Prediction of Acute Kidney Injury on Admission to Pediatric Intensive Care

OBJECTIVES: Up to 37% of children admitted to the PICU develop acute kidney injury as defined by Kidney Disease: Improving Global Outcomes criteria. We describe the prevalence of acute kidney injury in a mixed pediatric intensive care cohort using this criteria. As tools to stratify patients at risk of acute kidney injury on PICU admission are lacking, we explored the variables at admission and day 1 that might predict the development of acute kidney injury.

DESIGN: Single-center retrospective observational study.

SETTING: Thirty-six-bed surgical/medical tertiary PICU.

PATIENTS: Children from birth to less than or equal to 16 years old admitted between 2015 and 2018.

INTERVENTIONS: None.

MEASUREMENTS AND MAIN RESULTS: Clinical data were extracted from the PICU clinical information system. Patients with baseline creatinine at admission greater than 20 micromol/L above the calculated normal creatinine level were classified as "high risk of acute kidney injury." Models were created to predict acute kidney injury at admission and on day 1. Out of the 7,505 children admitted during the study period, 738 patients (9.8%) were classified as high risk of acute kidney injury at admission and 690 (9.2%) developed acute kidney injury during PICU admission. Compared to Kidney Disease: Improving Global Outcomes criteria as the reference standard, high risk of acute kidney injury had a lower sensitivity and higher specificity compared with renal angina index greater than or equal to 8 on day 1. For the admission model, the adjusted odds ratio of developing acute kidney injury for high risk of acute kidney injury was 4.2 (95% CI, 3.3-5.2). The adjusted odds ratio in the noncardiac cohort for high risk of acute kidney injury was 7.3 (95% CI, 5.5-9.7). For the day 1 model, odds ratios for high risk of acute kidney injury and renal angina index greater than or equal to 8 were 3.3 (95% CI, 2.6-4.2) and 3.1 (95% CI, 2.4-3.8), respectively.

CONCLUSIONS: The relationship between high risk of acute kidney injury and acute kidney injury needs further evaluation. High risk of acute kidney injury performed better in the noncardiac cohort.