Background
The concept of acute kidney injury (AKI) substages has been recommended to better phenotype AKI and identify high-risk patient groups and therefore improve the diagnostic accuracy of AKI. However, there remains a gap between the recommendation and the clinical application. The study aimed to explore the incidence of AKI substages based on a sensitive AKI biomarker of urinary cystatin C (uCysC), and to determine whether AKI substages were relevant with respect to outcome in critically ill children.
Results
The multicenter cohort study enrolled 793 children in pediatric intensive care unit (PICU) of four tertiary hospitals in China. Children were classified as non-AKI, sub-AKI and AKI substages A and B according to uCysC level at PICU admission. Sub-AKI was defined by admission uCysC level ≥ 1.26 mg/g uCr in children not meeting the KDIGO criteria of AKI. In children who fulfilled KDIGO criteria, those with uCysC < 1.26 was defined as AKI substage A, and with ≥ 1.26 defined as AKI substage B. The associations of AKI substages with 30-day PICU mortality were assessed. 15.6% (124/793) of patients met the definition of sub-AKI. Of 180 (22.7%) patients with AKI, 90 (50%) had uCysC-positive AKI substage B and were more likely to have classical AKI stage 3, compared to substage A. Compared to non-AKI, sub-AKI and AKI substages A and B were risk factors significantly associated with mortality, and the association of sub-AKI (adjusted hazard ratio HR = 2.42) and AKI substage B (adjusted HR = 2.83) with mortality remained significant after adjustment for confounders. Moreover, AKI substage B had increased risks of death as compared with sub-AKI (HR = 3.10) and AKI substage A (HR = 3.19).
Conclusions
Sub-AKI defined/based on uCysC occurred in 20.2% of patients without AKI and was associated with a risk of death close to patients with AKI substage A. Urinary CysC-positive AKI substage B occurred in 50% of AKI patients and was more likely to have classical AKI stage 3 and was associated with the highest risk of mortality.