Sensitivity and specificity of a single emergency department measurement of urinary neutrophil gelatinase-associated lipocalin for diagnosing acute kidney injury

Abstract

Background: A single serum creatinine measurement cannot distinguish acute kidney injury from chronic kidney disease or prerenal azotemia.

Objective: To test the sensitivity and specificity of a single measurement of urinary neutrophil gelatinase-associated lipocalin (NGAL) and other urinary proteins to detect acute kidney injury in a spectrum of patients.

Design: Prospective cohort study.

Setting: Emergency department of Columbia University Medical Center, New York, New York.

Participants: 635 patients admitted to the hospital with acute kidney injury, prerenal azotemia, chronic kidney disease, or normal kidney function.

Measurements: Diagnosis of acute kidney injury was based on the RIFLE (risk, injury, failure, loss, and end-stage) criteria and assigned by researchers who were blinded to experimental measurements. Urinary NGAL was measured by immunoblot, N-acetyl-beta-d-glucosaminidase (NAG) by enzyme measurement, alpha1-microglobulin and alpha(1)-acid glycoprotein by immunonephelometry, and serum creatinine by Jaffe kinetic reaction. Experimental measurements were not available to treating physicians.

Results: Patients with acute kidney injury had a significantly elevated mean urinary NGAL level compared with the other kidney function groups (416 microg/g creatinine [SD, 387]; P = 0.001). At a cutoff value of 130 microg/g creatinine, sensitivity and specificity of NGAL for detecting acute injury were 0.900 (95% CI, 0.73 to 0.98) and 0.995 (CI, 0.990 to 1.00), respectively, and positive and negative likelihood ratios were 181.5 (CI, 58.33 to 564.71) and 0.10 (CI, 0.03 to 0.29); these values were superior to those for NAG, alpha1-microglobulin, alpha1-acid glycoprotein, fractional excretion of sodium, and serum creatinine. In multiple logistic regression, urinary NGAL level was highly predictive of clinical outcomes, including nephrology consultation, dialysis, and admission to the intensive care unit (odds ratio, 24.71 [CI, 7.69 to 79.42]).

Limitations: All patients came from a single center. Few kidney biopsies were performed.

Conclusion: A single measurement of urinary NGAL helps to distinguish acute injury from normal function, prerenal azotemia, and chronic kidney disease and predicts poor inpatient outcomes.

Conflict of interest statement

Potential Financial Conflicts of Interest: Honoraria: P. Devarajan (Biosite, Abbott). Grants pending: P. Devarajan (Abbott, Biosite). Patents pending: T.L. Nickolas (NGAL for diagnosis of chronic renal failure), K. Mori (NGAL for diagnosis of chronic renal failure), P. Devarajan (NGAL for diagnosis of acute renal failure and NGAL for diagnosis of chronic renal failure), J. Barasch (NGAL for diagnosis of acute renal failure and NGAL for diagnosis of chronic renal failure). Cincinnati Children’s Hospital Medical Center and Columbia University have received licensing fees from Biosite and Abbott Diagnostics for technology to use NGAL as a biomarker of acute renal failure.

Figures

Figure 1

Study flow diagram.

Figure 2. Box plots of urinary neutrophil gelatinase–associated lipocalin (NGAL) and serum creatinine levels, by diagnostic group

Top. Patients with acute kidney injury had markedly elevated mean urinary NGAL levels compared with patients who had other forms of kidney dysfunction. Little overlap was present, except in 3 patients with chronic kidney disease and 1 patient with prerenal azotemia who had high urinary NGAL levels. Bottom. Patients with acute kidney injury had significantly elevated mean serum creatinine levels compared with patients who had other forms of kidney dysfunction, but values overlapped among the different categories of kidney function. To convert mg/dL to μmol/L, multiply by 88.402.

Figure 3. Kidney injury biomarkers versus clinical outcome

Measurements are normalized per gram of creatinine. Bars show the proportion of patients with biomarker levels above the selected cutoffs and clinical outcome. The total number of patients whose levels were above the cutoff was 66 for creatinine, 31 for neutrophil gelatinase–associated lipocalin (NGAL), 133 for N-acetyl-β-D-glucosaminidase (NAG), 143 for α1-microglobulin (A1M), and 294 for α1-acid glycoprotein (AAG). *P < 0.001 compared with NGAL.