B) with the predicted probability for the composite adverse clinical outcome of dialysis requirement or hospital death, using restricted cubic splines functions (16). The solid line represents the point estimate, and the hatched lines represent the 95% confidence interval of the predicted probability. Markings at the top of the graph indicate individual observation points.

Prognostic Performance Characteristics of Urinary Biomarkers Compared with Selected Clinical Variables

Table 4 displays the area under the ROC curve (AUC) of selected clinical and urinary predictor variables for the composite outcome of dialysis requirement or hospital death. Urinary NAG activity performed better than serum creatinine or urine output, two traditional clinical parameters. The APACHE II score, a composite score of 15 clinical variables, provided an AUC of 0.78, which was improved marginally by combining this score with urinary NAG activity (AUC 0.79) or KIM-1 level (AUC 0.80). The combination of urinary KIM-1 level or NAG activity with the four variables hepatic cirrhosis, sepsis, oliguria, and mechanical ventilation performed as well as the 15-variable APACHE II score for prediction of the composite outcome (AUC 0.78). The combination of urinary markers with APACHE II score or the four-variable model yielded areas under the ROC curves that were statistically significantly more discriminating than serum creatinine at enrollment alone (Table 4).