Shruthi Mohan, MBBS, MD

Acute Kidney Injury and Fluid Accumulation Following Neonatal Sepsis

Shruthi Mohan, MBBS, MD; Faris N. Al Gharaibeh, MD; Michael Santoro; Stuart L. Goldstein, MD, FAAP, FNKF; Cara L. Slagle, MD

Acute kidney injury and early fluid accumulation is associated with increased mortality in neonatal sepsis, like pediatric and adult sepsis. Future focus on judicious surveillance and prevention of acute kidney injury and fluid accumulation in neonatal sepsis can potentially mitigate some of the mortality.

Shruthi Mohan, MBBS, MD


Background: Acute kidney injury (AKI) and/or fluid overload are associated with increased morbidity and mortality in critically ill neonates. Data regarding neonatal AKI and fluid accumulation (FA) following sepsis are sparse in the neonatal intensive care unit (NICU).  

Objective: Describe the incidence of AKI and FA following neonatal sepsis and assess for associations with mortality.

Methods: This retrospective study included neonates with a positive blood culture, admitted to Cincinnati Children’s or University of Cincinnati NICUs from 6/2020 to 6/2021. Exclusion criteria: <5 days of antibiotics, no serum creatinine (SCr) data, or presence of a congenital kidney anomaly with dialysis dependence. AKI was defined by the neonatal modified Kidney Diseases: Improving Global Outcomes definition. FA was calculated for 7 days after positive culture by cumulative net fluid balance (L)/dry weight (kg). FA was not adjusted for insensible losses.

Results: 49 neonates had 54 sepsis events. Incidence of AKI was 33%(n=18/54). AKI diagnosis was by SCr criteria alone in 20%(n=11), urine output (UO) criteria alone in 2%(n=1) and combined SCr+UO criteria in 11%(n=6). AKI stage≥2 accounted for 55%(n=10). No subjects received renal replacement therapy. Mortality rate was 22%(n=11). Four subjects experienced death prior to 48 hours and were excluded from further FA analysis. Incidence of FA>10% occurring on day1 was 29%(13/45). Median time for FA>10% was 2 days[IQR:1,3].  FA>10% occurring on day1 was associated with AKI (50%vs15%, OR5.8, 95%CI 1.5-22.7, p=0.008). Subject demographics, nephrotoxic antibiotic exposure, and caffeine exposure did not differ between AKI and no AKI subgroups. Subjects with AKI were more likely to require vasopressor support (55%vs25%, OR:3.7, 95%CI 1.1-12.4, p=0.027) and had more SCr measurements (7vs3, p=0.013) (Table 1).  Mortality risk was increased in subjects with any AKI (47%vs9%, OR 8.5, 95%CI 1.9-39.4, p=0.003) and FA >10% occurring on day1(38%vs6%, OR: 9.4, 95%CI 1.5-57.6, p=0.007) (figure 1). Multivariable logistic regression model including AKI, gestational age at birth, birthweight, sex, and race showed AKI retained association with mortality, aOR 15.9, 95%CI 1.8-137.3, p 0.01 (table 2). 

Conclusions: AKI and early FA in neonatal sepsis was associated with higher mortality. Focus on AKI and FA prevention could potentially mitigate some of the mortality.

Table 1: Comparison of AKI and non-AKI groups

Variable AKI (n=18)No AKI (n=36)value
Male Sex6 (33%)21 (58%)0.083
Race  0.594
Black8 (44%)12 (33%) 
White9 (50%)23 (64%) 
Others1 (6%)1 (3%) 
Birth Gestational Age (weeks)26 [IQR: 24,36]28[IQR: 26,37]0.242
Corrected Gestational Age (weeks)31 [IQR: 28,43]35 [IQR: 28, 39]0.370
Small for Gestational Age 3 (17%)6 (17%)>0.999
Nephrotoxic Antibiotic Exposure   
Vancomycin8 (44%)19 (53%)0.564
Aminoglycoside4 (22%)5 (14%)0.439
Caffeine Exposure9 (50%)17 (47%)0.847
FA occurring on day 18 (50%)5 (15%)0.008
Vasopressor Exposure10 (55%)9 (25%)0.027
Number of SCr measurements7 [IQR:3,8]3 [IQR:2,5]0.013

Table 2: Multivariable logistic regression model for mortality in blood culture positive neonatal sepsis

Independent variablesb coefficientStandard ErrorWaldP valueExp (b)/OR95% C.I for Exp (b)
Lower Upper
Birth gestational age-0.2650.3800.4880.4850.7670.3651.614
Male Sex-1.9421.2102.5740.1090.1430.0131.538
Birth weight0.0030.0021.9920.1581.0030.9991.008
Small for gestational age1.1260.4540.4540.5003.0820.11781.30
white race-0.6081.2220.2480.6190.5440.0505.971