The Use of Renal Biomarkers in Pediatric Cardiac Patients With Acute Kidney Injury

Introduction

The assessment of acute renal injury (AKI) in critically ill pediatric patients has evolved during the past decade, with the goal to identify changes in renal function expeditiously and accurately for diagnostic and therapeutic interventions. Acute kidney injury occurs in 6% of pediatric hospitalized patients and 30% to 50% of postoperative cardiac patients in intensive care, with the prevalence rates dependent on the definition of AKI used.¹ The development of AKI has been demonstrated to increase the risk of mortality and contributes to morbidities, such as prolonged hospitalization, increased ventilation days, increased risks of acid-base disorders and endocrine abnormalities, and the potential development of chronic renal failure.² Historically, the true prevalence of AKI has been difficult to ascertain because of varying definitions of the disease. The definition of AKI has been standardized and has been expanded to encompass a broad array of clinical conditions, including pre-renal azotemia, acute tubular necrosis, acute interstitial nephritis, acute glomerular disease, and acute postrenal obstructive nephropathy.^3–5

Acute kidney injury leads to a rapid decline in the kidney's regulatory, secretory, and excretory functions, resulting in a decrease in glomerular filtration rate (GFR) and the retention of nitrogenous and other metabolic waste.⁴ In addition, medications that are renally eliminated have diminished clearance, resulting in drug-induced toxicities.⁵ To prevent these complications, it is important to identify early markers of AKI to implement appropriate treatment strategies. Urine output and serum creatinine are the primary indicators of AKI in both adult and pediatric patients. However, these markers have been found to be less sensitive when compared to actual renal function, with greater than 50% of kidney function loss required before significant changes in serum creatinine are observed.⁶ There remains a tremendous clinical need to identify biomarkers of AKI that can be efficiently incorporated into the care delivery model to allow for the early detection, localization, and treatment of AKI. The emergence of renal biomarkers as early indicators of renal dysfunction is beginning to be used in clinical care. This review will provide a brief overview of renal function in pediatric patients and the limitations with current clinical markers, the definition and classifications systems of pediatric AKI, and the emerging role of biomarkers specific to pediatric patients in diagnosing AKI. Finally, a review of key pediatric cardiothoracic surgery studies will examine the role of AKI biomarkers in the congenital heart disease patient.

Methods

A literature search was performed in the following databases: Medline (PubMed), SCOPUS (Elsevier), Embase (OVID), and Cochrane Central Register of Controlled Trials (CENTRAL), from January 1990 to present. An additional search was carried out using Google Scholar. Databases were searched using the following keywords: “pediatrics,” “renal function,” “acute kidney injury,” “AKI,” neutrophil gelatinase-associated lipocalin,” “NGAL,” kidney-injury molecule,” “liver-type fatty acid binding protein,” “cystatin C,” and “interleukin-18.” Articles were limited to English-language studies published in peer-reviewed journals, with additional publication identified from review articles published. Exclusion criteria were previous chronic kidney disease, adult patients, and transplant patients.

Clinical Markers of Pediatric Renal Function

Serum creatinine is the most commonly used marker of renal function and is a surrogate marker of GFR. Although administration of an exogenous marker is considered the “gold standard” (e.g., inulin) to measure GFR, it is rarely done in routine clinical practice.^7–9 Thus, the use of urine output and serum creatinine to estimate glomerular filtration remains the mainstay in the care delivery model, despite their inaccuracy and delayed diagnostic value in AKI.⁷ Challenges associated with the interpretation of serum creatinine specific to pediatric patients include: the influence of age, sex, nutritional status, and muscle mass composition on endogenous creatinine values⁴; serum creatinine does not identify specific location or extent of kidney damage; there is a delay of approximately 24 to 36 hours in the rise of serum creatinine values when compared to the time of kidney injury.⁷ The delay in creatinine rise is believed to be multifactorial, including the rate of endogenous creatinine generation, the rate of renal “reserve” to compensate for diminishing renal function, and the degree of renal compromise.^7,8

Despite the many limitations in the serum creatinine as a marker of renal function/dysfunction, it is a mainstay to estimate glomerular filtration in both pediatric and adult patients because of its ease and availability. Serum creatinine values are often incorporated into equations (e.g., Cockcroft-Gault, Schwartz) to calculate creatinine clearance (CrCl), an estimate of GFR (Table 1).^8,9 The Schwartz equation was developed in 1976 and was a frequently used standard to estimate GFR in children for approximately 3 decades.⁸ The Schwartz equation incorporated “k-values” to account for covariables such as age, sex, and height to standardize for ontogeny and patient size. Although often used in clinical practice, evidence suggests that the Schwartz equation overestimated GFR by 20% to 40%.⁹ The Conahan-Barratt equation was developed at approximately the same time as the Schwartz equation and uses serum creatinine to estimate GFR calculated with the variables of height and serum creatinine (along with the constant 0.43).⁹ Despite being available for 40 years, this equation has never been widely adopted into practice.⁹

Table 1. Equations to Estimate CrCl and GFR in Adult and Pediatric Patients

View Fullscreen

In 2009, the Schwartz equation was modified to the “bedside IDMS-traceable Schwartz equation” to address the standardization of serum creatinine measurements among laboratories using the isotope dilution mass spectrometry (IDMS)–traceable methodology. However, there remains insufficient evidence that the use of the new Bedside IDMS-traceable Schwartz equation is a better predictor for GFR in pediatric patients, especially when used to adjust medication dosing. A study by Padgett et al⁹ compared the Schwartz, Conahan-Barratt, and Bedside IDMS-traceable Schwartz equations in patients who simultaneously received either a 12- or 24-hour timed urine collection for creatinine clearance. A total of 91 stable pediatric patients (5 infants, 43 children, and 43 adolescents) without underlying renal dysfunction or AKI were enrolled in the study. There was a statistically significant difference between the timed urine creatinine collection and the mean calculated values for all 3 equations (Schwartz p = 0.016; Conahan-Barratt p = 0.026; and modified Schwartz p < 0.001).⁹ Further, the correlation to equation-based CrCl and GFR values was 0.71 for the Schwartz equation and 0.72 for both Bedside IDMS-traceable Schwartz and Conahan-Barratt, with inconsistent patterns among the 3 equations over-predicting and under-predicting CrCl and GFR values as compared with timed sampling.⁹ These data further support that the use of serum creatinine equations to estimate GFR must be cautiously interpreted in clinical care, especially during periods of rapidly changing renal function.

In an effort to standardize assessment of creatinine (both plasma and urine) across the wide range of age and weight across the pediatric population, investigators have used “normalization” approaches. For example, creatinine measurements are generally reported per volume (e.g., mg/dL), with a single “spot” collection assumed to be representative of a consistent endogenous production of creatine that measures GFR. However, researchers are beginning to examine the potential relationship between several endogenous biomarkers.¹ Baek and colleagues1 calculated 3 biomarkers-to-serum creatinine ratios to determine predictability of AKI, independent of volume. Although this is an emerging concept, the potential use of biomarker ratios could provide for the standardization of endogenous markers independent of intravascular or renal volume.

Urine output adjusted for weight per unit of time (e.g., mL/kg/hr) is frequently used to assess renal function, often combined with serum creatinine. A summary of pediatric urine output descriptions can be found in Table 2. Although urine output is an important clinical marker, there may be contributing factors that alter the clinical presentation. Diminished cardiac output, hypovolemia, prolonged fasting, and factors contributing to stress (e.g., pain) may stimulate the production of antidiuretic hormone, diminishing urine output independent of intrarenal insult.^3–6 Thus, it is critical that a sufficient time for urine collection be accounted for (>6 hours), with the goal duration of 12 hours for appropriate renal assessment. Finally, weight-based evaluation of urine output must take into consideration the potential role of obesity, especially in adolescents. Intravascular circulatory volume is independent of body weight in children and adolescents; thus, weight adjusted urine volumes may be falsely reduced in obese patients.⁶

Table 2. Urine Output Values in Pediatric Patients

View Fullscreen

AKI Classifications in Pediatrics

The current clinical guidelines to standardize the diagnosis and staging of AKI in pediatric patients are primarily based on 3 classifications: 1) the Pediatric Risk, Injury, Failure, Loss, End Stage Renal Disease (pRIFLE) criteria; 2) the Acute Kidney Injury Network (AKIN) criteria; and 3) the Kidney Disease: Improving Global Outcomes (KDIGO) classification.^3–5 A summary of the classifications can be found in Tables 3 through 5. In 2007, the Acute Dialysis Quality Initiative Group established the adult RIFLE criterion to assist in stratifying intensive care patients for risks of morbidity, mortality, hospital length of stay, and health care costs. These clinical markers were reevaluated for the pediatric patient population in 2011, and the pRIFLE criteria were established.³ The pRIFLE classification system combines urine output along with CrCl as determined by the Schwartz equation. The pRIFLE criteria suffer from several limitations, including: 1) the extrapolation of critically ill adult patients to pediatrics patients; 2) validation of the criterion in a limited sample of 150 pediatric patients; and 3) lack of baseline creatinine clearance in the pediatric sample population.

Table 3. Pediatric RIFLE (pRIFLE) Description and Criteria³

View Fullscreen

Table 4. Acute Kidney Injury Network (AKIN) Description and Criteria⁴

View Fullscreen

Table 5. Kidney Disease: Improving Global Outcomes (KDIGO) Description and Criteria⁵

View Fullscreen

The AKIN consensus guidelines were developed to expand on the pRIFLE guidelines, with small changes in criteria 1.⁴ The AKIN guidelines, however, were not validated because of a lack of prospective studies. The KDIGO criteria were established in 2012 as an international, multidisciplinary, clinical practice guideline for AKI.⁵ These evidence-based clinical practice guidelines are the most widely used standards in staging AKI in both adult and pediatric patients, although they were not originally intended for use in neonates. The National Institute of Diabetes and Digestive and Kidney Disease guidelines modified the KDIGO classification to include neonates.¹⁰ Although the pRIFLE, AKIN, and KDIGO guidelines established a standardized framework for the diagnosis of pediatric AKI, the staging criteria inherently suffer from the limitations described with serum creatinine and urine output. Nevertheless, the pRIFLE criteria are the most commonly used criteria to assess AKI in the critical care population, whereas KDIGO and pRIFLE are used in the general pediatric population.

Role of Biomarkers in Pediatric AKI

The early diagnosis and treatment of AKI are paramount in reducing AKI-associated morbidity and mortality. As previously discussed, routine diagnostic markers to identify AKI (serum creatinine, urine output) are not sensitive, specific, or timely in a patient with rapidly declining renal function; thus, there remains the need to identify biomarkers associated with early-stage AKI.^3–5 A biomarker is defined as a parameter of biochemical, genetic, or physiologic change that indicates the presence, severity, and/or progress of a disease. Because a biomarker can be measured in urine, serum, plasma, or other bodily fluids, the ideal pediatric AKI biomarker should be: non-invasive (being able to be done at the bedside or in a clinical lab), highly sensitive and specific, and able to monitor kidney function independent of age, and should not have interference with drugs or nutritional products unique to pediatric patients.¹¹ In addition, the ideal biomarker should demonstrate a wide dynamic range and values that optimize risk stratification. Biomarkers used for renal impairment should be able to identify the primary location of the injury, identify the duration of kidney failure (acute versus chronic), identify the underlying etiology, and monitor the response to interventions.¹¹ A summary of AKI biomarkers and associated location of injury can be found in Table 6. These biomarkers are measured in urine, often using <2 mL for analysis. Although neutrophil gelatinase-associated lipocalin (NGAL) has been evaluated in both plasma and urine, urine sampling remains the preferred method because of its non-invasive nature.

Table 6. Renal Biomarkers and Their Associated Injury Location^*

View Fullscreen

Traditionally, diagnostic tests and biomarkers are initially discovered and studied in adult patients with the eventual translation to pediatric patients. However, the extrapolation of biomarkers to pediatrics is confounded by growth and development and may not be appropriate. Kidney function is composed of 3 key elements: glomerular filtration, tubular secretion, and tubular resorption. Age-dependent renal changes associated with ontogeny often result in physiologic differences compared with adults. For example, development of GFR, tubular excretion, and tubular reabsorption are markedly different in a 32-week gestational newborn compared with a 1-year-old.⁷ Ontogeny-derived differences in normal physiologic functions can lead to variations in biomarker values, independent of pathophysiologic conditions. Thus, the application of AKI biomarkers in adults may not be translatable to pediatric patients. Even within the field of “pediatrics” there can be significant physiologic differences between a newborn compared with adolescents. In addition, the effect of cardiopulmonary bypass, volume expanders, and the invasiveness of cardiothoracic surgery can further alter endogenous biomarker production. It is important that continued research examine the influence of maturation, development, and underlying congenital anomalies on renal physiology and pathophysiology.¹² Although considerable research remains, the following discussion highlights several biomarkers currently under review for application in monitoring pediatric cardiac patients. NGAL has been the most extensively studied AKI bio-marker in the pediatric cardiothoracic patient population to date, with other biomarkers having been examined in this patient population.

Neutrophil Gelatinase-Associated Lipocalin. NGAL is a 25-kDa protein found primarily in human neutrophils, with limited expression in the liver, spleen, and kidneys. Its primary function is to serve as a bacteriostatic agent in the endogenous immune response, depriving bacteria of iron, which inhibits their growth; additional physiologic functions include scavenging iron and inducing epithelial cell growth.¹³ The NGAL protein exists in 3 structural forms: a monomer, a homodimer, and a heterodimeric, which is conjugated to gelatinase. The heterodimeric and monomeric forms are the predominant forms in the tubular epithelial cells, with the homomeric form primarily in neutrophils. Intravascular NGAL undergoes extensive renal glomerular filtration due to its low molecular weight and positive charge, with secretion in the proximal tubule in response to ATP depletion.¹³ The thick ascending limb and collecting duct produce intrarenal NGAL.

Intrarenal NGAL concentrations are upregulated following an ischemic or nephrotoxic kidney injury. An increase in NGAL concentrations is detectable in the urine within 3 hours after renal injury, peaking at 6 hours following the insult.¹⁴ Elevated NGAL concentrations can increase 10-fold in serum and 100-fold in urine after acute injury, with clinical studies demonstrating that increases in plasma and urine NGAL are a powerful and independent predictor of AKI independent of underlying etiology, including in the pediatric patient population.^13–18 Chronic kidney disease, cardiorenal syndrome, anemia, pregnancy, and cancer can also increase NGAL concentrations.¹⁹ The NGAL commercially available test has been approved for pediatric clinical use in Europe as an acute-phase biomarker. Although it has not yet been approved by the US Food and Drug Administration, the number of clinical studies demonstrating benefit as an acute-phase biomarker in pediatric patients continues to grow.

Kidney Injury Marker-1. Kidney injury marker-1 (KIM-1) is a transmembrane protein that contains immunoglobulin domains and extracellular mucin, which appears to have a role in the molecular and cellular biology associated with AKI. Expression of KIM-1 is low in the normal kidney but is upregulated after ischemia-reperfusion injury and proliferates to de-differentiated epithelial cells of the proximal tubule 48 hours (about 2 days) after injury.¹³ During injury, the extracellular component is shed from the membrane in a matrix metalloproteinase–dependent manner that serves as the biomarker after renal ischemic or toxic injury.¹³ It is believed that KIM-1 acts as a phagocyte, engulfing apoptotic bodies and necrotic debris. Because KIM-1 appears to play a role in renal recovery and tubular regeneration, it is understandable why the timing of peak change would occur 2 to 3 days after injury. Similarly to NGAL, KIM-1 demonstrates age-dependent differences in urine concentrations. Healthy children 3 to 5 years of age demonstrate a mean urine concentration of 336 pcg/mL, whereas those of adolescents are similar to that observed in adults (515 pcg/mL).¹⁷ Unlike NGAL, there does not appear to be significant difference in values based on sex. Because of the latent timeframe in KIM-1 expression, its role in AKI may be more limited to kidney recovery compared with AKI; however, additional research is warranted in the pediatric patient population.

Interleukin-18. Interleukin-18 (IL-18) is a 22-kDa member of the interleukin cytokine family that regulates innate and adaptive immunity. IL-18 is produced by macrophages, mononuclear cells, and non-immune cells, including proximal tubule cells. During AKI, IL-18 concentrations double as a result of its role as an important proinflammatory cytokine mediator in acute ischemia AKI.¹³ The increased IL-18 concentrations are a result of the release from proximal tubule cells and not immune-mediated carriers. The rise in urine IL-18 occurs within the first 6 hours of renal insult but does not peak until 12 to 18 hours.¹⁷ A rise in urine IL-18 concentrations appears to be more specific for acute tubular necrosis compared with other etiologies. With the delayed increase of IL-18, the potential use of targeted therapies to ablate the proinflammatory properties of IL-18 is currently being explored.²⁰ A meta-analysis of IL-18 concentrations in adults found it a predictive AKI biomarker in various settings, including cardiac surgery and intensive care units for children, adolescents, and adults.²¹ The 95th percentile of IL-18 concentrations in 368 pediatric patients was 87.9 pg/mL.¹⁷ Similarly to NGAL, IL-18 appears to have sex-related difference in adolescents, with females >10 years of age demonstrating higher concentrations than males (273 pg/mL vs 71 pg/mL).¹⁷ The widespread use of IL-18 has been limited because of the lack of commercially available assay kits for clinical use, the delayed rise as an acute-phase biomarker, and the effect of concurrent inflammatory conditions on AKI predictability.

Liver Fatty Acid–Binding Protein. Liver fatty acid-binding protein (L-FABP) was first discovered in the 1970s as a binder to long-chain fatty acids and other lipids, responsible for transportation of fatty acids to mitochondria or peroxisomes. Although L-FABP is expressed in the liver, it can also be found in the intestine, pancreas, lungs, stomach, and kidneys. This 14-kDa protein is expressed in normal and diseased kidneys in both the convoluted and proximal tubules and appears to play an important role in kidney injury and repair; thus, it is often considered a renoprotective protein.¹³ When exposed to renal hypoxia, L-FABP expression is upregulated to decrease the severity of renal ischemia-reperfusion injury.20 Patients who develop AKI demonstrate a rise in urine L-FABP concentration within 6 hours of insult, with a gradual decrease until 12 hours after the event.¹⁷ Of note, an increase in L-FABP urine concentrations resumes at 80 hours (about 3.5 days) after insult. The secondary increase during the maintenance and repair stage of AKI is reflective of the renal protective nature of this protein. The 95th percentile value of L-FABP is 17 ng/mL across all pediatric age groups.¹⁷ Pediatric studies have demonstrated a slight decrease in L-FABP values with older adolescents; however, there remains considerable variation in concentrations throughout childhood.¹⁷ As mentioned previously, the variability in L-FABP values following tubular injury has limited its utility as a biomarker for AKI.

Cystatin C. Cystatin C is a 13-kDa protein that serves as an endogenous inhibitor of cysteine proteinase and is generally found in all nucleated cells. Cystatin C is freely filtered by the glomerulus and reabsorbed in the proximal tubule.¹³ The concentration of cystatin C in the blood is related to an individual's GFR and can serve as a stronger indicator of GFRs between 60 and 90 mL/min compared with creatinine. Serum cystatin C is not affected by muscle mass and does not transfer across the placenta, making it a viable option as a renal marker in the first few days of life.²² Cystatin C has been demonstrated to be superior to creatinine in animal models and chronic kidney disease; however, its role in AKI remains uncertain.^13,21–25 A recent study by Inker and colleagues²¹ demonstrated in adults that estimated GFR equations that incorporate both creatinine and cystatin C were more accurate in measured GFR compared with race-based creatinine or cystatin C equations used separately.²¹ Covariables such as diabetes, corticosteroids, inflammation, hyperbilirubinemia, hyperthyroid, and rheumatic factor appear to affect cystatin C concentrations.²⁵ Cystatin C urine concentrations vary significantly in pediatric patients. Higher mean values (1354 ng/mL) are observed in preterm neonates <26 weeks gestational age compared with neonates >30 weeks gestational age (209 ng/mL).^22–25 Similar studies in premature neonates demonstrated higher concentrations in premature infants and a general decline to adult values at approximately 1 year of age.^22–25 For example, preterm neonatal values range from 1500 to 1750 ng/mL, with a significant decrease in serum concentrations at 4 to 11 months (1000 ng/mL), which remains stable throughout adolescents (800 ng/mL).^22–24 The age-related changes in cystatin C may be due to the anatomic length of the renal tubules, with the absolute proportion reabsorbed likely to be less than that seen in adults. A commercially available kit is available for clinical use in the United States and Europe; however, the detection time after renal injury (12–24 hours) and its limited ability to provide injury location within the kidney have minimized cystatin C's clinical utility.

Renal Biomarkers in Healthy Pediatric Patients

NGAL has been the most extensively studied AKI renal biomarker in pediatric patients, especially patients with congenital heart disease. Studies have evaluated plasma (pNGAL) and urine (uNGAL) NGAL concentrations in both healthy and non-healthy pediatric patients, ranging from prematurity to adolescents.^7,12,14–21 Studies conducted in healthy pediatric patients demonstrate both age and sex influences on endogenous NGAL concentrations. Premature neonates demonstrate higher median uNGAL values (298 ng/mL; IQR, 196–240) compared with term infants (45 ng/mL; IQR, 33–211) and older children (20 ng/mL; IQR, 3–27).¹⁷ Nephrogenesis is virtually complete by 36 weeks of gestation and continues postnatally; thus, higher uNGAL concentrations are reflective of immature nephron development, most likely due to greater glomerular filtration and the inability of the proximal and distal tubules to resorb filtrated NGAL. Data in older children and adolescents appear mixed. Children 3 to 4 years of age (3.6 ng/mL; IQR, 1.9–2.5 ) and ages 5 to 9 years (4.5 ng/mL; IQR, 2.1–9.4) demonstrate lower uNGAL concentrations compared with children 10 to 14 years (7.6 ng/mL; IQR, 3.3–21.7) and 15 to 18 years (12.1 ng/mL; IQR, 6.4–270).¹⁷ In addition, uNGAL concentrations appear to be higher in females compared with males when adjusted for age, although these results are mixed. Females older than 10 years appear to have higher uNGAL concentrations, possibly due to contamination from vaginal secretions that contain neutrophils. Data for pNGAL studies in healthy pediatric patients are sparse, with 1 study solely evaluating plasma concentrations of NGAL in infants 0 to 6 years of age.²⁶ Wheeler et al²⁶ describe a median pNGAL concentration of 80 ng/mL, independent of age and sex. Urine concentrations of NGAL have been the most extensively studied renal biomarker in the pediatric cardiac patient population. Studies consistently demonstrate that uNGAL is an early predictor of AKI. A pooled sensitivity and specificity analysis for 9 pediatric studies was conducted by Filho and colleagues¹⁸ and demonstrated a uNGAL sensitivity of 0.76 (95% CI, 0.62–0.86) and specificity of 0.93 (95% CI, 0.88–0.96).¹⁸ In addition, the pooled analysis of pNGAL demonstrated a sensitivity of 0.8 (95% CI, 0.64–0.94) and specificity of 0.87 (95% CI, 0.74–0.94). As noted in Table 7, there are several limitations to implementing uNGAL monitoring into the routine care delivery model. Currently, the NGAL diagnostic kits are not indicated for use in pediatric patients. In addition, the age and sex variability in NGAL reference ranges in healthy children can be complex, independent of underlying disease. The specific urinary values defining AKI have not been determined across the pediatric age ranges; thus, further research is warranted for implementation in routine clinical practice.

Table 7. Summary of Acute Kidney Injury (AKI) Biomarkers^*

View Fullscreen

KIM-1 demonstrates age variation similar to that seen with NGAL. KIM-1 urinary concentrations are lower in 3- to 4-year-old patients (336 pg/mL; IQR, 217–513) and 5- to 9-year-old patients (386 pg/mL; IQR, 217–620) compared with 15- to 18-year-old adolescents (515 pg/mL; IQR, 195–866).¹⁷ Multivariate analysis has demonstrated a 3.8% (95% CI, 1.6–5.9) increase in KIM-1 urine concentrations for each increase in year of age. IL-18 also demonstrates a 2.2% increase for each year of age (95% CI, 0.4–4.0), but it also demonstrates a 49% increase based on female sex. IL-18 urine concentrations appear to be higher in 10- to 14-year-old patients (25.4 pg/mL; IQR, 15.6–37.3) compared with 3- to 4-year-old children (18 pg/mL; IQR, 10–30).¹⁷ L-FABP demonstrates an inverse relationship between age and urine concentration, with a 3.5% decrease in concentration for each increase in year (95% CI, −5.6 to −1.4). Mean L-FABP urine concentrations have been shown to approximate 3.4 ng/mL, with slightly higher concentrations in children 3 to 5 years of age (5.2 ng/mL) compared with 15- to 18-year-old adolescents (3.8 ng/mL).¹⁷ Similarly to NGAL, KIM-1, IL-18, and L-FABP do not have commercially available diagnostic kits for clinical practice. Further, there remains inconsistent association between baseline endogenous concentrations and the development of AKI in pediatric patients.

Renal Biomarkers in Pediatric AKI Secondary to Cardiac Surgery

The pediatric cardiac surgery patient population has been the most extensively studied AKI pediatric population to date. Cardiac surgery patients are more likely to develop AKI, baseline serum creatinine information is routinely available prior to surgery, sampling (urine and blood) is often performed within the context of the surgery, and the timing of AKI after surgery (onset 48–72 hours after surgery) can be predicted.^{12,15,19,27–29} A summary of key studies evaluating renal biomarkers in pediatric cardiac patients is presented below.

Fadel et al Study. Fadel et al¹⁵ examined pNGAL concentrations in 40 pediatric patients (2 to 78 months of age) who received cardiopulmonary bypass during pediatric cardiac surgery. Patients were divided into 2 groups: 1) 19 patients who developed pRIFLE-defined category II or III AKI; 2) 21 patients who did not develop AKI or were categorized as pRIFLE I. Spot serum creatine and pNGAL concentrations were obtained at 2, 12, and 24 hours following surgery. Mean pNGAL concentrations in patients with AKI were statistically higher compared with those patients without AKI at 2 hours (154 vs 66 ng/dL; p < 0.0001), 12 hours (201 vs 84 ng/dL; p < 0.0001), and 24 hours (238 vs 108 ng/dL; p < 0.0001) following surgery. Further, there were strong correlations between pNGAL concentrations and AKI at 2 hours (r² = 0.893), 12 hours (r² = 0.873), and 24 hours (r² = 0.795) following surgery. A cutoff concentration of 100 ng/mL at 2 hours and 125 ng/mL at 12 hours recorded the highest accuracy (95% for both), sensitivity (100% and 89.5%, respectively), and specificity (90.5% and 100%, respectively) for diagnosing AKI. The correlation of serum creatinine with NGAL concentrations was statistically significant at 24 hours after surgery (a predictable outcome due to the delayed rise in serum creatinine following renal insult). The results of this study demonstrated that pNGAL could be an early marker of AKI following cardiopulmonary bypass in the pediatric patient population

Krawczeski et al Study. Krawczeski et al¹² investigated the pattern and potential value of 4 urinary biomarkers for predicting cardiac surgery–associated AKI: NGAL, IL-18, L-FABP, and KIM-1. Spot urine samples were obtained prior to cardiopulmonary bypass and at 2, 6, 12, and 24 hours following bypass initiation. In addition, concurrent serum creatinine concentrations were obtained. Acute kidney injury was defined as a >50% increase in serum creatinine from baseline within 48 hours after bypass. The results of 220 patients (age range, 0.4 months to 6 years) demonstrated that AKI occurred in 27% of patients, with uNGAL significantly increased at 2 hours in those patients with AKI. Elevated concentrations of IL-18 and L-FABP were demonstrated at 6 hours, whereas KIM-1 concentrations were increased at 12 hours. At 6 hours, uNGAL, IL-18, and L-FABP were significant predictors of AKI, with NGAL demonstrating the highest discrimination. The authors calculated “receiver-operator characteristics” curves for each biomarker at each time point, with the resultant area under the curve (AUC) compared among the biomarkers. At 12 hours, uNGAL demonstrated the highest predictive AUC compared with the other 3 biomarkers. At 24 hours, all 4 biomarkers were significantly higher than the control group. The uNGAL concentration at 2 hours was independently correlated with ventilator days and hospital length of stay (p = 0.001). Finally, the investigators examined the combination of biomarkers to determine if the combination of biomarkers enhanced the predictive abilities. The authors did note that combination of the renal biomarkers led to an overall improvement in AKI prediction. The combination of uNGAL and IL-18 provided the best predictability of AKI at 6 hours, with a combination of all 4 biomarkers demonstrating the best predictability at 24 hours.

Baek et al Study. Baek and colleagues¹ examined AKI in 30 patients with congenital heart disease who underwent cardiopulmonary bypass surgery.¹⁹ Urine and blood samples of NGAL, KIM-1, and IL-18 were collected at baseline, 6, 24, and 48 hours after surgery. Serum and urine creatinine and blood urea nitrogen were also collected. Acute kidney injury was defined by the KDIGO criteria, although the specific stage was not identified. Of the 30 patients (mean age, 1.14 months), 12 patients developed AKI within 48 hours after surgery, with 8 developing AKI within 24 hours after surgery. The investigators collected serum creatinine and urinary biomarker concentrations with and without urinary creatinine adjustments. The median baseline uNGAL concentrations was 400 ng/dL at baseline, 357 ng/dL at 6 hours, 275 ng/dL at 24 hours, and 257 ng/dL at 48 hours in the 12 patients with AKI, which were not statistically different from the control group. KIM-1 urine concentrations ranged from 26 ng/dL at baseline to 195 ng/dL at 24 hours in the AKI cohort, subsequently falling to 62 ng/dL at 48 hours. There was no statistical significance demonstrated with IL-18 concentrations through all 4 time periods (range, 6 ng/dL at baseline vs 17 ng/dL at 6 hours) When urine biomarker concentrations were adjusted for urine creatine concentrations, the uNGAL/Cr ratio ranged from 34 ng/mg at baseline to 40 ng/mg at 48 hours. Urine KIM-1/serum creatinine ratio met AKI criteria at 24 hours after surgery (15 ng/mg) but did not at 48 hours following surgery. The uNGAL/serum creatinine ratios and uIL-18/serum creatinine ratios did not demonstrate a significant trend for the first 48 hours following surgery. Although the investigators concluded that KIM-1/Cr concentrations could be considered a strong predictor for AKI in younger children, the baseline uNGAL concentrations were considerably higher in this cohort compared with other studies.^7,17,18 Thus, there are concerns that the higher baseline uNGAL concentrations could not differentiate between AKI and non-AKI patients. The concept of standardizing urine biomarkers to another endogenous biomarker (e.g., urine creatinine) remains an interesting concept that warrants further investigation in the future.

Yoneyama et al Study. Yoneyama and colleagues²⁷ evaluated 103 patients (4 neonates, 35 infants <1 year of age, and 64 children between 1 and 18 years) to evaluate urinary biomarkers in patients who underwent cardiac surgery.²⁷ Acute kidney injury was defined as >50% increase in serum creatinine from baseline. Urine L-FABP and NGAL were obtained at intensive care unit admission, and at 4, 12, and 24 hours after surgery. Areas under the curve were calculated at each assessment time. Approximately 50% of patients developed AKI by the second postoperative day, with univentricular status, aortic cross-clamp time, and interoperative fluid balance independently associated with AKI development. Urine NGAL concentrations were statistically elevated in the AKI cohort compared with the non-AKI group at intensive care unit admission (45 vs 5 ng/dL), 4 hours (15 vs 5 ng/dL), 12 hours (22 vs 8 ng/dL), and 24 hours (23 vs 8 ng/dL). The uNGAL AUC ranged from 0.9 upon intensive care admission to 0.62 at 12 hours (4-hour AUC, 0.8; 24-hour AUC, 0.73). A statistically significant increase in L-FABP was demonstrated at all time points compared with non-AKI cohort: baseline (55 vs 20 ng/dL), 4 hours (40 vs 15 ng/dL), 12 hours (50 vs 20 ng/dL), and 24 hours (40 vs 5 ng/dL). Similarly to the study by Krawczeski et al, the investigators calculated “receiver-operator characteristics” curves for each biomarker at each time point, with the resultant AUC compared between L-FABP and NGAL. The L-FABP AUC ranged from 0.82 upon intensive care admission to 0.63 at 12 hours (4-hour AUC, 0.78; 24-hour AUC, 0.72). Patients with higher uNGAL concentrations at baseline and at 4 and 24 hours after ICU admission demonstrated longer intubation days, intensive care duration, and hospitalization compared with the control group. The authors concluded that uNGAL and L-FABP can be useful biomarkers to detect AKI in the pediatric cardiac patient population. Further, this study was unique compared with previous AKI studies because of its analysis of clinical outcomes, demonstrating worsening clinical outcomes with elevated uNGAL concentrations.

Summary

Acute kidney injury continues to be a common and serious condition, especially in the pediatric patient population undergoing congenital cardiac repair. Unfortunately, diagnostic and therapeutic interventions are often hampered, contributing to significant morbidity and mortality in this patient population. Enhanced understanding of the factors associated with AKI have led to the investigation of renal biomarkers that can identify AKI in a more timely manner: NGAL, KIM-1, L-FABP, IL-18, and cystatin C. Each biomarker has a unique profile that when evaluated in aggregate can assist in the diagnosis and treatment of AKI. As demonstrated in Table 7, the timing of biomarker expression following renal insult, the association between AKI and biomarker values in the pediatric patient, the ability to identify the location of the renal insult, and the commercial availability of assay kits all play an important role in determining their place in therapy.

Acute kidney injury occurs in 4 phases: 1) initiation of the injury; 2) extension of the injury; 3) maintenance of the ongoing injury; and 4) resolution.²⁰ Early identification of renal injury can facilitate therapeutic treatments to facilitate resolution. For example, during the initiation phase, therapeutic interventions, such as vasodilator therapy, antioxidant therapy, and/or iron chelation therapy, may assist in limiting disease progression. Because uNGAL appears to be an early biomarker of AKI, the routine use of uNGAL in postoperative patients could more rapidly identify patients progressing to AKI. Renal tubular cells undergo reperfusion-related death, with amplified inflammation during the extension phase. Biomarkers such as IL-18 and L-FABP could facilitate therapies with anti-inflammatory and antiapoptotic interventions. Finally, patients in the maintenance/resolution phases of the AKI demonstrate simultaneous renal cell death and regeneration. The use of KIM-1 may be the best biomarker to measure progression and the effect of therapeutic interventions, such as growth factors and stem cells.

The adoption of renal biomarkers into routine pediatric clinical practice still warrants significant research. Challenges associated with using biomarkers include the variability of reference values based on age and sex, as demonstrated by NGAL and IL-18. This variability illustrates the importance of growth and development on renal physiology and pathophysiology. The standardization of age cohorts based on renal ontogeny (e.g., neonates, 1 month to 2 years of age, and >2 years of age) could assist in the establishment of standardized reference ranges.

The use of urine sampling that requires small volumes (e.g., 1–3 mL) is particularly advantageous in pediatric patients, especially neonates. Reducing the need for repeated blood draws, especially in the postoperative period, is an attractive diagnostic option when screening for the potential development of disease (such as AKI). The use of an AKI urine panel that includes the aforementioned biomarkers has the potential to affect the outcomes associated with pediatric AKI. This is particularly relevant when considering concomitant pharmacologic therapy. Kidney function and renal clearance are important factors that can alter medication pharmacokinetics, leading to drug accumulation and adverse events. Earlier detection of AKI could help prevent drug-related kidney injury (acute and chronic) due to possibly avoiding repeated dosing of offending agents. Future pediatric studies are warranted to study the relationship between AKI biomarkers, glomerular filtration, and pharmacokinetic parameters. It is the goal that earlier identification and treatment strategies can reduce the morbidity and mortality in this vulnerable patient population.