Pediatric Critical Care Fellow University of California, San Francisco San Francisco, California, United States
Background: One in four children admitted to the pediatric intensive care unit (ICU) develops acute kidney injury (AKI) within the first week, with many progressing to severe AKI. Severe AKI requiring dialysis (AKI-D) is linked to prolonged hospital stays, higher mortality, and long-term risks of hypertension, chronic kidney disease (CKD), and death. However, factors influencing kidney recovery in children and young adults remain unclear. This study examines associations between clinical and demographic factors and the likelihood of kidney recovery in children and young adults with AKI-D. By assessing factors such as achieved blood pressure, nephrotoxic exposure, dialysis modality, and illness severity, we aim to identify predictors of kidney recovery that could provide anticipatory guidance and prognostication. Objective: Identify demographic and clinical factors associated with kidney recovery after AKI-D in critically ill children and young adults. Design/Methods: This single-center retrospective cohort study uses de-identified electronic health records from 2012 onwards, including children and young adults (≤25 years) diagnosed with AKI-D during ICU stays. Exclusion criteria include pre-existing end-stage kidney disease. IRB approval was not required. Measurements: The primary outcome is time to AKI-D recovery, defined as survival and dialysis cessation for >7 days, with follow-up from dialysis initiation to cessation or death. Predictors include dialysis modality, nephrotoxic exposures (e.g., ARB, ACEI, NSAIDs, antibiotics), achieved blood pressure, pre-dialysis kidney function (urine output, eGFR, serum creatinine), and illness severity (PEWS, lactate, bicarbonate, ventilation status, vasopressor use) prior to dialysis. Demographic data (age, race/ethnicity, language) and comorbidities (hypertension, diabetes, malignancy, transplant history, congenital heart surgery) will be collected. Cox proportional hazards regression models will assess predictor’s association with cessation of dialysis. Data collection is expected by November, with analysis completion by January.
