Clinical Determinants and Mortality Predictors of Acute Kidney Injury in Neonates: A Five-Year Retrospective Cohort Study

Abstract

Objectives

Acute kidney injury (AKI) is a frequent and clinically significant complication in critically ill neonates and is associated with increased morbidity and mortality. Although awareness of neonatal AKI has increased in recent years, data on mortality predictors and the trajectory of renal recovery remain limited, particularly in single-center cohorts. This study aimed to comprehensively evaluate the incidence, antenatal, perinatal, and postnatal risk factors, clinical characteristics, and short-term outcomes of neonates diagnosed with AKI in a neonatal intensive care unit, with a particular focus on deaths attributable to AKI during the neonatal intensive care unit (NICU) stay.

Methods

In this single-center retrospective cohort study, the medical records of neonates admitted to the NICU between January 2019 and January 2024 were retrospectively reviewed. Neonates diagnosed with AKI according to the KDIGO criteria were included, and relevant demographic, clinical, laboratory, and outcome data were collected. Survivors and non-survivors were compared. Univariate and multivariable logistic regression analyses were performed to identify predictors of mortality, and model performance was assessed using receiver operating characteristic (ROC) curve analysis.

Results

Among 3,456 NICU admissions, 121 neonates (3.5%) were diagnosed with AKI. The mean gestational age was 31.0±5.5 weeks, and 78.5% of patients were preterm. The overall in-hospital mortality rate among neonates with AKI was 30.6%. Sepsis (54.5%) and perinatal asphyxia (40.5%) were common risk factors. Non-survivors had significantly lower gestational age and birth weight and higher serum creatinine levels on postnatal day 7 (p=0.04). In multivariable analysis, sepsis (adjusted OR: 2.64, 95% CI: 1.19–5.86) and serum creatinine on postnatal day 7 (adjusted OR: 7.98, 95% CI: 2.47–25.80) were independent predictors of mortality. The multivariable model demonstrated good discrimination, with an AUC of 0.857.

Conclusion

Neonatal acute kidney injury-related mortality in the neonatal intensive care unit remains high. In this cohort, sepsis and persistently elevated serum creatinine levels on postnatal day 7 were identified as independent predictors of in-hospital mortality. The observed divergence in creatinine trajectories between survivors and non-survivors highlights the prognostic value of serial renal function assessment for risk stratification. These findings underscore the importance of early identification of high-risk neonates and support the need for future prospective studies to determine whether targeted monitoring and intervention strategies can improve clinical outcomes.

Acute kidney injury (AKI) is a common and clinically significant complication among critically ill neonates admitted to neonatal intensive care units (NICUs). Owing to the immaturity of renal structure and function, neonates—particularly preterm and low-birth-weight infants—are uniquely vulnerable to renal insults during the perinatal period.^[1-4] In recent years, growing awareness of neonatal AKI has led to a substantial increase in research focusing on its epidemiology, risk factors, and short- and long-term outcomes.

Neonatal AKI has been consistently associated with adverse clinical outcomes, including prolonged mechanical ventilation, extended hospital stays, increased health care costs, and higher mortality rates.^[2,5-7] Moreover, accumulating evidence suggests that neonates who experience AKI are at increased risk of developing chronic kidney disease (CKD), hypertension, and reduced renal functional reserve later in life, underscoring the long-term consequences of early-life kidney injury.^[8-11]

The pathogenesis of AKI in neonates is multifactorial and typically involves a complex interplay of antenatal, perinatal, and postnatal risk factors. Antenatal conditions such as preeclampsia, intrauterine growth restriction, chorioamnionitis, and premature rupture of membranes may impair fetal renal perfusion and nephrogenesis.^[12,13] Perinatal factors, particularly perinatal asphyxia, low Apgar scores, and the need for cardiopulmonary resuscitation at birth, contribute significantly to renal hypoxic–ischemic injury.^[14-16] In the postnatal period, sepsis, hemodynamic instability, exposure to nephrotoxic medications, and the need for invasive respiratory support further exacerbate the risk of AKI development.^[5,17-20]

Historically, the lack of a standardized definition of neonatal AKI limited the comparability of studies and hindered progress in this field. The adaptation of the Kidney Disease: Improving Global Outcomes (KDIGO) criteria for neonates has represented a major advance, providing a unified framework based on changes in serum creatinine and urine output. However, its applicability in neonates remains challenging because of non-steady-state creatinine kinetics.^[21-26] The widespread adoption of KDIGO criteria over the past decade has facilitated more accurate identification of AKI and enabled meaningful comparisons across studies and populations.

One of the most influential contributions to the field is the Assessment of Worldwide Acute Kidney Injury Epidemiology in Neonates (AWAKEN) study, a large multicenter, multinational cohort study that demonstrated that neonatal AKI was independently associated with increased mortality and prolonged hospitalization.^[6] The AWAKEN study also highlighted that AKI occurs across all gestational age groups, challenging the earlier perception that it is predominantly a problem among extremely preterm infants.^[6]

As the clinical impact of neonatal AKI has become clearer, research efforts have increasingly focused on early risk stratification, surveillance strategies, and preventive approaches. These include the identification of high-risk populations, such as neonates exposed to nephrotoxic medications, as well as the exploration of novel biomarkers and diagnostic tools aimed at detecting kidney injury before overt increases in serum creatinine occur.^{[1,12,25,27-30]} Additionally, recent studies have suggested that targeted pharmacologic interventions, such as the judicious use of methylxanthines, may reduce the incidence of AKI in selected high-risk neonatal populations.^{[17,24,26,29,30]}

Despite these advances, neonatal AKI remains underrecognized in daily clinical practice, and data from single-center cohorts continue to play a critical role in elucidating local epidemiology, management strategies, and outcome determinants. In this context, the present study aims to evaluate the incidence, antenatal and postnatal risk factors, treatment modalities, and morbidity and mortality outcomes of neonates diagnosed with AKI in a tertiary-level NICU over a five-year period. Furthermore, we sought to identify clinical and laboratory factors associated with mortality among neonates with AKI, thereby contributing to improved risk assessment and targeted management strategies in this vulnerable population.

Methods

Study Design and Setting

This study was designed as a single-center, retrospective observational cohort study conducted in the Neonatal Intensive Care Unit (NICU) of University of Health Sciences, Şişli Hamidiye Etfal Education and Research Hospital. Medical records of neonates admitted to the NICU between January 2019 and January 2024 were retrospectively reviewed. No additional laboratory analyses or interventional procedures were performed, as all data were obtained from existing patient files.

All neonates hospitalized in the NICU during the study period were screened for a diagnosis of acute kidney injury (AKI). Infants who met the diagnostic criteria for AKI at any time during hospitalization were included in the study. The study was approved by the Institutional Review Board of the University of Health Sciences, Şişli Hamidiye Etfal Education and Research Hospital (approval no. 4358; date: March 26, 2024). All procedures were conducted in accordance with the Declaration of Helsinki and relevant national regulations. Because the study involved neonates, informed consent for participation and the use of clinical data was obtained from the parents or legal guardians of all infants included in the study.

Inclusion and Exclusion Criteria

Neonates admitted to the NICU between January 2019 and January 2024 who were diagnosed with AKI during their NICU stay according to the Kidney Disease: Improving Global Outcomes (KDIGO) criteria and had sufficient clinical and laboratory data available in their medical records were included in the study. Neonates with incomplete or missing clinical or laboratory data, infants who died before an adequate diagnostic evaluation could be performed, and those who were transferred to another medical center before completion of follow-up were excluded from the analysis.

Definition of Acute Kidney Injury

AKI was defined according to the modified neonatal KDIGO criteria, based on changes in serum creatinine and/or urine output. Specifically, AKI was identified by an increase in serum creatinine of ≥0.3 mg/dL within 48 hours or a rise to ≥1.5 times the baseline value and/or urine output <1 mL/kg/h.^[21]

Data Collection

Data were extracted from hospital electronic medical records and patient charts using a standardized case report form. Recorded variables included demographic characteristics (sex, gestational age, birth weight, and Apgar scores at 1 and 5 minutes) and antenatal risk factors, such as maternal conditions and pregnancy-related complications, when available in the medical records. Perinatal factors included the need for positive-pressure ventilation (PPV), cardiopulmonary resuscitation (CPR), and evidence of perinatal asphyxia, whereas postnatal factors comprised the presence of sepsis and the requirement for invasive or non-invasive respiratory support. Renal parameters included serum creatinine levels at diagnosis and during follow-up, as well as urine output patterns. Treatment modalities included fluid restriction, diuretic therapy, peritoneal dialysis, and hemodialysis. Clinical outcomes included associated morbidities and in-hospital mortality.

Subgroup Analysis and Outcome Measures

For outcome analysis, neonates diagnosed with AKI were categorized into two subgroups: survivors and non-survivors. Comparative analyses were performed between these groups to identify clinical, laboratory, and treatment-related factors associated with in-hospital mortality. The primary outcome of the study was in-hospital mortality among neonates diagnosed with AKI. Secondary outcomes included the identification of antenatal, perinatal, and postnatal risk factors associated with mortality, as well as treatment modalities used in AKI management.

Statistical Analysis

Statistical analyses were performed using IBM SPSS Statistics v27 (IBM Corp., Armonk, NY, USA). The distribution of continuous variables was assessed using visual methods (histograms and Q–Q plots) and analytical methods (Kolmogorov–Smirnov and Shapiro–Wilk tests). Continuous variables were expressed as mean±standard deviation for normally distributed data and as median (minimum–maximum) for non-normally distributed data. Categorical variables were summarized as frequencies and percentages. Comparisons between the survivor and non-survivor groups were conducted using the independent-samples t-test for normally distributed continuous variables and the Mann–Whitney U test for non-normally distributed variables. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate.

To identify factors associated with mortality, univariate logistic regression analysis was initially performed for clinically relevant variables, including gestational age, birth weight, sex, Apgar scores, perinatal asphyxia, presence of sepsis, urine output pattern, need for invasive respiratory support, and serum creatinine levels at diagnosis and during follow-up. Variables with a p value <0.10 in univariate analysis and those considered clinically significant were subsequently entered into a multivariable logistic regression model to determine independent predictors of mortality. Multicollinearity among independent variables was assessed using the variance inflation factor (VIF), and variables showing significant collinearity were excluded from the final model. The results of logistic regression analyses were reported as odds ratios (ORs) with 95% confidence intervals (CIs). Model calibration was evaluated using the Hosmer–Lemeshow goodness-of-fit test (p>0.05 indicating good fit), and model discrimination was assessed by calculating the area under the receiver operating characteristic (ROC) curve. A two-sided p value <0.05 was considered statistically significant for all analyses.

Results

During the five-year study period, a total of 3,456 neonates were admitted to the NICU. Among these, 121 neonates (3.5%) were diagnosed with acute kidney injury (AKI) and constituted the final study population (Fig. 1).

Figure 1.

Flowchart of patient selection and study population.

The baseline demographic, perinatal, and clinical characteristics of neonates with AKI are shown in Table 1. The mean gestational age was 31.0±5.5 weeks, and the mean birth weight was 1,650±973 g. The majority of patients were preterm infants (78.5%), and 52 neonates (43.0%) were male. Cesarean section was the predominant mode of delivery (58.7%). Median Apgar scores were 4 (range: 1–9) at 1 minute and 6 (range: 2–10) at 5 minutes. Delivery room resuscitation was frequently required, with positive-pressure ventilation (PPV) administered to 52.9% and cardiopulmonary resuscitation (CPR) to 18.2% of neonates. AKI was diagnosed at a mean postnatal age of 4.8±10.1 days. The mean estimated glomerular filtration rate at diagnosis was 11.9±5.8 mL/min/1.73 m². The overall AKI-related neonatal mortality rate was 30.6% (Table 1).

Table 1.

Baseline demographic and perinatal characteristics of neonates diagnosed with acute kidney injury

Variable	Patients (n=121)
Gestational age, weeks (mean±SD)	31±5.5
Birth weight, g (mean±SD)	1650±973
Male sex, n (%)	52 (43.0)
Mode of delivery
Cesarean section, n (%)	71 (58.7)
Vaginal delivery, n (%)	50 (41.3)
Gestational status
Preterm, n (%)	95 (78.5)
Term, n (%)	26 (21.5)
Apgar score, median (range)
1 minute	4 (1–9)
5 minutes	6 (2–10)
Delivery room resuscitation
Positive pressure ventilation (PPV), n (%)	64 (52.9)
Cardiopulmonary resuscitation (CPR), n (%)	22 (18.2)
AKI diagnosis time, postnatal day (mean ± SD, range)	4.8±10.1 (1–78)
Estimated GFR, mL/min/1.73 m2 (mean ± SD, range)	11.9±5.8 (2–28)
Urine output pattern
Normal, n (%)	42 (34.7)
Oliguria, n (%)	28 (23.1)
Polyuria, n (%)	27 (22.3)
Anuria, n (%)	24 (19.8)
Treatment modalities
Hydration only, n (%)	40 (33.1)
Diuretic therapy, n (%)	24 (19.8)
Peritoneal dialysis, n (%)	9 (7.4)
Continuous veno-venous hemodialysis (CVVHD), n (%)	2 (1.7)
Length of NICU stay, days (mean ± SD, range)	47.2±38.6 (1–180)
AKI-related neonatal mortality, n (%)	37 (30.6)

AKI: Acute kidney injury; GFR: Glomerular Filtration Rate; NICU: Neonatal intensive care units.

Antenatal, perinatal, and postnatal risk factors observed in neonates with AKI are shown in Table 2. At least one antenatal risk factor was identified in 52 neonates (43.0%). The most common antenatal factors were preeclampsia (32.2%) and premature rupture of membranes (PROM) (30.6%), followed by oligohydramnios (11.6%) and gestational diabetes mellitus (5.0%). Perinatal and postnatal risk factors were highly prevalent. Perinatal asphyxia was observed in 40.5% of cases, whereas hypoxic–ischemic encephalopathy (HIE) occurred in 17.4%. Postnatally, sepsis was the most frequent risk factor, affecting 54.5% of neonates. Exposure to nephrotoxic medications was common, including aminoglycoside use in 62.0% and vancomycin use in 18.2%. Mechanical respiratory support was required in 81.0% of patients. Among the patients, normal urine output was observed in 34.7%, whereas oliguria, polyuria, and anuria were present in 23.1%, 22.3%, and 19.8% of neonates, respectively (Table 2, Fig. 2).

Table 2.

Antenatal, perinatal, and postnatal risk factors in neonates with acute kidney injury (AKI) (n=121)

Risk factors	N	%
Antenatal (Prenatal)
Preeclampsia	39	32.2
Chorioamnionitis	3	2.5
Premature rupture of membranes (PROM)	37	30.6
Oligohydramnios	14	11.6
Polyhydramnios	1	0.8
ACE inhibitor exposure	1	0.8
Gestational diabetes mellitus (GDM)	6	5.0
Maternal sepsis	1	0.8
≥1 antenatal risk factor present	52	43.0
Perinatal / Postnatal
Perinatal asphyxia	49	40.5
Hypoxic–ischemic encephalopathy (HIE)	21	17.4
Therapeutic hypothermia	15	12.4
Aminoglycoside use	75	62.0
Vancomycin use	22	18.2
Antifungal therapy	6	5.0
Sepsis (clinical / culture-proven)	66	54.5
Dehydration	30	24.8
Patent ductus arteriosus (PDA)	32	26.4
Inotropic support	72	59.5
Respiratory support (mechanical ventilation)	98	81.0

Figure 2.

Distribution of urine output patterns among neonates with acute kidney injury.

The spectrum of morbidities associated with AKI is shown in Table 3. The most frequent comorbid conditions were intraventricular hemorrhage (61.2%), respiratory distress syndrome (46.3%), and necrotizing enterocolitis (37.2%). Bronchopulmonary dysplasia developed in 28.9% of neonates.

Table 3.

Laboratory parameters of neonates diagnosed with acute kidney injury

Parameter	Mean±SD	Min–Max
Urea (mg/dL)	83.01±53.32	14.00–272.00
Serum creatinine (mg/dL)	1.01±0.43	0.20–2.50
Sodium (mmol/L)	142.81±10.35	120.00–180.00
Potassium (mmol/L)	4.93±1.01	3.00–9.00
Fractional excretion of sodium (fENA, %)	3.17±2.96	1.00–14.00
Uric acid (mg/dL)	3.44±2.69	1.00–12.00
Lactate (mmol/L)	5.58±4.66	1.00–22.00
Bicarbonate (HCO3-, mmol/L)	15.87±5.81	0.00–38.00
Base excess (mmol/L)	−8.95±7.70	−29.00–5.00

Laboratory parameters at diagnosis are shown in Table 3. The mean serum creatinine level was 1.01±0.43 mg/dL, and the mean blood urea level was 83.0±53.32 mg/dL. Electrolyte disturbances and metabolic derangements were common, with elevated lactate levels (5.58±4.66 mmol/L) and reduced bicarbonate levels (15.87±5.81 mmol/L) (Table 3).

A comparison of clinical, laboratory, and treatment-related variables between survivors and non-survivors is shown in Table 4. Non-survivors had significantly lower gestational age (28.3±5.3 vs. 32.2±5.2 weeks, p<0.01) and birth weight (1,094±798 vs. 1,895±945 g, p<0.01). Perinatal asphyxia and sepsis were significantly more frequent among non-survivors than among survivors (p<0.01 and p=0.002, respectively). Serum creatinine levels on postnatal day 7 were significantly higher in non-survivors (p=0.04) (Table 4).

Table 4.

Comparison of clinical, laboratory, and treatment-related variables between survivors and non-survivors

Variable	Survivors (n=84)	Non-survivors (n=37)	p
Gestational age, weeks (mean±SD)	32.2±5.2	28.3±5.3	<0.01
Birth weight, g (mean±SD)	1895±945	1094±798	<0.01
Perinatal asphyxia, n (%)	26 (31.0)	23 (62.2)	<0.01
Sepsis, n (%)	11 (13.1%)	31 (83.8%)	0.002
Mechanical ventilation, n (%)	2 (5.4)	17 (20.2)	0.009
Serum creatinine day 3, mg/dL (mean±SD)	1.2±0.4	1.4±0.5	0.07
Serum creatinine day 7, mg/dL (mean±SD)	1.0±0.4	1.5±0.6	0.04

Trends in serum creatinine levels over time for survivors and non-survivors are depicted in Figure 3. In survivors, mean serum creatinine levels showed a gradual decline from 1.2±0.4 mg/dL at diagnosis to 1.2±0.4 mg/dL on postnatal day 3 and further decreased to 1.0±0.4 mg/dL on postnatal day 7. In contrast, non-survivors exhibited persistently elevated creatinine levels, increasing from 1.4±0.5 mg/dL at diagnosis to 1.4±0.5 mg/dL on postnatal day 3 and reaching 1.5±0.6 mg/dL on postnatal day 7 (Fig. 3).

Figure 3.

Trends in serum creatinine levels at diagnosis and during follow-up.

Univariate logistic regression analysis of factors associated with in-hospital mortality in neonates with AKI is shown in Table 5. Lower gestational age was significantly associated with an increased risk of mortality (OR 0.87, 95% CI 0.80–0.94, p=0.001), as was lower birth weight (OR 1.00 per 100 gram decrease, 95% CI 1.00–1.00, p<0.001). Perinatal asphyxia was associated with a more than threefold increase in mortality risk (OR 3.48, 95% CI 1.54–7.83, p=0.003). Similarly, the presence of sepsis significantly increased the risk of in-hospital mortality (OR 2.18, 95% CI 1.28–3.71, p=0.004). The requirement for mechanical ventilation was strongly associated with mortality (OR 6.76, 95% CI 1.46–31.25, p=0.014). Elevated serum creatinine levels on postnatal day 3 (OR 2.95, 95% CI 1.16–7.49, p=0.023) and postnatal day 7 (OR 8.68, 95% CI 3.11–24.23, p<0.001) were also significant predictors of mortality (Table 5).

Table 5.

Univariate logistic regression analysis of factors associated with in-hospital mortality in neonates with AKI

Variable	Odds Ratio (OR)	95% Confidence Interval	p
Gestational age (weeks)	0.87	0.80–0.94	0.001
Birth weight (per 100 g decrease)	1.00	1.00–1.00	<0.001
Perinatal asphyxia	3.48	1.54–7.83	0.003
Sepsis	2.18	1.28–3.71	0.004
Mechanical ventilation	6.76	1.46–31.25	0.014
Serum creatinine day 3 (mg/dL)	2.95	1.16–7.49	0.023
Serum creatinine day 7 (mg/dL)	8.68	3.11–24.23	<0.001

Odds ratio for birth weight is expressed per 100 g decrease.

Multivariable logistic regression analysis identifying independent predictors of mortality is shown in Table 6. Sepsis (adjusted OR: 2.64, 95% CI: 1.19–5.86, p=0.017) and serum creatinine level on postnatal day 7 (adjusted OR: 7.98, 95% CI: 2.47–25.80, p=0.001) emerged as independent predictors of in-hospital mortality. The discriminative performance of the multivariable model was assessed using ROC curve analysis. The model demonstrated good predictive ability, with an AUC of 0.857, sensitivity of 0.909, and specificity of 0.750 at the optimal cutoff value determined by the Youden index (Table 6, Fig. 4).

Table 6.

Multivariable logistic regression analysis identifying independent predictors of AKI-related neonatal mortality

Variable	Adjusted Odds Ratio (aOR)	95% Confidence Interval	p
Gestational age (weeks)	1.22	0.96–1.56	0.106
Birth weight (g)	1.00	1.00–1.00	0.066
Perinatal asphyxia	3.23	0.97–10.77	0.057
Sepsis	2.64	1.19–5.86	0.017
Mechanical ventilation	7.07	0.94–53.26	0.058
Serum creatinine day 7 (mg/dL)	7.98	2.47–25.80	0.001
Model	AUC	Optimal cut-off (Youden index)	Sensitivity	Specificity
Multivariable logistic regression model*	0.857	0.358	0.909	0.750

^*Model variables: gestational age, birth weight, perinatal asphyxia, sepsis, mechanical ventilation, and serum creatinine at day 7.

Figure 4.

Receiver operating characteristic (ROC) curve of the multivariable logistic regression model for prediction of mortality. AUC=0.857. * Model variables: gestational age, birth weight, perinatal asphyxia, sepsis, mechanical ventilation

AUC: Area under the curve

Discussion

In this single-center retrospective cohort of neonates with AKI, the key findings were: (i) an AKI incidence of 3.5% among NICU admissions; (ii) a high in-hospital mortality rate of 30.6%; (iii) substantial exposure to perinatal and postnatal stressors, notably respiratory support and sepsis; and (iv) in multivariable analysis, sepsis and serum creatinine on day 7 emerged as independent predictors of mortality, with the multivariable model demonstrating good discrimination (AUC 0.857).

Our observed AKI incidence (3.5%) is markedly lower than the approximately one-third incidence reported in large multicenter and meta-analytic studies. Jetton et al.^[6] (AWAKEN) reported an overall AKI incidence of approximately 30% and demonstrated that AKI was independently associated with worse outcomes. Meena et al.^[4] similarly reported that AKI affects a substantial proportion of neonates and is associated with increased mortality. A likely explanation is case definition and ascertainment. Our operational definition relies on absolute creatinine criteria (serum creatinine of ≥0.3 mg/dL within 48 hours or a rise to ≥1.5 times the baseline value and/or urine output <1 mL/kg/h), which, while pragmatic, can underestimate AKI in early neonatal life when baseline creatinine is influenced by maternal creatinine and non-steady-state physiology. Starr et al. ^[1] highlighted that neonatal AKI epidemiology varies substantially according to the definition used and the timing of creatinine measurement and emphasized the importance of harmonized approaches. This definitional sensitivity can plausibly shift incidence estimates downward in single-center retrospective datasets, particularly if urine output criteria or standardized surveillance are not uniformly applied. This discrepancy may also reflect differences in surveillance intensity and diagnostic practices across centers.

Antenatal factors were common in our cohort, most notably preeclampsia and PROM, supporting the concept that disrupted placental perfusion, inflammation, and altered nephrogenesis can predispose the kidney to postnatal injury. Perico et al.^[18] reviewed how maternal and environmental factors, including hypertensive disorders and infection/inflammation, may influence neonatal renal vulnerability and long-term kidney health.

Perinatal hypoxia-related stress was also prominent: perinatal asphyxia was frequent, was strongly associated with mortality in univariate analysis, and remained borderline significant in multivariable modeling. This finding aligns with broader evidence that renal hypoperfusion and ischemia–reperfusion injury are central mechanisms in neonatal AKI, particularly in the context of perinatal compromise. Meena et al.^[4] reported comparable AKI incidence in sepsis and perinatal asphyxia subgroups but highlighted higher mortality in the asphyxia group, consistent with our observation that perinatal asphyxia is a clinically meaningful mortality-associated factor even if it narrowly misses statistical significance after adjustment.

Sepsis was prevalent and emerged as an independent predictor of mortality (aOR 2.64) in our study. This finding is highly consistent with multicenter literature. Jetton et al.^[6] demonstrated that neonatal AKI is closely intertwined with the severity of critical illness and that AKI confers additional risk beyond baseline illness. Coleman et al.^[5] underscored sepsis and hemodynamic instability as major postnatal drivers of AKI and highlighted the clinical importance of early recognition and supportive management. Mechanistically, sepsis can exacerbate kidney injury through inflammation, microvascular dysfunction, and impaired perfusion; it also often necessitates exposure to vasoactive agents and nephrotoxic antimicrobials, further compounding renal risk. In our cohort, inotropic support and broad antimicrobial exposure were common, reflecting the typical “AKI ecosystem” in NICU practice.

Exposure to nephrotoxic agents, including aminoglycosides, vancomycin, and antifungals, was substantial in our cohort. This pattern is consistent with the AWAKEN nephrotoxin analysis, in which Steflik et al.^[17] demonstrated frequent nephrotoxic medication exposure in early life and explored its association with early AKI. From a preventive standpoint, this reinforces the value of “nephrotoxin stewardship,” including dose optimization, therapeutic drug monitoring, limiting combinations when possible, and structured renal surveillance. Beyond acute outcomes, longer-term concerns are increasingly evident: Ameta et al.^[11] reported that children with nephrotoxic medication exposure and AKI in infancy may show markers of CKD years later, supporting the need for follow-up strategies after neonatal renal injury.

Our cohort demonstrated heterogeneity in urine output patterns, including normal urine output, oliguria, polyuria, and anuria, highlighting that neonatal AKI is not uniformly oliguria-driven. This is clinically relevant because reliance on oliguria alone may result in missed cases of non-oliguric AKI and potentially delay recognition. Starr et al.^[1] emphasized improved surveillance and risk-stratification approaches, including attention to urine output and early biomarkers, to mitigate delayed diagnosis.

The morbidity profile in our AKI cohort—characterized by high rates of IVH, RDS, NEC, and BPD—suggests that AKI frequently coexists with systemic instability and multisystem immaturity, especially in predominantly preterm populations. This mirrors patterns described in contemporary neonatal AKI reviews and reinforces the notion that AKI often represents both a kidney-specific event and a marker of global physiologic stress. Chirico et al.^[29] reviewed how prematurity-related complications, respiratory failure, infection, and hemodynamic instability cluster with AKI in NICU settings and contribute to short- and long-term adverse outcomes.

A major contribution of our analysis is the demonstration that serum creatinine on day 7—reflecting failure of renal recovery—was a strong independent predictor of mortality (aOR 7.98). We also demonstrated divergent creatinine trajectories between survivors and non-survivors, with persistent elevation among non-survivors and improvement among survivors. This is conceptually aligned with the emerging distinction between transient and sustained AKI. Yen et al.^[2] reported that sustained AKI is associated with worse downstream outcomes, including neurodevelopment and growth, emphasizing that AKI persistence carries prognostic significance beyond an isolated increase in creatinine. Our findings extend this concept to acute hospital mortality: persistent creatinine elevation into the first week may serve as a pragmatic bedside marker for high-risk infants who may benefit from intensified monitoring, careful fluid and vasoactive management, nephrology consultation, and avoidance of further nephrotoxic exposure.

The multivariable model yielded an AUC of 0.857, with high sensitivity and reasonable specificity at the Youden-derived cutoff. This level of discrimination suggests meaningful clinical utility for risk stratification, particularly in resource-limited settings where early identification of high-risk infants may guide escalation of care, such as tighter hemodynamic targets, renal-protective antimicrobial choices, or earlier discussions regarding renal support. Nevertheless, overfitting cannot be excluded given the relatively limited sample size and single-center design; therefore, external validation is essential before any clinical implementation.

Our findings support three practical messages. First, sepsis prevention and prompt treatment remain central to reducing mortality in neonatal AKI. Second, day-7 creatinine appears to be a simple, high-yield prognostic marker and may help identify infants who require intensified renal-protective strategies. Third, systematic nephrotoxin stewardship and structured kidney surveillance protocols could represent high-impact, modifiable approaches in the NICU. Future work should prioritize prospective multicenter validation of our mortality model, incorporate standardized urine output criteria, explore early biomarkers, as described by Askenazi et al., ^[27,28] and evaluate whether protocolized kidney-support pathways improve both short-term survival and long-term kidney health.

Limitations

Strengths of this study include a multi-year cohort with granular clinical variables, subgroup comparisons between survivors and non-survivors, and a structured regression framework linking bedside features to mortality. Key limitations include the retrospective design, single-center setting, and lack of external validation of the predictive model, which may limit the generalizability of the findings. In addition, AKI may have been under-ascertained because of reliance on creatinine-based criteria and the variable completeness of urine output data. Some predictors, such as mechanical ventilation and perinatal asphyxia, showed borderline significance in multivariable analysis, possibly reflecting a limited number of events per variable and collinearity with markers of illness severity. Finally, because our analysis included only AKI cases, we describe the frequencies of risk factors within the AKI population rather than estimating population-level predictors relative to non-AKI controls.

Conclusion

In conclusion, the current study demonstrates that neonatal acute kidney injury is associated with a substantial burden of morbidity and a high in-hospital mortality rate. Among neonates with AKI, sepsis and persistently elevated serum creatinine levels on postnatal day 7 emerged as independent predictors of mortality, highlighting the prognostic importance of systemic infection and failure of early renal recovery. The observed divergence in creatinine trajectories between survivors and non-survivors underscores the clinical value of dynamic renal function monitoring rather than reliance on a single creatinine measurement at diagnosis. Additionally, the good discriminative performance of the multivariable model suggests that combining readily available clinical and laboratory parameters may aid in early risk stratification. Overall, our findings emphasize the need for heightened renal function surveillance, aggressive management of sepsis, and renal-protective strategies in high-risk neonates to potentially improve short-term outcomes.

Disclosures

Ethics Approval

Ethics committee approval for this study was obtained from the Şişli Hamidiye Etfal Training and Research Hospital (Approval no: 4358; date: March 26, 2024).

Informed consent

Since the study included newborns, informed consent for participation and the use of clinical data was obtained from the parents or legal guardians of all infants included in the study.

Conflict of Interest

The authors declare that there is no conflict of interest.

Financial Disclosure

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Use of AI for Writing Assistance

No artificial intelligence-assisted technologies were used in any section of the study.

Authorship Contributions

Concept – H.A., A.B., E.K.B., A.D.; Design – H.A., A.B., E.K.B., A.D.; Supervision – H.A., E.K.B., A.D., A.Y.T.; Fundings − H.A.; Materials − H.A., A.B., E.K.B., A.D., A.Y.T.; Data collection &/or processing – H.A., A.B., E.K.B., A.D., A.Y.T.; Analysis and/or interpretation – H.A., A.B., E.K.B., A.D., A.Y.T.; Literature search – H.A., A.D., A.Y.T.; Writing – H.A., A.B.; Critical review – H.A., E.K.B., A.D., A.Y.T.

Peer-review

Externally peer-reviewed.