Pediatric Nephrology

, Volume 26, Issue 1, pp 29–40

Biomarkers of acute kidney injury in children: discovery, evaluation, and clinical application


  • Zubaida Al-Ismaili
    • Department of Pediatrics, Division of Nephrology, Montreal Children’s HospitalMcGill University Health Centre
  • Ana Palijan
    • Department of Pediatrics, Division of Nephrology, Montreal Children’s HospitalMcGill University Health Centre
    • Department of Pediatrics, Division of Nephrology, Montreal Children’s HospitalMcGill University Health Centre

DOI: 10.1007/s00467-010-1576-0

Cite this article as:
Al-Ismaili, Z., Palijan, A. & Zappitelli, M. Pediatr Nephrol (2011) 26: 29. doi:10.1007/s00467-010-1576-0


Acute kidney injury (AKI) in children is associated with increased mortality and prolonged length of hospital stay and may also be associated with long-term chronic kidney disease development. Despite encouraging results on AKI treatment in animal studies, no specific treatment has yet been successful in humans. One of the important factors contributing to this problem is the lack of an early AKI diagnostic test. Serum creatinine, the current main diagnostic test for AKI, rises late in AKI pathophysiology and is an inaccurate marker of acute changes in glomerular filtration rate. Therefore, new biomarkers of AKI are needed. With great advancements in genomics, proteomics, and metabolomics, new AKI biomarkers, mainly consisting of urinary proteins that appear in response to renal tubular cell injury, have been, and continue to be, discovered. These new biomarkers offer promise for early AKI diagnosis and for the depiction of severity of renal injury occurring with AKI. This review provides a summary of what a biomarker is, why we need new biomarkers of AKI, and how biomarkers are discovered and should be evaluated. The review also provides a summary of selected AKI biomarkers that have been studied in children.


Acute renal failureChildrenDiagnostic testChronic kidney diseaseGenomicsProteomics


Acute kidney injury (AKI) has long been known to be an important clinical problem in hospitalized children. Thanks to the recent standardization of AKI definitions, our understanding of the incidence and risk on poor outcome of AKI has increased in the last 5 to 7 years. Despite encouraging therapeutic targets identified from animal studies, there are currently no specific treatments for AKI. The current clinical AKI diagnostic test is serum creatinine (SCr). Many studies have shown SCr to be a poor diagnostic test and an inaccurate biomarker for AKI, which has greatly hindered treatment trials.

The area of biomarkers of pediatric AKI is exciting for several reasons. First, it is a great example of bringing the bench to the bedside, wherein animal studies have taught what genes and proteins are involved in tubular cell injury and repair, leading to the discovery of biomarkers and eventual application to humans. Secondly, biomarker research will provide the tools needed to adequately study new AKI treatments. Finally, AKI biomarker research is not only relevant to children, but to hospital patients of all ages. This review will provide the reader with a brief summary of what a biomarker is and of biomarker discovery, followed by a summary of selected AKI biomarkers studied in children.

What is a biomarker?

A biomarker has been defined as “a characteristic that is objectively measured and evaluated as an indicator of normal biological process, pathogenic process, or pharmacologic response to a therapeutic intervention” [1]. Some commonly used “biomarkers” include: height, a biomarker of growth (biologic process); proteinuria, a biomarker of disease severity in IgA nephropathy (disease progression); urine dipstick for nitrites in urinary tract infection (diagnostic); anti-glomerular basement membrane antibodies in patients with Goodpasture’s syndrome (therapeutic response).

Potentially, biomarkers may be used as surrogate end-points, or “a biomarker intended to substitute for a clinical end-point aiming to predict clinical benefit or harm (or lack of benefit or harm) on the basis of epidemiological, therapeutic, pathophysiological, or other scientific evidence” [1], in longitudinal studies or clinical trials. In pediatric research, where hard end-points (death or dialysis requirement) are rare or take years to achieve, the use of biomarkers as surrogate end-points is incredibly appealing. However, great caution must be used and extensive validation of a biomarker must be performed. The latter is done by displaying a strong association with accepted clinical end-points prior to accepting a biomarker as a surrogate end-point.

Why do we need novel biomarkers for acute kidney injury?

Some current biomarkers of AKI include the presence of SCr rise (the main biomarker), urinary casts, or clinical markers of renal dysfunction (fluid overload/oliguria). Discovering and validating new biomarkers of disease is a long, exhaustive, and expensive endeavor. Therefore, the disease of interest must be important (estimated by its association with poor outcomes, its incidence or prevalence) and there should be a substantial deficit in our ability to either diagnose the disease or prognosticate outcome. AKI satisfies these criteria.

Two very similar standardized definitions of AKI have recently been derived and described for use in children: the pediatric Risk, Injury, Failure, Loss, End-Stage Kidney Disease criteria [pRIFLE] or the Acute Kidney Injury Network Staging [AKIN] (Table 1) [2, 3]. AKI in hospitalized children using both definitions is common [2, 4, 5] and is associated with poor short-term outcomes like mortality and length of stay [2, 5, 6]. Early studies suggest that children with AKI may be at risk for long-term chronic kidney disease [7].
Table 1

Two recent definitions of acute kidney injury: pediatric Risk, Injury, Failure, Loss, End-Stage Kidney Disease criteria and the Acute Kidney Injury Network Staging [2, 3]

Acute Kidney Injury Network (AKIN) classification of AKI


Serum creatinine (SCr) criteria

Urine output (UO) criteria

Stage 1

≥0.3 mg/dl (26.5 μmol/l) rise or rise to 1.5–1.99 X baseline

UO <0.5 ml/kg/h for 6 h

Stage 2

Rise to ≥ 2–2.99 X baseline

UO <0.5 ml/kg/h for 12 h

Stage 3

Rise to ≥ 3 X baseline or ≥ 4 mg/dl (354 μmol/l) rise with an acute rise of at least 0.5 mg/dl (44 μmol/l)

UO < 0.3 ml/kg/h for 24 h or anuria for 12 h

Modified pediatric Risk, Injury, Failure, Loss, End-stage kidney disease (pRIFLE) criteria


Estimated creatinine clearances (eCCl)

Urine output (UO) criteria


eCCl decreased by 25%

UO < 0.5 ml/kg/h for 8 h


eCCl decreased by 50%

UO < 0.5 ml/kg/h for 16 h


eCCl decreased by 75% or eCCl <35 ml/min/1.73 m2

UO < 0.3 ml/kg/h for 24 h or anuria for 12 h


Persistent failure >4 weeks


 End-stage kidney diseasea

Persistent failure >3 months


a The pRIFLE stages “loss” and “end-stage kidney disease” are not AKI stages per se; they describe chronic outcomes of AKI

SCr is an inadequate biomarker of AKI due to its late rise in AKI pathophysiology. It is a functional marker of glomerular filtration rate, only rising once substantial renal injury has already occurred and a large proportion of renal filtration capacity has been lost. This is a major problem and may explain why clinical trials for AKI have failed in which SCr rise was used for AKI diagnosis. Animal studies have shown that to successfully treat AKI, early intervention is essential [8]. SCr concentrations are influenced by several non-renal factors like muscle mass, diet, medications and tubular secretion, lending to inaccuracies in depicting SCr rise with AKI and making SCr rise as the sole diagnostic AKI test inadequate. New AKI biomarkers of renal tubular injury, which are involved in early pathophysiologic events of AKI or in molecular events of renal tubular damage or repair, can serve to provide early AKI diagnosis and potentially, to provide information on the location and cause of AKI (i.e. affected renal tubule segment; nephrotoxic versus ischemic AKI). A good AKI biomarker should be non-invasive, easily obtainable and measurable using standardized assays with fast results and incur reasonable cost to perform. Figure 1 displays potential roles of novel AKI biomarkers.
Fig. 1

Currently studied and possible roles for novel renal injury biomarkers. AKI acute kidney injury, FSGS focal segmental glomerular sclerosis, HUS hemolytic-uremic syndrome, ATN acute tubular necrosis, CKD chronic kidney disease

The role of genomics, proteomics, and metabolomics in biomarker development

Most recent AKI biomarkers were discovered using genomic and proteomic studies in animal models. Expression of newly identified biomarkers results from either upregulation or downregulation of genes or proteins in the kidney by different mechanisms of renal tubular cell damage and death (e.g. ischemia-reperfusion or nephrotoxic injury). A gene is a segment of DNA that encodes for an mRNA, which in turn encodes for a protein. Genomics refers to the analysis of gene expression of the entire genome of an organism [9]. Recent technological advances have permitted the development of this field, making it the forefront of biological research. An example of such technology is the microarray. Gene microarray assays, also known as chips, are a means to quantify the presence of genes in biological material. This is done in a large-scale fashion, referred to as high-throughput, in which the entire genome is analyzed simultaneously. The data obtained from high-throughput methods provide quantitative information on biological processes by linking genes to function, also known as functional genomics [8, 9].

Proteomics is the study of protein identification, and functional assessment using technology to separate, identify, and quantify proteins from biological samples [10]. Biological and pathological conditions cause variations in protein expression and function. Proteomics offers many tools useful in discerning these. Currently, the field of proteomics can be divided into two categories [11]. The first is an unbiased form of search for which no prior knowledge of the biological system is needed, which is referred to as discovery-oriented proteomics (using two-dimensional gel electrophoresis, liquid chromatography coupled to mass spectrometry, capillary electrophoresis coupled to mass spectrometry, and surface-enhanced or matrix-enhanced laser desorption/ionization time-of-flight mass spectrometry). The second category of proteomics is referred to as systems-oriented proteomics for which the investigator has an interest in a subset of proteins that are linked by function, location, or homology [11], using high-throughput method protein-detecting microarrays, with which many candidate proteins in a biological fluid can be evaluated at one time.

A relatively new area of biomarker research called metabolomics is emerging. Metabolites are unique low-molecular-weight molecules (e.g. amino acids or carbohydrates) that are produced by intracellular metabolism in various cellular processes. Differential levels of metabolites can be considered as the end product of downstream events occurring due to a disease. The study of metabolomics refers to systematic study of metabolites and their changes in biological samples due to physiologic or genetic changes. These metabolites or metabolite expression profiles may also serve as biomarkers of disease or physiological state [12, 13].

Once a novel candidate biomarker (such as a protein) has been identified, it must be evaluated in a careful, stepwise fashion prior to clinical use in humans. The series of events leading to the use of a new biomarker at the bedside may be referred to as “biomarker development”.

Biomarker development

Contrary to the fast pace of biomarker discovery in a laboratory setting, the rate at which these new biomarkers are actually being implemented into research studies or clinical medicine is slow. The biomarker validation process from bench to bedside is both complex and time-consuming. Phases for biomarker development and validation were proposed by Pepe et al. [14] in the context of cancer biomarkers. These “phases” were adapted for AKI research by Coca et al. [8] and are displayed in Table 2.
Table 2

Proposed phases of biomarker development, modified from Coca [8] and Pepe [14]



Biomarker aims

Discovery phase

Phase 1

• Identify leads for AKI biomarker

• Prioritize identified leads

Translational phase

Phase 2

• Evaluate diagnostic accuracy

• Assess the ability to distinguish AKI from non-AKI

Phase 3

• Evaluate the capacity of the biomarker to detect preclinical AKI or other AKI characteristic (e.g. severity, prognosis)

• Define criteria for a positive screening test (e.g. cut-off)

Validation phase

Phase 4

• Determine the operating characteristics of the biomarker-based screening test in a relevant population by determining the detection rate and the false referral rate (prospective study)

Phase 5

• To estimate the reduction in AKI morbidity and mortality afforded by the screening test (effects on population)

Suppose, hypothetically, that a new potential protein biomarker of AKI thought to be involved in mediating renal tubular cell injury was identified in mice exposed to ischemic renal injury. The first phase of biomarker development would explore and confirm the presence of these biomarkers in humans with AKI. A phase 2 biomarker study would develop a clinically useful assay for this protein, compare concentrations of this biomarker in patients with established AKI to patients without AKI, evaluate which other conditions might affect biomarker concentrations (for example sepsis, urinary tract infection, age groups, gender) and determine concentration norms in individuals without disease in order to elucidate AKI disease specificity. In a phase 3 study, the biomarker is evaluated on the role for which it was intended, in this case early AKI diagnosis, and to identify concentration cut-offs that appear to be useful for diagnosis. Diagnostic characteristics of the biomarker to predict AKI development should be calculated. If the biomarker appears to be a good early diagnostic test for AKI, a phase 4 study is performed. Phase 4 studies require the prospective follow-up of a relevant at-risk population (for example children admitted to the critical care unit), with multiple assessments of the biomarker concurrent with outcome assessment (AKI), in order to truly determine if the biomarker diagnoses AKI prior to SCr rise and performs with high diagnostic accuracy. Phase 4 biomarker studies usually require larger sample sizes and meticulous planning. The purpose of a phase 5 study is to evaluate to what extent the use of this biomarker actually leads to improvement in important outcomes, for example mortality and hospital costs at the population level.

These phases of biomarker development are guidelines and do not necessarily require that studies occur in the exact order presented. For example, urinary NGAL was initially discovered to be highly upregulated in mice with ischemic AKI. A human phase 3 biomarker study in children having cardiac surgery closely followed and revealed that urine NGAL had high diagnostic accuracy for predicting AKI [15]. However, subsequent other Phase 2 studies in other populations have revealed variable success of NGAL at early AKI diagnosis, exemplifying the importance of careful, stepwise biomarker evaluation.

Evaluation of biomarker diagnostic performance

During phase 2 biomarker evaluation, a biomarker is evaluated for its ability to discriminate between the presence or absence of disease and during phases 3 and 4 for diagnosing a disease at a clinically relevant time-point (e.g. before SCr rise). Success of a biomarker to perform these tasks is measured by its diagnostic characteristics, typically sensitivity, specificity, and area under the receiver operating characteristic curve (ROC) or AUC [16, 17].

When a biomarker is expressed as being above or below a certain cut-off value (normal versus abnormal), sensitivity and specificity are used to express its diagnostic accuracy. Sensitivity is the probability of a positive test result (e.g. high biomarker concentration) among patients who have the disease (such as in a cross-sectional phase 2 study) or who will develop the disease (phase 3 or 4 study). In other words, it is the depiction of the extent to which a test can detect the disease if it is present (Table 3). The specificity is the probability of a negative test result (e.g. low biomarker concentration) among those who do not have the disease (Table 3). An ideal but unlikely scenario is for a test to have 100% sensitivity and 100% specificity. In the true clinical setting, attaining high sensitivity often occurs at the expense of decreasing specificity and vice versa and a decision is sometimes required to prioritize these two characteristics (sensitivity vs. specificity), depending on the disease in question and the proposed utility of the biomarker being studied.
Table 3

Summary of calculating diagnostic characteristics: sensitivity, specificity, positive predictive value, and negative predictive value


Disease present

Disease not present


Biomarker test positive

True positive (TP)

False positive (FP)

TP + FP = total with positive test

Biomarker test negative

False negative (FN)

True negative (TN)

FN + TN = total with negative test


TP + FN = total with disease

FP + TN = total with no disease


How to calculate diagnostic characteristics


TP/(TP + FN): TP / total with disease



TN/(TN + FP): TN / total without disease


 Positive predictive value (PPV)

TP/(TP + FP): TP / total with positive test


 Negative predictive value (NPV)

TN/(TN + FN): TN / total with negative test

When a biomarker value is evaluated on a continuous scale, the ROC curve is used to evaluate diagnostic performance. The ROC curve is a plot of the true positive rate (sensitivity) versus the false-positive rate (1-specificity) at multiple different concentration cut-off values of a biomarker (Fig. 2). The AUC summarizes the information obtained from the ROC curve; an AUC of 1 denotes a perfect test, with 100% sensitivity and 100% specificity at all biomarker cut-off values. An AUC of 0.5 denotes that the biomarker does not increase diagnostic ability beyond chance alone. The ROC may be used to determine the best biomarker concentration cut-point which maximizes sensitivity and specificity for diagnosis.
Fig. 2

Three hypothetical receiver operating characteristic curves: plot of the sensitivity of a test versus 1-specificity for many different cut-off values of a biomarker. An AUC = 0.9 (red curve) represents an excellent diagnostic test for diagnosis of a disease; an AUC = 0.5 (black) represent a test which provides no more information than that provided by chance guessing alone. AUC area under the curve

Diagnostic testing studies often report two other diagnostic characteristics: the positive and negative predictive values. The positive predictive value is the probability of actually having the disease of interest when the test is positive. Conversely, the negative predictive value is the probability of not having the disease when the test is negative [17]. Great caution must be used when calculating and reporting these characteristics, since they are highly affected by the prevalence of the disease in the population studied. When the proportion of patients with the disease of interest is very high, the positive predictive value becomes highly overestimated, and vice versa for the negative predictive value.

Novel urine AKI biomarkers

Several candidate biomarkers discovered from animal models have passed the phase 1 stage of biomarker discovery and have been studied in humans, showing significant promise for use as early diagnostic tests of AKI and for portending outcome. The remainder of this review focuses on specific biomarkers that have been studied in pediatric AKI (summarized in Table 4) highlighting the most relevant adult studies.
Table 4

Representative summary of studies on biomarkers of acute kidney injury performed in children

First author, year N

Biospecimen source

Clinical setting

AKI definition

Time of biomarker measurement


Outcome used for biomarker


Sensitivity (cut-off value)

Specificity (cut-off value)


 Yilmaz 2009, 89 [29]



Not applicable

At presentation

ELISA (Cat no:CY-8070; CircuLex)

Predict UTI diagnosis


0.97 (20 ng/ml)

0.76 (20 ng/ml)

 Lavery 2008, 20 [68]


Premature infants

Not applicable

First 24 h of life and daily for 4 days

ELISA (AntibodyShop, Gentofte Denmark)

Detection of uNGAL in premature infants


 Bennet 2008, 196 [24]


Cardiopulmonary bypass (CPB)

≥50% SCr increase

At baseline and every 2 h for 12 h. Twice per day for up to 48 h, then daily for up to 72 h

ELISA (AntibodyShop, Gentofte, Denmark) ARCHITECT

Early prediction of AKI development and its severity


0.82 (100 ng/ml)

0.9 (100 ng/ml)

 Wheeler 2008, 143 [69]


Critically ill children with SIRS or septic shock

AKI: BUN>100 mg/dl or SCr>2 mg/dl or need for dialysis

First 24 h of ICU admission and on third day of admission

ELISA (AntibodyShop, Gentofte, Denmark)

Early prediction of AKI

At 24 h (0.71) At 48 h (0.64)


 Dent 2007, 120 [25]


Cardiopulmonary bypass

≥50% SCr increase

At baseline and at 2 h, 6 h, 12 h, and 24 h after CPB

ELISA (AntibodyShop, Gentofte, Denmark) Triage kit

Early prediction of AKI


0.84 (150 ng/ml)

0.94 (150 ng/ml)

 Zappitelli 2007, 140 [70]


Critically ill children

≥50% SCr increase

At 14:00 everyday for 4 days

ELISA (AntibodyShop, Gentofte, Denmark)

Early prediction of AKI and its severity

AKI diagnosis at first 48 h (0.78)

AKI 0.77 (0.2 ng/mg Cr)

AKI 0.72 (0.2 ng/mg Cr)

 Hirsch 2007, 91 [26]

Urine and serum

Contrast-induced nephropathy

50% SCr increase

At baseline, and at 2, 6, and 24 h after contrast

ELISA (AntibodyShop, Gentofte, Denmark)

Prediction of contrast induced nephropathy

At 2 h:

At 2 h:

At 2 h:


Plasma (0.91),

Urine (0.92)

Plasma (0.73),Urine (0.73) (100 ng/ml)

Plasma (0.98),Urine (1.00) (100 ng/ml)

 Trachtman 2006, 34 [28]

Urine and serum

Diarrhea + hemolytic uremic syndrome

Not applicable

Daily during the first 7 days, day 10 of admission. Days 7, 14, 28 and 60 post-discharge

ELISA (AntibodyShop, Gentofte, Denmark)

Predictions of severe renal injury and dialysis need


Urine need of dialysis: 0.9 (200 ng/ml)

Urine need of dialysis: 0.54 (200 ng/ml)

 Mishra 2006, 25 [71]

Kidney tissue

Kidney transplant

Not applicable

1 h of reperfusion


Early prediction of ischemic renal injury


 Parikh 2006, 53 [72]


Kidney transplant

Delayed graft function (DGF): dialysis need within the first week post-transplant

Spot urine sample at first 24 h (day 0) following transplantation

ELISA (AntibodyShop, Gentofte, Denmark)

Early biomarker of early graft function


0.9 (1,000 ng/mg Cr)

0.83 (1,000 ng/mg Cr)

 Mishra 2005, 71 [15]

Urine and serum

Cardiopulmonary bypass

50% Cr increase

Urine: every 2 h in first 12 h, then every 12 h. Serum: 2 h after CPB, every 12 h for the first day, and then once daily for 5 days

ELISA (AntibodyShop, Gentofte, Denmark) Immunoblot

Early biomarker of ischemic renal injury after CPB

At 2 h

At 2 h

At 2 h


Urine (0.998),

Urine (1.00),

Urine (0.98),


Serum (0.906)

Serum (0.5) (50 μg/l)

Serum (1.0) (50 μg/l)



 Washburn 2008, 137 [38]


Critically ill children


2:00 pm each day for 4 days

ELISA kit (Medical and Biological Laboratories, Nagoya, Japan)

Prediction of AKI development and severity

AKI within 24 h 0.54

AKI within 24 h 0.38 (75 pg/ml)

AKI within 24 h 0.78 (75 pg/ml)

 Parikh 2006, 53 [72]


Kidney transplant

Delayed graft function (DGF): dialysis need within the first week post-transplant

Spot urine sample at first 24 h (day 0) following transplantation

ELISA kit (Medical and Biological Laboratories, Nagoya, Japan)

Early biomarker of early graft function



 Parikh 2006, 55 [36]


Cardiopulmonary bypass

50% SCr increase

At baseline, every 2 h for the first 12 h, then once every 12 h for 5 days

ELISA kit (Medical and Biological Laboratories, Nagoya, Japan)

Early AKI prediction after CPB

At 12 h 0.75

At 12 h 0.5

At 12 h 0.94


At 24 h 0.73

At 24 h 0.4 (50 ng/ml)

At 24 h 0.94 (50 ng/ml)



 Nepal 2008, 2 [73]

Kidney tissue

Urate Nephropathy

Not applicable



Marker for proximal tubular injury


 Han 2008, 40 [50]


Cardiopulmonary bypass

50% SCr increase

At baseline. Every 2 h post-CPB in the first 12 h and then once every 12 h

ELISA (MaxiSorp; Nunc, Naperville, IL, USA)

Early AKI diagnosis

At 12 h 0.83

At 12 h 0.74

At 12 h 0.9


At 24 h 0.78

At 24 h 0.65 (2 ng/mg Cr)

At 24 h 0.8 (2 ng/mg Cr)



 Nguyen 2008, 106 [74]


Cardiopulmonary bypass

50% SCr increase

Before surgery and at 2 and 6 h


Prediction of AKI and its adverse clinical outcome

At 2 h 0.92

At 2 h 0.85 (2.4)

At 2 h 0.96 (2.4)

Cystatin C and Beta2-microglobulin (B2M)

 Herrero-Morin 2007, 25 [75]


Critically ill children

GFR of < 80 ml/min/1.73 m2 (measured by CrCl and estimated by Schwartz)

Every morning

Endpoint nephelometry in a BN-ΙΙ device

Detection of AKI in critically ill children

GFR by CrCl: S. cystatin C 0.802, B2M 0.851

GFR by CrCl: S. cystatin C 0.85 (0.6 mg/l), B2M 0.85 (1.5 mg/l)

GFR by CrCl: S. cystatin C 0.63 (0.6 mg/l), B2M 0.54 (1.5 mg/l)

Metabolomics (HAV-SO4)

 Beger 2008, 40 [64]


Cardiopulmonary bypass

50% SCr increase

At baseline. At 4 and 12 post-CPB

Ultra performance liquid chromatography/mass spectrometry

Predicting early AKI

At 12 h 0.95

At 12 h 0.9 (24 ng/μl)

At 12 h 0.95 (24 ng/μl)


 Liu 2009, 39 [63]


Cardiopulmonary bypass

50% Cr increase

At baseline, then at 2, 12, and 24 h after CPB

Bead-based cytokine multiplex kit (Bio-Rad) Luminex LabMPA multiplex system

Early AKI prediction

IL-6 at 2 h 0.76, IL-8 at 2 h 0.74

IL-6 at 2 h 0.78 (125 pg/ml), Il-8 at 2 h 0.83 (40 pg/ml)

IL-6 at 2 h 0.8 (125 pg/ml), Il-8 at 2 h 0.45 (40 pg/ml)

Neutrophil gelatinase-associated lipocalin (NGAL)

Lipocalins are proteins with binding sites that allow them to carry small lipophilic molecules (mainly iron-binding molecules) between cells in the body. NGAL is a 25-kDa lipocalin secreted by activated neutrophils and expressed in many cells and is upregulated in several injury settings, including infection, cancer, and renal tubular injury. NGAL appears to be highly involved in several cellular responses, including bacteriostasis, cell proliferation and differentiation, and apoptosis [18, 19]. NGAL is mainly expressed in the loop of Henle and distal tubule segments, but is also filtered by the glomerulus and reabsorbed by the proximal tubule, leading to increased proximal tubule expression in settings of high filtered NGAL [18, 20]. Using genomic and proteomic technologies, NGAL was found to be highly upregulated in mouse models of ischemic and nephrotoxic AKI, with urine concentrations rising within a few hours of injury [21, 22]. Plasma NGAL also rises dramatically with AKI [15, 23].

The first NGAL-AKI human study was a prospective study in 71 children having cardiac surgery [15]. Urine and plasma NGAL rose dramatically within 2 h after surgery only in subjects who subsequently developed AKI. The same group has confirmed these results using different NGAL assays and has demonstrated a strong association of NGAL with outcomes such as length of stay, dialysis requirement, and mortality [24, 25]. In a similar patient population undergoing cardiac catheterization, urine and plasma NGAL from 2 h post-nephrotoxic radiocontrast, predicted occurrence of contrast nephropathy with AUC of 0.92 [26]. Interestingly, in similar studies performed in adults undergoing cardiac surgery, while NGAL was still predictive of AKI, AUC’s were lower (in the 0.7 range), possibly due to the presence of more chronic kidney disease and comorbidities in adults [27]. In a phase 2–3 study in a heterogeneous group of critically ill children (non-cardiac surgery), mean and peak urine NGAL were associated with the presence of AKI in a dose-response fashion (increasing NGAL with worsening AKI severity). In this population, the AUC for urine NGAL to predict AKI in 48 h was 0.78. It is not surprising that the predictive ability of urine NGAL for AKI was less in the critically ill population, since the cardiac surgery population represents a much more homogeneous group with a known onset and mechanism of AKI. In the above studies, AKI was most commonly attributable to acute tubular necrosis. NGAL has also been studied in other forms of kidney injury, such as hemolytic uremic syndrome [28], urinary tract infection [29], IgA nephropathy, lupus nephritis, or chronic kidney disease [30]. Further research in different acute or chronic clinical pediatric settings will aid in understand how best to use this very promising marker.

Interleukin-18 (IL-18)

Interleukin (IL)-18, an 18-kDa pro-inflammatory cytokine initially synthesized in its inactive form (24-kD) and subsequently cleaved by caspase-1 into its active form, is produced systemically and in renal tubular epithelial cells with AKI [31]. In mice exposed to ischemic AKI, high urine IL-18 concentrations were present along with increased caspase-1 expression and activity [32]. Administration of IL-18-neutralizing antiserum protected mice against ischemic AKI (determined by less histologic features of acute tubular necrosis), supporting the role of IL-18 in the pathogenesis of ischemic AKI [33].

A phase 2 study initially demonstrated that hospitalized patients with acute tubular necrosis (ATN) (including children and adults) had higher median urine IL-18 concentrations than other patients [34]. The first phase 3 biomarker study demonstrating the potential for urine IL-18 as an early AKI biomarker was in critically ill, ventilated adults previously enrolled in the Adult Respiratory Distress Syndrome (ARDS) trial [35], where urine IL-18 rose at least 24 h before SCr did in AKI. Subsequently, urine IL-18 concentrations were found to be higher in children with AKI, only 4–6 hours after cardiac surgery, peaking at over 25-fold from baseline at 12 h and remaining elevated up to 48 h after bypass [36]. However, a study in adults undergoing cardiac surgery showed less predictive ability of early post-operative IL-18 to predict AKI [37].

Washburn et al. found that peak urine IL-18 concentrations increased with worsening AKI severity in critically ill children (non-cardiac), but performed poorly as an early predictor of AKI [38]. In addition, IL-18 was much more elevated in children who were diagnosed with sepsis. Further studies will hopefully elucidate which patients and in what clinical settings an inflammatory marker such as IL-18 can be useful for early AKI prediction, and if this utility is independent of the effect of systemic inflammation as is seen with sepsis.

Kidney injury molecule-1 (KIM-1)

Kim-1 (104-kDa) is a type-1 transmembrane glycoprotein expressed in low levels of the normal kidney, which is highly upregulated in the proximal tubules after ischemic or toxic AKI. The extracellular domain has characteristics suggesting that KIM-1 is involved in cell–cell or cell–matrix interactions. With renal tubular injury, the extracellular domain of KIM-1 is cleaved from the transmembrane domain by proteolytic enzymes and released into the urine [3941]. KIM-1 gene and protein was originally found to be upregulated in rats exposed to nephrotoxic medication [42]. Zhang et al. found that gentamicin- and cisplatin-treated rats had increased KIM-1 protein expression in the proximal tubular epithelial cells (mainly S1/S2 segments) [43]. In other animal studies, the upregulation of the KIM-1 protein in animals with polycystic kidney disease, toxic renal injury, or massive proteinuria were also described [4449].

Almost no studies of KIM-1 have been published for children. However, early studies of adults (mostly phase 2) suggest that KIM-1 appears to discriminate patients with different types of acute tubular necrosis (hospitalized patients, critically ill patients, patients with acute graft rejection) from those without AKI [50, 51]. In addition, urine KIM-1 is associated with the adverse outcomes of mortality or dialysis requirement [52] and it predicts AKI development in adults undergoing cardiac surgery [53, 54]. The first reported pediatric study in 40 children undergoing cardiac surgery revealed that urine KIM-1 at 12 h post-operatively had an AUC of 0.83 for detection of subsequent AKI [50].

Liver-type fatty acid-binding protein

Proximal tubule cells express the 14-kDa protein liver-type fatty acid binding protein (L-FABP), which binds, helps transport and facilitates metabolism of urinary filtered free fatty acids. In mouse models of cisplatin-induced injury and unilateral renal obstruction models, L-FABP appears to play a reno-protective role [55, 56], potentially via anti-oxidant effects of binding fatty-acid oxidation products [57, 58]. In mice exposed to nephrotoxic cisplatin and to renal ischemia, L-FABP increased within minutes to hours of the AKI exposure and correlated with histological severity of the injury [59]. In clinical adult studies, L-FABP appears to be a promising early AKI biomarker of contrast-induced nephropathy [60]. In a comprehensive phase 2 study of adult hospitalized patients, L-FABP discriminated between AKI/non-AKI with an AUC of 0.93, was found to be higher in patients with established AKI, compared to several non-AKI hospitalized groups and healthy controls, and was associated with adverse hospital outcomes [61]. Portilla et al. studied 40 children undergoing congenital heart surgery, collecting urine specimens pre-operatively and at 4 and 12 h post-operatively with daily assessment of SCr [62]. They found that L-FABP concentrations rose significantly at 4 and 12 h in both patients with and without AKI, but that this rise was significantly higher in children who subsequently developed AKI. L-FABP rose at least 1 day before SCr did and predicted subsequent AKI with an AUC of 0.81. L-FABP also correlated with a longer length of hospital stay. More studies on the use of urine L-FABP are certain to emerge and hopefully in different types of AKI causes and patient populations.

Other studied or potential AKI biomarkers in children

After demonstrating in prior animal studies that several inflammatory cytokines were increased with AKI, Liu et al. evaluated IL-6 and IL-8 as early AKI biomarkers in a phase 3 study of 71 children undergoing cardiac surgery, using a previously studied prospective cohort [63]. Both IL-6 and IL-8 were shown to be good potential early AKI biomarkers (AUCs ranging from 0.71 to 0.84) and were highly associated with prolonged mechanical ventilation. In the only study examining the use of metabolomics in pediatric AKI, Beger et al. studied 40 children who had cardiac surgery, prospectively collecting urine (at 4 and 12 h after surgery) and SCr. They identified a metabolite of dopamine (homovanillic-acid-SO4) to be higher in patients who developed AKI, with a 90% sensitivity and 95% specificity to diagnose AKI at concentrations of 24 ng/µl [64]. Studies in adults have suggested that urinary enzyme or low-molecular-weight proteins may correlate with poor short-term renal outcome and hospital outcomes [53].

Serum cystatin C (CysC) has been evaluated in some studies as a potential early AKI biomarker. However, CysC is not a direct marker of renal tubular damage, but simply an alternative marker (to SCr) of glomerular filtration rate. CysC has been found to be more accurate at estimating glomerular filtration rate in the outpatient setting and possibly for detecting early signs of chronic kidney disease [65]. Therefore, it is reasonable to think that CysC may also be more accurate to reflect acute renal dysfunction. Though controversial, some adult studies support that CysC rises approximately 1 day before SCr does with AKI [66, 67]. However, much further research is required, especially in children.


As has been displayed by this review, AKI biomarker discovery is a rapidly advancing field. One of our biggest hurdles will be to make the next step from “study” to “clinical application”. As more candidate biomarkers prove useful for early AKI diagnosis and prognostication, we will need to address issues around cost-effectiveness and outcomes. As may be evident from the summary provided, it is unlikely that one single biomarker will perform well at diagnosing AKI in all patients with different types of AKI. Moreover, different biomarkers appear to rise at differing time-points along the AKI pathophysiology processes. Therefore, research on understanding how to combine information from different biomarkers and how to combine this information with currently used clinical data to improve prediction and prognosis of AKI is needed. Multi-center studies with the goal of recruiting large numbers of children with a variety of diagnoses will soon be required to assess the utility of biomarkers in the clinical setting. In the nearer future, it is likely that AKI biomarkers will be incorporated into clinical trials in order to risk-stratify subjects who should receive the intervention. The starting point of such studies is to incorporate the newly defined AKI biomarkers into clinical trials in order to risk-stratify subjects to determine who should receive the intervention. Finally, as understanding of renal injury biomarkers increases, future research will hopefully enable the use these novel markers not only for AKI diagnosis, but for identifying location of renal injury (i.e. tubular segment) and mechanism (e.g. nephrotoxic versus ischemic).


Dr. Zappitelli receives salary support from the Kidney Research Scientist Core Education and National Training Program and the Fondation de Recherche en Sante du Quebec.

Copyright information

© IPNA 2010