The investigators ran a prospective observational cohort in the pediatric intensive care unit at Vietnam National Children’s Hospital between October 2024 and June 2025. Children aged 1 month to under 18 years were eligible when the Renal Angina Index (RAI) could be calculated within 24 hours of admission. A total of 314 patients met inclusion criteria and were followed for the pre-specified outcomes. The primary endpoint was severe acute kidney injury (AKI) at 72 hours, defined as KDIGO stage 2–3. Secondary endpoints included incident AKI, need for kidney replacement therapy, and PICU mortality. The study received Institutional Review Board approval (Approval No. 2424/BVNTW-HĐĐĐ; September 23, 2024) and parents or legal guardians were informed prior to participation.
RAI demonstrated good discrimination for the primary outcome. Data indicated an area under the receiver operating characteristic curve (AUC) of 0.790 (95% CI 0.741–0.834) for predicting severe AKI at 72 hours. Using a conventional cutoff of RAI ≥8, the sensitivity was 78.6% and specificity 70.5% in this cohort. Severe AKI developed in 56 of the 314 patients (17.8%). For the secondary outcome of incident AKI, the AUC was reported as 0.752 (95% CI 0.643–0.861). The study team observed that RAI offered a clinically meaningful early signal of risk within the first 24 hours of PICU admission.
Investigators compared RAI performance against two commonly used bedside markers: serum creatinine change and fluid overload. The RAI outperformed both serum creatinine and fluid overload in predicting severe AKI in this cohort. The abstract reports higher discrimination for RAI compared with those individual measures, suggesting that the composite RAI score captured risk elements not reflected by creatinine or fluid status alone. The source did not provide detailed numeric AUCs for serum creatinine and fluid overload side-by-side beyond the statement that RAI outperformed them, nor did it report exact delta-creatinine thresholds used for incident AKI in the comparison.
In multivariable logistic regression, an RAI ≥8 remained independently associated with severe AKI: adjusted odds ratio 7.40 (95% CI 2.51–21.78; p < 0.001). That association persisted after adjustment for covariates the study team included in the model. The cohort showed that RAI discrimination for incident AKI was lower than for severe AKI but still acceptable (AUC 0.752). The abstract lists secondary outcomes—incident AKI, kidney replacement therapy, and PICU mortality—but does not provide detailed event counts for kidney replacement therapy or mortality within the text of the source provided. The source likewise did not report the exact components and point-scoring algorithm values used to compute RAI in this cohort, beyond indicating the index was calculated within 24 hours of admission and evaluated at the ≥8 cutoff.
The study followed ethical standards and obtained appropriate approval; parents or guardians provided information before enrollment. Limitations reported (implicitly by the scope of the abstract) include the observational design and a single-center setting in Vietnam, which may affect generalizability. The source did not report granular data on how baseline creatinine was defined when unavailable, nor did it detail criteria or timing for initiation of kidney replacement therapy, or provide the full contingency of other clinical covariates included in the multivariable model. The sample included 314 patients with 56 events of severe AKI, but the abstract does not present subgroup analyses by age, primary diagnosis, or illness severity scores. These gaps identify concrete questions for clinicians: how RAI performance varies by disease subgroup, how to integrate RAI with biomarkers in low-resource settings, and whether RAI-guided pathways alter use of kidney replacement therapy or PICU outcomes. The study stops at these evidence gaps rather than extending into protocolized interventions.
Ethics approval details and registry information are documented in the source. The results reported here are limited to the data and analyses presented in the abstract; the source did not report further methodological granularity or additional outcomes beyond those summarized.