Abstract
BACKGROUND: Acute kidney injury (AKI) is a common complication in critically ill children. However, the common lack of baseline serum creatinine values affects AKI diagnosis and staging. Several approaches for estimating baseline creatinine values in those patients were evaluated. METHODS: This single-center retrospective study enrolled pediatric patients with documented serum creatinine measurements within 3 months before admission and more than two serum creatinine measurements within 7 days after admission to the pediatric intensive care unit of a tertiary care children's hospital between January 2016 and April 2020. Four different approaches for estimating AKI using serum creatinine measurements were compared: 1) back-calculation using age-adjusted normal reference glomerular filtration rates, 2) age-adjusted normal reference serum creatinine values, 3) minimum values measured within 7 days after admission, and 4) initial values upon admission. RESULTS: The approach using minimum values showed the best agreement with the measured baseline value, with the largest intraclass correlation coefficient (0.623), smallest bias (-0.04), and narrowest limit of agreement interval (1.032). For AKI diagnosis and staging, the minimum values were 80.8% and 76.1% accurate, respectively. The other estimated baseline values underestimated AKI and showed poor agreement with baseline values before admission, with a misclassification rate of up to 42% (p < 0.001). CONCLUSION: Minimum values of serum creatinine measured within 7 days after hospital admission showed the best agreement with creatinine measured within 3 months before admission, indicating the possibility of using it as a baseline when baseline data are unavailable. Further large-scale studies are required to accurately diagnose AKI in critically ill children.