Acute Kidney Injury in Neonatal Intensive Care Unit: Epidemiology, Diagnosis and Risk Factors

Acute kidney injury (AKI) is associated with long-term consequences and poor outcomes in the neonatal intensive care unit. Its precocious diagnosis represents one of the hardest challenges in clinical practice due to the lack of sensitive and specific biomarkers. Currently, neonatal AKI is defined with urinary markers and serum creatinine (sCr), with limitations in early detection and individual treatment. Biomarkers and risk factor scores were studied to predict neonatal AKI, to early identify the stage of injury and not the damage and to anticipate late increases in sCr levels, which occurred when the renal function already began to decline. Sepsis is the leading cause of AKI, and sepsis-related AKI is one of the main causes of high mortality. Moreover, preterm neonates, as well as patients with post-neonatal asphyxia or after cardiac surgery, are at a high risk for AKI. Critical patients are frequently exposed to nephrotoxic medications, representing a potentially preventable cause of AKI. This review highlights the definition of neonatal AKI, its diagnosis and new biomarkers available in clinical practice and in the near future. We analyze the risk factors involving patients with AKI, their outcomes and the risk for the transition from acute damage to chronic kidney disease.


Introduction
Acute kidney injury (AKI), frequently involving patients admitted to the neonatal intensive care unit (NICU), is associated with poor outcomes and affects 18-70% of critically ill neonates [1,2].Although the real incidence of AKI in neonates is uncertain, with wide and various ranges, due to several definitions and diagnostic methods, more immature and ill neonates are characterized by the highest risk of AKI [3].Patients with congenital heart disease, perinatal asphyxia, premature birth or a low birth weight, and necrotizing enterocolitis, as well as neonates who receive nephrotoxic medications or require extracorporeal life support, represent high-risk populations.
According to the Worldwide Acute Kidney Injury Epidemiology in Neonates (AWAKEN) data, one episode of AKI, related to an increased length of hospitalization and mortality rate, was observed in 30% of critically ill neonates [4].
To apply personalized therapeutic approaches to AKI, an adequate and early diagnosis could reduce or avoid subclinical and precocious renal injuries and their related effects, requiring chronic follow-up, as suggested by the Clinical Practice Guidelines for Acute Kidney Injury (KDIGO), recommending a nephrological evaluation after three months, if an AKI event occurs [5,6].The precocious diagnosis of AKI based on urinary output and serum creatinine (sCr) levels represents one of the hardest challenges in clinical practice.Several biomarkers and clinical scores were assessed to predict neonatal AKI, to identify by a reduced glomerular filtration rate (GFR) and tubular reabsorption of creatinine [19].For these reasons, creatinine cannot be considered an adequate marker for the diagnosis of acute kidney neonatal injury.
AKI management in the NICU is related to the fluid management of patients and the assessment of water balance.Fluid overload and AKI are cause-and-effect potentiators, with a high risk of detecting falsely reduced creatinine, underestimating AKI [20].
In intensive care, fluid management is often based on massive volume expansion, especially following surgery or resuscitation for sepsis, increasing the total body water by more than 10% within 72 h, reducing sCr concentrations and underestimating the severity of kidney injury [21].
Renal replacement therapy (RRT) "timing" is based on the severity of fluid overload, as well as severe acidosis or electrolyte imbalance, i.e., refractory hyperkalemia, behind a single value of sCr.All confounding risk factors to renal failure determined individually are analyzed, and the treatment is personalized [22].
In this context, peritoneal dialysis (PD) is preferred in neonates with AKI, whereas major abdominal surgery, severe hyperammonaemia, necrotizing enterocolitis and abdominal wall defects represent the main contraindications.The risk of leaking is higher, and thereafter, so is infection after catheter placement [23].
Newborns undergoing cardiopulmonary bypass were the most studied population, but further research should evaluate these biomarkers in other critical pediatric populations, such as very-low-birth-weight (VLBW) neonates [30].
For the same biomarker, urine evaluation can provide more specific information about renal injury than serum analysis, avoiding the influence of systemic inflammation or alterations due to concomitant non-renal diseases [31].
However, more works are needed to validate all these following urinary and sera biomarkers, which are not used in common clinical practice, with the exception of cystatin C.

Neutrophil Gelatinase-Associated Lipocalin (NGAL)
Neutrophil gelatinase-associated lipocalin (NGAL) is a component of innate immunity, released by activated neutrophils and epithelial cells of several tissues, including the renal thick limbs of Henle and collecting ducts [32].NGAL, detectable in the blood and urine precociously, represents a potential AKI biomarker, highlighting hypoxic-ischemic damage involving renal tubular epithelial cells [33].
In particular, high urine and serum NGAL levels early revealed AKI after cardiac surgery [30].In a prospective trial that enrolled critically ill children, urine NGAL predicted AKI before an increase in SCr [34][35][36], as well as in premature infants with AKI, in whom high urinary NGAL concentrations were independently associated with mortality [37,38].Similarly, Merve showed higher NGAL concentrations in 70 premature calves with respiratory distress syndrome and AKI, closely related to a 4-fold high mortality risk [39].
Another clinical application of this biomarker is the detection of hypoxic-ischemic AKI, responsible for renal tubular epithelial cell damage, with increased urine and serum NGAL concentrations [33].Moreover, NGAL is an early predictive biomarker of contrast-induced nephropathy in children [40].
In particular, nephrotoxic medication exposure in NICU induces an increase in urine NGAL, with a risk for AKI of 2.76 for urine NGAL levels ≥ 400 ng/mL [41].
A prospective observational study involving very low birth weight assessed the reference range of urine NGAL concentrations.The median, the 25th, 75th and 95th quantiles and the range of pooled urine NGAL concentrations were 5 ng/mL, 2 ng/mL, 10 ng/mL, 50 ng/mL and 2-150 ng/mL, respectively.These data are not valid in patients with gestations < 26 weeks or with birth weights ≤ 750 g with uncomplicated clinical courses [35].

Kidney Injury Molecule-1 (KIM-1)
After ischemic or toxic damage, renal proximal tubules release kidney injury molecule-1 (KIM-1), a type I transmembrane protein [42,43].The role of KIM-1, associated with tubulointerstitial inflammation and fibrosis, has not been fully elucidated; however, its role is related to apoptosis and the regeneration of injured tubules, participating in the restoration of tubular structure and function.Moreover, macrophages other than proximal tubular epithelial cells are implicated in tubular repair [44].
This biomarker has been evaluated in the same pathological fields of NGAL, from cardiac surgery to AKI related to nephrotoxic medications.However, contrasting data are available in the literature, without a clear indication to use this biomarker alone in the pediatric setting, but it is hypothesized that a panel of biomarkers with KIM-1 can reveal damage in proximal renal tubules.

Cystatin C
Serum cystatin C (CC) is entirely removed from the blood by glomerular filtration because of its low molecular weight and positive charge, reflecting a better GFR than those of other low-molecular-weight proteins, including creatinine.It is then reabsorbed and catabolized in the proximal renal tubule [45].Infections, inflammatory diseases and liver diseases do not alter filtration, suggesting that CC represents a marker of GFR more accurately than serum creatinine [46].It has been demonstrated that sCr and CC levels are significantly higher in AKI patients, correlating with each other [47].There is a strong correlation between serum CC concentrations and sCr levels in pediatric AKI patients, revealing a better diagnostic profile of serum CC than that of sCr [48].Similarly, in a prospective, observational study, serum CC and βeta-2 microglobulin (β2MG) had more sensitive and specific properties than those of sCr in detecting AKI in critically ill children [49,50].Moreover, serum CC levels detected AKI early in preterm neonates with respiratory distress syndrome (RDS) [51].
Reference values for CC in preterm infants were established and compared with sCR, and they ranged from 1.25 to 2.84 mg/L at the first day of life, significantly decreasing after 3 days [50].
CC concentrations in children reach adult values by the age of 1 year (range, 0.50-1.27mg/L; adult range, 0.51-0.98),whereas higher CC levels are assessed during the first year of life, reflecting the immaturity of the kidneys (<1 year: 0.59-1.97mg/L) [52].
CC levels in healthy Turkish neonates were determined to be 1.41 mg/L, 1.22 mg/L and 1.21 mg/L for neonates born in gestational weeks 28-32 and 33-36 and in neonates born after gestational week 37, respectively [53].
Serum CC level was found to have a statistically significant association with AKI development in preterm neonates with respiratory distress syndrome (RDS) [51].
In more than fifty thousand Chinese neonates, cystatin C-related criteria, independent from gestational age and birth weight, defined AKI by a serum Cys-C of ≥2.2 mg/L or when an increase in Cys-C of ≥25% occurred, with a positive correlation with an increased risk of in-hospital mortality [54].
However, contrasting data refer also to CC, as assessed in neonates after cardiac surgery in whom, despite a significant difference in CC levels from baseline and after 13 h, only urine output was independently associated with AKI [55].
Further studies enrolling more homogenous pediatric and neonatal populations are needed to understand the role of CC as a precocious renal biomarker of AKI.

Interleukin-18
Interleukin (IL)-18, a pro-inflammatory cytokine that stimulates interferon-gamma production in T cells and natural killer cells, is upregulated in patients with AKI.In particular, during ischemic acute renal damage, a precursor of IL-18 is activated by caspase-1-mediated conversion and released by tubular cells, representing a marker of proximal tubular injury in ischemic acute tubular necrosis [56].High urinary IL-18 levels are characterized by ischemia, acute tubular damage and delayed graft function compared with other renal diseases, allowing for the early diagnosis of AKI [57].Other data suggest a prognostic role of urine IL-18 in patients with RDS in intensive care, reflecting the mortality risk [58].Similarly, urinary IL-18 predicted the severity of AKI and was an independent predictor of mortality in non-septic critically ill children, revealing a better diagnostic profile than that of sCr [59,60].

TIMP2 × IGFBP7
Tissue inhibitor of metalloproteinase-2 (TIMP-2) and insulin-like growth factor-binding protein 7 (IGFBP7) are two urinary G1 cell cycle arrest biomarkers for the early diagnosis of kidney damage.Following injury, renal tubular cells enter a period of G1 cell cycle arrest, leading to cellular repair, cell death or cellular senescence [61,62].
The product of these urinary markers [TIMP-2] × [IGFBP7] performed better than any other biomarker in detecting AKI according to the KDIGO criteria, including neonatal and pediatric patients [63].Pediatric studies using these biomarkers to predict AKI are limited.
In post-cardiac surgery, the diagnostic properties of these markers were comparable to NGAL, with better profiles assessed at 4 h and 12 h post-cardio-pulmonary bypass for predicting AKI [64,65].
Moreover, severe AKI was assessed in 230 patients undergoing complex cardiac surgery using [TIMP-2] × [IGFBP7], predicting late severe AKI but not precocious AKI, as early as 2 h after the start of cardiopulmonary bypass [66].[TIMP-2] × [IGFBP7] was tested in critically ill neonates to predict AKI and severe AKI, revealing its independent predictive role, independent of sex, illness severity and gestational age, in the heterogeneous critical neonatal population [67].
Further prospective studies are required in larger pediatric populations to evaluate the role of these biomarkers in different subtypes of AKI.

Prematurity and Low Birth Weight
Preterm neonates represent a population at a high risk for AKI due to a potential incomplete and abnormal process of nephrogenesis, resulting in decreased nephron mass [68].The reduction in renal mass in the preterm became the subject of follow-up studies on cohorts of preterm infants.These studies identifyied subjects at risk, and implemented preventive interventions and revealed the connection between low birth weight, decreased nephron mass and blood hypertension [69,70].
According to Barker's hypothesis, low-birth-weight (LBW) babies could have a reduced nephron-related mass, altering the sodium transport and metabolism in the glomeruli, with a high risk of developing blood hypertension [71].However, no guidelines suggest the prescription of antihypertensive drugs for LBW patients with a risk of hypertension.In addition, neonates could have reduced renal blood flow associated with low GFR, increasing from 10 to 20 mL/min/1.73m 2 during the first week of life to 120 mL/min/1.73m 2 in the first year, reaching adult levels by 2 years of age [72][73][74].
The proximal tubule reabsorbs various substances filtered at the glomerular level (amino acids, bicarbonates and glucose) through active and passive transport mechanisms.These transport systems are immature in preterm infants, causing metabolic acidosis and glycosuria; the tubular phosphorus reabsorption system is more active in newborns than that in adults and can therefore lead to hyperphosphatemia [75].Hyperkalemia is frequently observed in preterm newborns due to the immaturity of sodium and potassium transport systems and the regulation of levels in the collecting duct.Moreover, intrauterine growth retardation can alter kidney maturation, interfering with nephrogenesis and resulting in a reduction in kidney size compared to reference values [76].Chen revealed that oliguric AKI had a higher mortality risk than that of non-oliguric AKI, independent of sCr levels and the severity of AKI.In particular, prenatal small-for-gestational-age and perinatal and postnatal adverse events were often associated with oliguric AKI, and nephrotoxin exposures caused non-oliguric AKI [77].
Mazhaeri confirmed this link between AKI, prematurity and LWB in association with low Apgar scores, mechanical ventilation and sepsis [78].

Sepsis
Sepsis-related AKI (sAKI) represents one of the main mortality risks in the NICU, of around 70% [79].Neonatal sepsis is frequently assessed in preterm newborns, secondary to pathogens acquired after birth, and is defined as early (ES) or late (LS) sepsis if diagnosis occurs within or after 72 h of birth, respectively [80].Neonatal sepsis is inversely related to birth weight and gestational age, with a high prevalence in very-low-birth-weight neonates [81] and with gestational age (GA) < 28 weeks [82].
Sepsis therapy is based on the administration of drugs that can have nephrotoxic action and worsen the renal damage caused by hypotension and hypoperfusion [83].However, contrasting data are reported about the subtypes of sepsis, underlining that Munyendo revealed that AKI complicates the clinical course of neonates with LS, whereas Holda demonstrated a high incidence of AKI in ES patients [84,85].
Moreover, primiparity and the presence of maternal fever within a week of delivery are maternal factors associated with AKI, as well as peripartum sepsis and exposure to antibiotics in the immediate peripartum period [86].

Post-Neonatal Asphyxia, Respiratory Distress and Mechanical Ventilation
Neonatal asphyxia represents a risk factor for AKI, closely related to its severity, and it induces multiorgan dysfunction, including renal failure, due to ischemic processes.
A wide range of AKI prevalence, between 30 and 56% [87,88] due to different definitions, involves neonates with asphyxia, representing, however, a non-negligible cause of AKI, with incidence also higher than that related to sepsis, according to some epidemiological reports [89,90].El-Kalioby reported other epidemiological data, including perinatal asphyxia (59.5%) as a risk factor of AKI, independent from RDS, shock, prematurity, sepsis and nephrotoxic drugs [91].Similarly, mechanical ventilation (MV) represents another risk of AKI due to hemodynamic factors or ventilator-induced lung injury, inducing pulmonary and systemic inflammation [92,93].Moreover, these data were confirmed in a retrospective study in which invasive MV, associated with a lower GA and birth weight, were independent risk factors for AKI in newborns [94].Underlining the potential role of systemic inflammation and several inflammatory pathways involved in this acute lung-kidney crosstalk, the relation between MV and AKI remains unclear.

Nephrotoxic Drugs
Nephrotoxic drugs represent a frequent cause of AKI in critically ill neonates with long-term complications, leading to potential chronic kidney disease (CKD).
Of neonates, 75% are treated with at least one nephrotoxic medication, i.e., aminoglycosides, during the first postnatal week in the NICU, with a close relationship with AKI [95].Behind the high rate of treated patients, this cause of AKI is potentially preventable.Nephrotoxic drugs can be avoided, their prescription can be modulated according to kidney function, or concomitant comorbidities that expose neonates in the NICU at a high risk of AKI can be evaluated.The link between drugs and acute renal dysfunction remains understudied, having only been researched by few studies, revealing, especially in premature neonates, that the cumulative exposition of drugs is closely related to AKI risk [96][97][98][99][100].
Nephrotoxicity depends on several factors, including age, comorbidities, the dosage of the drug and concomitant medications, considering that more than 80% of neonates often receive one or more nephrotoxic drugs, such as vancomycin or gentamicin [101,102].Infants with very low birth weights (VLBWs) are the most studied population in the NICU, revealing acute interstitial nephritis and acute tubular necrosis as the main physiopathological mechanisms of renal injury [103].The effects of these processes often lead to hypokalemia due to a reversible renal tubular waste of potassium, also including phosphate and calcium, with full recovery within 1-2 weeks after the discontinuation of the drug [96].
Behind bacterial infections, VLBW infants are characterized by a non-negligible risk of systemic fungal infection.When Candida sp.represents the major pathogen, amphotericin B is the treatment of choice, highlighting that the encapsulated liposome (AmBisome) forms do not induce nephrotoxic effects [105,106].
A preventive strategy to reduce the risk of nephrotoxicity related to drug administration was developed by Goldstein, who created the Nephrotoxic Injury Negated by Just-in-time Action (NINJA) program, identifying pediatric patients at a high risk for AKI according to sCr monitoring [107].Similarly, Stoops applied the NINJA program to the NICU, creating the 'Baby NINJA' and reducing nephrotoxic medication exposure and AKI rates in the NICU [108].These programs, associated with electronic alerts, could help clinicians reduce nephrotoxic medication exposures and AKI events.

Patent Ductus Arteriosus and Cardiac Surgery
Persistently patent ductus arteriosus (PDA) in preterm infants induces hemodynamic instability due to a left-to-right shunt, leading to chronic lung disease and AKI, necrotizing enterocolitis with a high risk of perforation [109].
PDA can cause a condition of relative hypovolemia and hypoperfusion with an increase in the synthesis of vasodilatory prostaglandins in an attempt to compensate for the hemodynamic effects of the duct.In this condition, non-steroidal anti-inflammatory drug (NSAD) administration reduces this vasodilation effect, but it may worsen renal perfusion, with consequent renal injury and damage [110].Although it is traditional to manage PDA with NSAD, guidelines for therapeutical management are lacking, with a high risk of adverse events related to these drugs [111][112][113].In particular, indomethacin can cause transient renal failure due to a reduction in prostaglandins, which act at the renal level, causing vasodilation mediated by the inhibition of vasopressin [74].Ibuprofen inhibits the vasodilator effects of prostaglandins on the afferent glomerular arterioles, reducing the GFR and diuresis [114].However, it induces fewer adverse effects on renal hemodynamics, if compared to indomethacin, representing the first choice for PDA treatment in preterm infants [115,116].
Cardiac surgery represents one of the most frequent causes of AKI in neonates, with a high risk of a longer stay and mortality [117].Different mechanisms are related to this type of surgery for AKI, such as a prolonged cross-clamp time, a prolonged MV, a longer time to reach a negative fluid balance and higher inotropic support [118].

Recurrent AKI and Transition to CKD
CKD represents a potential long-term complication in neonates who suffer from AKI during NICU stays [119,120].Moreover, AKI events can also relapse after an initial resolution before discharge from the NICU, defined as recurrent AKI (rAKI) according to the KDIGO definition.rAKI incidence remains largely unknown, ranging between 7% and 24% [2,121-123] due to few single-center studies that have assessed it in neonates.
GA and LBW are risk factors for rAKI, which is related to a longer stay in the NICU and to increased mortality if compared to patients who suffer from only one AKI event [121].rAKI is also related to a prolonged time of mechanical ventilation [123].These data, obtained from these single-center studies, were confirmed by the multicentric trial AWAKEN, which analyzed more than two thousand critically ill neonates, revealing a higher incidence of rAKI among neonates with the youngest GAs and lowest birth weights; rAKI also occurred in patients with severe first AKI episodes with low Apgar scores and if respiratory failure, sepsis or congenital heart disease occurred [124].
rAKI represents a prognostic factor that can be used to predict potential evolution to chronic kidney disease (CKD), as assessed both in adults and in childhood [5,[125][126][127][128][129].The progressive loss of nephrons, leading to initial hyperfiltration, associated with endothelial dysfunction, represents the final effect of maladaptive repair involving the kidneys, with pro-inflammatory processes, tubulo-interstitial fibrosis and glomerulosclerosis [119].
A literature review showed a wide range (10-70%) of CKD rates after AKI in children [130,131] due to the lack of trials analyzing outcomes with an insufficient follow-up time, requiring an analysis of CKD prevalence during adulthood.In particular, birth weight is negatively related to sCr levels in adults born very preterm.Similarly, intrauterine growth retardation (IUGR) is associated with CKD in subjects born very prematurely, suggesting that IUGR represents a non-negligible risk factor for renal failure, increasing systemic arterial stiffness and mean blood pressure [132,133].
These hypotheses were confirmed by Franke, who highlighted that, in more than four hundred children, SGA and prematurity represent risk factors for CKD [134].Other studies have revealed that proteinuria, albuminuria, a loss of renal mass and nephrocalcinosis characterize preterm infants with extreme LBW and AKI over an 18-year period [76,135].

Conclusions
AKI complicates the clinical course in critically ill neonates in the NICU, inducing prolonged hospitalization and a high mortality risk.Accurate and precocious diagnosis of AKI is required to evaluate the prevalence of AKI, the risk factors and a personalized therapeutical strategy and to create follow-up plans for infants at risk of rAKI and transitioning to CKD.Observational studies based on prospective physiological findings, including a large number of premature and term babies stratified according to gestational weeks and risk factors, are required.
To achieve these goals, new biomarkers should be tested and validated in neonatal and pediatric populations.However, to date, only cystatin C is suitable for clinical use, whereas all reported renal biomarkers are not commonly used in clinical practice.
Large multicenter studies with long-term follow-ups are needed to understand the physiopathological mechanisms and the risk of CKD in patients with AKI events that occur during childhood.