**3.3 Acute kidney injury**

Acute kidney injury (AKI; previously referred to as acute renal failure) is reported to occur in 8% to 24% of preterm neonates admitted to neonatal intensive care units (Stapleton et al., 1987; Hentschel et al., 1996). The current diagnosis of AKI is primarily based on the RIFLE system, which categorises the stages of increasing AKI severity: Risk, Injury, Failure, Loss and End-stage kidney disease (ESKD) (Bellomo et al., 2004). This system was further modified following recommendations from the acute kidney injury network (AKIN) (Mehta et al., 2007). The initial clinical indication of AKI risk includes a 50% increase in serum creatinine (or ≥ 0.3 mg/dl within a 48 hour period), and/or a urine output less than 0.5 mg/kg/hr for a period of six hours (Bellomo et al., 2004; Mehta et al., 2007), which are changes indicative of a significantly reduced GFR. Classifications for the definition of AKI in a neonatal specific population, however, have not been developed.

The causes of AKI in the preterm neonate are primarily pre-renal in origin, arising from conditions which affect renal perfusion such as hypotension, hypoxia and sepsis (Stapleton et al., 1987; Cataldi et al., 2005). These in turn lead to apoptotic, necrotic and inflammatory processes within the kidney (Ueda and Shah, 2000; Bonventre, 2007). Importantly, AKI in the preterm neonate may subsequently lead to long-term chronic renal disease (Abitbol et al., 2003).

In a study involving 172 preterm neonates by Cataldi *et al.* 2005, the risk factors for AKI were found to be maternal and neonatal drug administration (non-steroidal antiinflammatory drugs (NSAIDs) and antibiotics, especially ceftazidime), a low Apgar score, and a patent ductus arteriosus. Interestingly, gestational age did not affect risk of AKI, however, the majority of AKI cases (79%) weighed < 1.5 kg at birth (Cataldi et al., 2005). In a larger study by Cuzzolin *et al.* (2006), involving 281 preterm neonates, a number of risk factors for AKI were also identified. These included maternal NSAID administration, low Apgar score, respiratory distress syndrome, neonatal drug administration (antibiotics and NSAIDs), and a number of clinical interventions (intubation at birth, catheterization, phototherapy, and mechanical ventilation).

Given the importance of the early diagnosis and treatment of AKI, there has been much recent focus on the discovery of novel urinary biomarkers. The expectation of a new biomarker is to enable the diagnosis of cellular injury before a decline in renal function occurs. For example, serum creatinine is not elevated until 48-72 hours after an acute injury has occurred (Moran and Myers, 1985); such a prolonged delay before diagnosis and treatment likely results in further renal injury. As Rosner (2009) describes, it would be optimal if a biomarker could be developed to: 1) assess the response to, and any adverse effects of therapeutic interventions 2) indicate the severity of renal injury 3) inform on the etiology of the injury and 4) identify the location of injured cells. In a systematic review of the current literature, Parikh *et al.* (2010) determined that the molecules with the most promise for the diagnosis of established AKI include interleukin-18 (IL-18), kidney injury molecule-1 (KIM-1), N-acetyl-beta-D-glucosaminidase (NAG) and neutrophil gelatinaseassociated lipocalin (NGAL). NGAL, IL-18, fatty acid binding protein (FABP), and cystatin-C are the most encouraging biomarkers for the early diagnosis of AKI, given that the upregulation of these molecules following injury onset precedes the rise in serum creatinine by many hours (Parikh et al., 2010).

In the preterm neonate, a small number of studies have been conducted for the assessment of urinary NGAL levels, with mixed results. These studies have shown that the highest NGAL levels are evident in the neonates that are critically ill, with and without evidence of renal dysfunction (Lavery et al., 2008; Parravicini, 2010); in particular, NGAL shows potential as a promising biomarker of late-onset sepsis (Lavery et al., 2008; Parravicini et al., 2010). Urinary NGAL levels also strongly correlated with gestational and postnatal age (Lavery et al., 2008; Huynh et al., 2009), perhaps reflecting the renal production of NGAL during nephrogenesis (Gwira et al., 2005) which is often still ongoing during the early postnatal period. Normative values for urinary NGAL in preterm neonates with uncomplicated clinical courses have also been published, with the results indicating a greater variation in females than males (Huynh et al., 2009).
