This case control study was conducted at Level III NICU of Shree Krishna Hospital (SKH) in Anand District of Gujarat, India. It is a referral center for government and private establishments of the surrounding districts. The NICU is a well-equipped, 20 bedded state of the art facility with average newborn: nurse ratio of 2:1. The hospital has an electronic registry system for all admissions and central diagnostic laboratory. Screening criteria of serum creatinine >1.5mg/dl was used to identify the study population from the online registry of the central diagnostic laboratory. The medical records of all the newborns admitted, between January 2008 to December 2010, who had the diagnosis of AKI were reviewed. These were compared with the medical records of 100 NICU admissions without AKI during the same period. Controls were randomly selected from the list of hospital numbers showing serum creatinine <1.5mg/dl. Matching was not possible as rest of the details were not available in the online registry.

The diagnosis of AKI was made on the basis of serum creatinine levels >1.5 mg/dl [12–14]. The diagnosis of Respiratory Distress Syndrome (RDS), Meconium aspiration Syndrome (MAS), birth asphyxia, clinical sepsis, Culture proven sepsis, NEC, hyperbilirubinemia were as per World Health Organization (WHO) [15]. The babies with AKI were managed by fluid adjustment, adjusting dose of medications based on renal function, peritoneal dialysis and supportive care as per need. Data regarding clinical and demography profile and laboratory investigations were obtained from the admission register of NICU, admission files, labor register of obstetrics and gynaecology department and electronic registry.

Data was entered in Microsoft Excel 2007 and STATA 14 was used to analyze descriptive data. Chi square/independent sample t-test as applicable and logistic regression were used to establish an association of various factors and outcome with AKI. The p-value less than 0.05 were considered significant. Approval was taken from the Institutional Human Research Ethics Committee (HREC) of Shree Krishna Hospital (SKH).

During the study period, out of 1745 NICU admissions (819 inborn; 926 outborn) in the hospital, 74 (4.24%) were diagnosed having AKI. Male: female ratio of 1.69:1 was observed in the admissions during the study period (1096 males; 649 females). Out of 74 neonates, 20 were inborn, 61 were male and 52 were born through vaginal delivery.

Distribution of various factors across both the groups are given in [Table/Fig-1a,b]. The difference between mean birth weight and gestational age of AKI and non-AKI babies was statistically significant (2.35 kg v/s 1.94 kg, p-value < 0.001; 37.44 weeks v/s 34.96 weeks, p=0.001).

The female gender was less associated with AKI (28.9% v/s 47.3 %; p=0.032). AKI was more common in term babies (54.1% v/s 28.4 %; p=0.001), outborn babies (p=0.011) and low ApGAR score at one minute and five minute Appropriate for Gestational Age (AGA) newborns had more chances of AKI than Small for Gestational Age (SGA)(53.7% v/s 28.9%; p=0.001). One baby out of 22 babies <1500 gm had AKI, 30 babies out of 84 babies between birth weights 1500 - <2500 gm had AKI, and 39 out of 59 babies with birth weight 2500 gm or more had AKI. Birth weights of four babies were not available. There were no cases of congenital structural anomaly of urinary tract in either group.

On multivariate logistic regression analysis, variables found significantly associated with AKI were male gender (OR=2.84, CI=1.12-7.21) and sepsis (OR=14.46, CI=4.5-46.46) [Table/Fig-2]. RDS, NEC were not included in analysis because of collinearity. ApGAR scores were not included because it reduced the sample size.

Out of those 74 neonates, 21(28.4%) were discharged, 38(51.3%) newborns went Discharge against Medical Advice (DAMA) and 15(20.3%) newborns died. Survival (discharged babies) amongst AKI was significantly associated with no requirement of mechanical ventilation and absence of shock. DAMA were considered as non survivors as being a tertiary care center, non-affordability with poor prognosis were the reasons for DAMA [Table/Fig-3].

We provided a descriptive overview of AKI in newborns that were admitted to NICU. The incidence of AKI in our unit was 4.24 % during the study period. A previous study from India found the incidence of AKI in newborns to be 3.9 in 1000 live births and 34.5 in 1000 newborns admitted in the NICU [16]. A Turkish study using criterion of serum creatinine >1.5 mg/dl showed an incidence of AKI as 3.4% [12].

The wide variability of incidence of AKI in the available data from different units can be attributed to demographic characteristics of population studied, and secondly no consensus definition of AKI was used. There have been two recent studies in similar population (critically ill neonates); one using urine output with serum creatinine as the criteria and the other one only using serum creatinine. The incidence of AKI was 20% and 6.3% respectively; highlighting the importance of having fixed definitions of AKI [17,18].

Most of the published studies, especially older, have used arbitrary definitions of AKI; one frequently used is absolute serum creatinine >1.5 mg/dl [12–14], other studies have used risk, injury, failure, loss of kidney function, and End-stage kidney disease (RIFLE) and Acute Kidney Injury Network (AKIN) criteria, which are not meant for neonatal population [13]. Recently, Jetton JG and Askenazi [19] proposed a new definition in 2012 [Table/Fig-4]. It graded severity of AKI using changes in serum creatinine and urine output. Subsequently, in April 2013 the group of neonatologists and paediatric nephrologists at the National Institute of Health (NIH) neonatal AKI workshop recommended the use of this definition [14]. However, the group stressed on the need of large multicenter cohort studies to validate its ability to predict clinical outcomes; Assessment of Worldwide Acute Kidney Injury Epidemiology in Neonates (AWAKEN) trial [20]. We could not use the new definition as we used retrospective data collection and could not find quantitatively documented urine output for the babies.

Various studies suggest that AKI is common in VLBW/ELBW newborns and is associated with poor prognosis [6,13,21]. Koralkar et al., reported incidence of AKI using modified KDIGO criteria to be 18% amongst 229 VLBW infants. They also reported higher mortality in the AKI group (p-value <0.001) [6]. Vishwanathan S et al., and Carmody JB et al., also reported similar findings [7,22]. A previous study from India showed that the percentage of babies with birth weight of <2500 gm in AKI group was higher than in healthy neonates [9]. Interestingly, we observed higher incidence in term babies, this could be attributed to the fact that a major portion of full term neonates catered in our study were referred for sepsis or asphyxia, which also form a high risk group for AKI.

In current study, we observed predominance of male newborns (n= 61; 82.4%) in the AKI group, in accordance with previous study [23]. Another recent NICU study although reported higher prevalence of AKI among females [10].

As a result of unique neonatal renal physiology, maternal exposures and perinatal events can lead to AKI. Perinatal risk factors associated with neonatal AKI include intubation at birth, low ApGAR scores, low cord pH and asystole [13]. Recently Bolat F et al., confirmed association between intubation at birth and AKI in NICU population [24].

Sepsis has been consistently associated as a risk factor for development of AKI in various studies conducted around the world; contributing to as high as 78% cases in some neonatal studies [10,18,24]. Another study from India by Mathur NB et al., shows that out of 200 newborns with sepsis, 26% developed AKI. The study also concluded that those with AKI had lower birth weight, and were prone for meningitis, disseminated intravascular coagulation and septic shock [9]. The newborns with sepsis are thought to be predisposed for AKI as a result of hypotension secondarily to sepsis and a direct damaging effect on renal microvasculature [13].

We found RDS to be more common in non-AKI group. This finding however, can also be due to more proportion of premature babies in the control group as compared to AKI group. The limited literature supports a positive association between AKI and RDS. A recent study by Momtaz HE et al., reported RDS as a third most common association with AKI (34.6%) after sepsis and dehydration [10].

We found significant association low ApGAR scores at one minute and five minutes with AKI. Surprisingly asphyxia did not show any association. Newborns with perinatal asphyxia have been associated with a higher risk of AKI. As Apgar scores are dependent on asphyxia, low Apgar scores have also been shown to be associated with AKI [13]. Several previous studies have found birth asphyxia to be the most common cause of AKI of neonatal period [12,16,25]. Perinatal asphyxia is associated with acute tubular injury which is the most common cause of intrinsic AKI. Two recent studies reported an association between asphyxia and AKI using modern definition for AKI [8,26]. Selewski DT et al., reported an incidence of 38% of AKI and Kaur S et al., reported 41.67%. The second study was conducted amongst neonates undergoing therapeutic hypothermia for perinatal asphyxia [8,26].

We observed an association of need of mechanical ventilation and shock with AKI for mortality. Several theories have been proposed in support of this hypothesis. Major mechanisms involved compromised renal blood flow because of hypercapnia or hypoxemia; and barotrauma induced pulmonary inflammatory reaction leading to secondary systemic inflammatory reaction [27,28]. A recent NICU study from Turkey also reported similar association [24].

In the study by Mathur NB et al., amongst various clinical factors, only shock was found to be as a significant predictor for mortality [9]. In the study by Tellier B et al., age <24 hours, underlying diseases, low urine output and multiorgan failure were shown as prognostic factors [29]. Another study by Agras PI [12], demonstrates that intrinsic AKI, need for dialysis and mechanical ventilation were associated with higher mortality rates and no significant correlation was found between mortality rate and prematurity, serum blood urea nitrozen and creatinine level and perinatal factors. In the current study peritoneal dialysis did not have any prognostic implication. This might be because of low sample size.

A recent study by Esfandier N et al., mentions hyaline membrane disease (HMD), using mechanical ventilation, the need to use surfactant, low Apgar score, high blood PCO₂, high serum creatinine level, and low birth weight being related to mortality [30].

The inability to use the most accepted definition for neonatal AKI. Single center data reduces generalizability. The retrospective design of the study is a limitation that restricts our understanding to certain associations only. The control group was not matched with the cases; this may have restricted the clear delineation of associated factors. Absence of long term follow up precludes comments on delayed renal sequelae.

Male gender and sepsis came out as significant risk factors for AKI. Major factors associated with mortality in AKI were presence of shock and need for mechanical ventilation. As we move to gentler ventilation modes, better care, and smaller neonates, continued review of prospectively collected data are necessary to establish prognostic factors. Further studies are needed to assess the impact of low ApGAR scores on AKI.