Neutrophil gelatinase–associated lipocalin (NGAL) has been described in chronic heart failure (HF) as marker of tubular damage and renal dysfunction; however, less data are available in patients with acute HF. Because of high rate of acute kidney injury (AKI) development, we aimed to investigate the role of NGAL in predicting early AKI development; second, we compared NGAL with respect to cystatin C, B-type natriuretic peptide (BNP), renal function, and blood urea nitrogen (BUN) for outcome prediction. We measured admission serum NGAL, cystatin C, and BNP in 231 patients affected to acute HF; all patients were submitted to daily creatinine, estimated glomerular filtration rate, and measurement to identify inhospital AKI defined by Risk, Injury, Failure, Loss, End-Stage Kidney Disease and Acute Kidney Injury Network criteria. We also measured admission and discharge estimated glomerular filtration rate, creatinine, and BUN to evaluate their prognostic role during a 6-month follow-up period; 78 patients developed AKI during hospitalization. In these subjects, NGAL levels were significantly increased respect to patients without AKI (295 ± 228 vs 129 ± 108 ng/ml, p <0.001). A cutoff of 134 ng/ml has been related to AKI with good sensibility and specificity (85% and 80%, respectively; area under the curve 0.81, p <0.001). BNP was also mildly increased (1,000 ± 906 vs 746 ± 580 pg/ml, p = 0.03) but not cystatin C. Patients with chronic kidney disease demonstrated higher NGAL levels compared with subjects with preserved renal function (258 ± 249 and 120 ± 77 ng/ml, p <0.001). The receiver-operating characteristic curve analysis demonstrated that increased NGAL values were associated with increased mortality (cutoff 170 ng/ml, sensibility 60%, specificity 82%, accuracy 71%, area under the curve 0.77, p <0.001). The same significant correlation was also found for BUN at discharge (cutoff 100 mg/dl, sensibility 65%, specificity 85%, accuracy 71%, area under the curve 0.77, p <0.001). Multivariable Cox regression analysis showed that cutoff 170 ng/ml was related with adverse outcome (hazard ratio 1.77, confidence interval 1.24 to 2.83, p = 0.01). In conclusion, NGAL measurement is a sensible tool to predict AKI during hospitalization. Elevated NGAL levels appear to be related to BUN increase and post-discharge outcome. This suggests a prognostic role of tubular damage beyond renal dysfunction.
Neutrophil gelatinase–associated lipocalin (NGAL) has been recently proposed as an early marker to estimate the risk of acute kidney injury (AKI) and chronic kidney disease (CKD). Data from several sources demonstrated that approximately 20% to 40% of patients with acute heart failure (HF) develop AKI, defined on the basis of Risk, Injury, Failure, Loss, End-Stage Kidney Disease and Acute Kidney Injury Network criteria. Because of the high rate of AKI development and its prognostic impact in these patients, we aimed to investigate: (1) the role of early NGAL in predicting AKI in patients with acute HF; (2) plasma admission NGAL, B-type natriuretic peptide (BNP), cystatin C, blood urea nitrogen (BUN) at admission and discharge for AKI prediction during hospitalization; and (3) the prognostic role of admission NGAL compared with cystatin C, BNP, BUN, creatinine, and estimated glomerular filtration rate (eGFR) during the 6-month post-discharge period.
Methods
This is a prospective, randomized, single-center pilot study evaluating 231 patients with primary diagnosis of acute HF. Patients were enrolled consecutively from the Department of Internal Medicine, Cardiology Section Center (Siena, Italy) from April 2010 to September 2013. This is a substudy including patients with primary diagnosis of acute HF admitted for continuous or intermittent loop diuretic therapy (Diuretic1 NCT01441245). Patients were eligible if they were admitted with a primary diagnosis of acute HF, could be randomized within 24 hours from hospital presentation, and had evidence of volume overload (pulmonary congestion on chest x-ray or BNP greater than the upper limit of normal for age). Patients were evaluated to determine HF status according to the New York Heart Association. All recruited patients underwent echocardiographic examination (Hewlett-Packard 5500 Sonos; Philips, Andover, Massachusetts). Changes in BUN and creatinine were monitored during their hospital stay for 5 days or until hospital discharge. eGFR was calculated using the Modification of Diet in Renal Disease formula. AKI was defined as an increase in serum creatinine >0.3 mg/dl with respect to baseline values in accordance with previous studies following Acute Kidney Injury Network criteria.
CKD was defined by basal cutoff >1.4 mg/dl for creatinine and eGFR <60 ml/min/1.73 m 2 . Patients were followed up to 6 months starting from the first day of hospital admission by ambulatory visit every 2 months or telephone call. Six patients were lost during follow-up. There were 2 co-primary end points: (1) to evaluate the role of NGAL, BNP, and cystatin C for early AKI prediction during hospitalization period and (2) to identify the potential prognostic significance of biomarkers during 6-month follow-up period; secondary end point was the relation evaluation among these biomarkers, baseline and discharge BUN, creatinine, and eGFR. Patients were categorized on the basis of renal function in one of the following diagnostic categories: (1) preserved renal function, patients with eGFR ≥60 ml/min/1.73 m 2 , and creatinine ≤1.2 mg/L that did not meet the criteria of any of the other categories; (2) stable CKD, patients with a sustained elevation of creatinine level indicative of a reduced eGFR of <60 ml/min/1.73 m 2 and creatinine ≥1.4 mg/dl that was calculated before hospitalization ; and (3) AKI, defined on the basis of Acute Kidney Injury Network criteria during the hospitalization period. In particular, a new-onset 1.5-fold increase in creatinine level or a 25% decrease in eGFR from baseline were included. Patients were excluded if they had received >2 intravenous doses of furosemide or any continuous infusion of furosemide before randomization, if they had end-stage renal disease or the need for renal replacement therapy (dialysis or ultrafiltration), isolated diastolic dysfunction, and recent myocardial infarction. Patients with systolic blood pressure of <90 mm Hg or a baseline serum creatinine level that was >5.0 mg/dl were also excluded. Moreover, patients with secondary HF because of systemic, neoplastic, or inflammatory diseases were also excluded. Once written informed consent had been obtained, baseline demographic data were recorded, and blood samples were obtained for measurements of NGAL, cystatin C, and BNP. These measurements were repeated at admission and soon before discharge. Samples for NGAL and BNP assays were obtained at admission within the first 24 hours from hospitalization. The Triage NGAL test (Alere Inc., San Diego, California) is an immunoassay in a single-use plastic cartridge that contains a fluorescently labeled monoclonal antibody against NGAL labeled with a fluorescent dye and NGAL. The test procedure involves the addition of several drops of whole blood or plasma to the sample port on the test device. After addition of the sample, the cells are automatically separated from the plasma by a filter. The sample reacts with fluorescent antibody, conjugates within the reaction chamber, and flows down the device detection lane by capillary action. The fluorescent antibody conjugates are captured on discrete solid-phase zones resulting in binding immunoassays that are specific for NGAL or the control antigens. BNP testing was also performed on the Triage platform using a standard commercially available assay (Triage Assay; Alere Inc.). Cystatin C was measured with a particle-enhanced immune nephelometric method (Dade Behring, Margburg, Germany); the reference intervals for normal value are 0.53 to 0.95 mg/L without differences between genders.
A complete blood count with hemoglobin, hematocrit, red blood cell count, serum creatinine, BUN, sodium, and potassium were performed at baseline, every day during hospital time period, and before discharge. This trial was approved by the Local Investigational Human Review Board, and all patients gave their signed informed consent.
Continuous variables are presented as mean ± SD and were compared using the t tests for paired data. Differences between groups for the categorical variables were analyzed using the chi-square tests. We repeated the tests in subgroups to discriminate patients with CKD with respect to those with preserved renal function and patients with AKI with respect to non-AKI subjects. The correlation analysis was performed using spearman’s rho coefficient. The diagnostic utility of BNP, NGAL, BUN, cystatin C, and other laboratory values and risk factors in differentiating patients with CKD, AKI, and adverse events were studied using receiver-operating characteristic curve. The results are expressed in terms of the area under the curve and 95% confidence interval for this area. Sensitivity, specificity, and accuracy were computed for significant variables. The impact of significant biomarkers on adverse event was evaluated with Cox multivariable regression models adjusted for demographic variables and risk factors (hypertension, diabetes, dyslipidemia, and smoke). We used Kaplan-Meier curves to display survival using a selection of possible cut-off points for significant markers. p Values <0.05 were considered significant. Statistical analyses were performed using SPSS, version 20.0 (SPSS, Chicago, Illinois).
Results
The initial number of recruited patients was 231; however, 6 were excluded because of cardiac death during hospitalization, 16 patients for incomplete laboratory data, and 6 patients were lost during follow-up. Of the remaining 203 patients at admission, 107 had CKD (with eGFR <60 ml/min/1.73 m 2 ) and 96 had preserved renal function.
Seventy-eight patients developed AKI during hospitalization period, defined as an increase of creatinine ≥0.3 mg/dl with respect to baseline (or decrease of eGFR ≤25% at baseline). In this subgroup, NGAL levels were significantly increased with respect to patients without AKI (p <0.001). Patients with AKI had higher BUN levels at discharge with respect to those without AKI (p <0.001). The same trend was found for BNP (p = 0.03). Cystatin C was not significantly increased in this subgroup with respect to patients without AKI ( Table 1 ).
| Variable | All patients (n=203) | ||
|---|---|---|---|
| Acute Kidney Injury (n=78) | No Acute Kidney Injury (n=125) | p-value | |
| Age (years) | 78±9 | 80±8 | 0.242 |
| Male | 35 | 48 | 0.361 |
| Female | 43 | 77 | |
| Ejection Fraction | 35%±10 | 40%±9 | <0.001 |
| Hypertension | 45 (58%) | 42 (34%) | <0.001 |
| Diabetes mellitus | 45 (58%) | 72 (58%) | 0.989 |
| Dyslipidemia | 32 (41%) | 51 (41%) | 0.974 |
| Smoker | 40 (51%) | 90 (72%) | 0.004 |
| Increased Blood Urea Nitrogen at discharge (upper 20% than baseline) | 71 (92%) | 7 (6%) | <0.001 |
| Creatinine at discharge (mg/dL) | 2.3±0.7 | 1.5±0.7 | <0.001 |
| Estimated Glomerular Filtration Rate at discharge (mL/min/1.73 m 2 ) | 29±12 | 53±21 | <0.001 |
| Neutrophil Gelatinase Associated Lipocalin (ng/mL) | 295±228 | 129±108 | <0.001 |
| Blood Urea Nitrogen at discharge (mg/dL) | 155±56 | 70±37 | <0.001 |
| B-type Natriuretic Peptide (pg/mL) | 1000±906 | 746±580 | 0.03 |
| Serum Cystatin C (mg/L) | 1.2±0.5 | 1.1±0.5 | 0.204 |
The receiver-operating characteristic curve analysis for predicting AKI confirmed that an NGAL cutoff >134 ng/ml had high sensibility 85%, specificity 80%, and accuracy 82% (area under the curve 0.81 [0.76 to 0.88], p <0.001); similarly, a cutoff of 110 mg/dl for BUN at discharge had good sensibility 70%, specificity 90%, and accuracy 83% (area under the curve 0.91 [0.87 to 0.95], p <0.001). BNP and cystatin C were not statistically significant in predicting AKI ( Figure 1 ). The percentage of patients who had BUN at discharge increased >20% compared with that the admission value was significantly higher in patients with AKI (92% vs 6%, p <0.001). The receiver-operating characteristic curve analysis for NGAL, BNP, and cystatin C in predicting a BUN increase >20% during hospitalization demonstrated that NGAL values were linked with this condition (area under the curve 0.79 [0.73 to 0.85], p <0.001). BNP and cystatin C did not predict worsening BUN during hospitalization.
Of 78 patients who developed AKI during hospitalization, 45 patients had CKD at admission, defined as admission values of creatinine ≥1.4 mg/dl with respect to baseline (or eGFR ≤60% at baseline). Clinical and laboratory characteristics of patients with both CKD and AKI with respect to those with AKI without baseline CKD are described in Table 2 . Patients with CKD had increased levels of both tubular and glomerular damage compared with those with preserved renal function (258 ± 249 vs 120 ± 77 ng/ml, p <0.001 for NGAL; 1.5 ± 0.5 mg/L vs 0.7 ± 0.2, p <0.001 for cystatin C). BNP levels showed only a mild increase in CKD group (946 ± 909 vs 729 ± 676 pg/ml, p = 0.05).
| Variable | Patients with Acute Kidney Injury (n=78) | ||
|---|---|---|---|
| No Chronic Kidney Disease (n=33) | Chronic Kidney Disease (n=45) | p-value | |
| Age (years) | 75±8 | 80±9 | 0.01 |
| Male | 19 | 16 | 0.05 |
| Female | 14 | 29 | |
| Hypertension | 15 (48%) | 28 (64%) | 0.141 |
| Diabetes mellitus | 21 (66%) | 28 (64%) | 0.896 |
| Dyslipidemia | 15 (48%) | 28 (64%) | 0.141 |
| Smoker | 14 (45%) | 24 (55%) | 0.340 |
| Basal Creatinine (mg/dL) | 1.2±0.2 | 2.1±0.7 | 0.001 |
| Basal estimated Glomerular Filtration Rate (mL/min/1.73 m 2 ) | 61±14 | 29±10 | 0.001 |
| Basal Blood Urea Nitrogen (mg/dL) | 71 ± 32 | 127 ± 46 | 0.001 |
| Neutrophil Gelatinase Associated Lipocalin at admission (ng/mL) | 175±84 | 383±301 | 0.001 |
| Cystatin C at admission (ng/mL) | 0.7±0.2 | 1.6 ±0.5 | 0.001 |
| Creatinine at discharge (mg/dL) | 1.8±0.4 | 2.6±0.7 | 0.001 |
| Estimated Glomerular Filtration Rate at discharge (mL/min/1.73 m 2 ) | 38±10 | 22±7 | 0.001 |
| Blood Urea Nitrogen at discharge (mg/dL) | 120 ± 42 | 180 ± 50 | 0.001 |
| B-type Natriuretic Peptide (pg/mL) | 882±854 | 1086±1130 | 0.387 |
| Ejection Fraction | 35%±11 | 34%±10 | 0.913 |
In all patients, we found positive significant correlations among NGAL and creatinine at baseline and at discharge (r = 0.40, p <0.001; r = 0.62, p <0.001, respectively) and NGAL and BUN at admission and at discharge (r = 0.54, p <0.001; r = 0.65, p <0.001, respectively). We found a mild significant correlation among NGAL and cystatin C (r = 0.41, p <0.001). We revealed an inverse significant correlations among NGAL and GFR at baseline and at discharge (r = −0.40, p <0.001; r = −0.64, p <0.001, respectively). In line with previous correlation, we found important significant correlations among cystatin C and basal renal parameters: cystatin C and creatinine (r = 0.92, p <0.001), cystatin C and GFR (r = −0.85, p <0.001), and cystatin C and BUN (r = 0.67; p <0.001).
The receiver-operating characteristic curve analysis for NGAL, BNP, cystatin C, and renal parameters at baseline and at discharge for predicting adverse events at 6-month follow-up demonstrated that increased NGAL values were linked with increased mortality (cutoff 170 ng/ml, sensibility 60%, specificity 82%, accuracy 71%, area under the curve 0.77, p <0.001). The same significant correlation was also found for BUN at discharge (cutoff 100 mg/dl, sensibility 65%, specificity 85%, accuracy 71%, area under the curve 0.77, p <0.001). BNP, basal creatinine, creatinine at discharge, BUN at admission, and cystatin C also predicted adverse events but were less significant (BNP: area under the curve 0.59, p = 0.01; admission BUN: area under the curve 0.68, p = 0.001; cystatin C: area under the curve 0.67, p = 0.001; basal creatinine: area under the curve 0.64, p = 0.001; creatinine at discharge: area under the curve 0.76, p <0.001). As previous markers, the decrease in eGFR also predict death and rehospitalization (eGFR at admission: area under the curve 0.38, p = 0.003; eGFR at discharge: area under the curve 0.26, p <0.001) ( Figure 2 ). Univariate and multivariable Cox regression analyses for all-cause mortality, cardiac death, and rehospitalization adjusted for age, gender, and traditional cardiovascular risk factors showed that NGAL values >170 ng/ml were associated with the combined end point. Similarly, BUN at discharge >100 mg/dl and creatinine at discharge ≥2 mg/dl predicted poor outcome. BNP was also a predictor for adverse outcome. In contrast, cystatin C was not predictor for adverse outcome ( Table 3 ).
