Diagnostic accuracy of renal angina index alone or in combination with biomarkers for predicting acute kidney injury in children

Early recognition of patients at risk for severe acute kidney injury (AKI) by renal angina index (RAI) may help in the early institution of preventive measures. Objective was to evaluate performance of RAI alone or in combination with biomarkers in predicting severe AKI (KDIGO stage 2 and 3 or equivalent) and receipt of kidney replacement therapy (KRT) in critically ill children. We searched PubMed, EMBASE, Web of Sciences, and CENTRAL for studies published till May 2021. Search terms included acute kidney injury, pediatrics, adolescent, renal angina index, and biomarker. Proceedings of relevant conferences and references of included studies were also scrutinized. Two reviewers independently assessed the study eligibility. Cohort and cross-sectional studies evaluating the diagnostic performance of RAI in predicting AKI or receipt of KRT in children were included. Eligible participants were the children less than 18 years with RAI assessment on day 0 ofadmission. We followed PRISMA-DTA guidelines and used the QUADAS-2 tool for quality assessment. A bivariate model for meta-analysis was used to calculate the summary estimates of diagnostic parameters. Major outcomes were the diagnostic accuracy of RAI (>= 8) alone or with biomarkers in predicting severe AKI and KRT receipt. Diagnostic accuracy was reported using summary sensitivity, specificity, and area under the curve (AUC). Overall, 22 studies (24 reports, 14,001 participants) were included. RAI >= 8 on day 0 has summary sensitivity, specificity, and AUC of 0.86 (95% CI, 0.77-0.92), 0.77 (0.68-0.83), and 0.88 (0.85-0.91) respectively for prediction of severe AKI on day 3. In comparison, a combination of RAI and urinary neutrophil gelatinase-associated lipocalin (NGAL) showed summary sensitivity, specificity, and AUC of 0.76 (0.62-0.85), 0.89 (0.74-0.96), and 0.87 (0.84-0.90) respectively for predicting severe AKI. The sensitivity, specificity, and AUC of RAI for predicting receipt of KRT were 0.82 (0.71-0.90), 0.74 (0.66-0.81), and 0.85 (0.81-0.88) respectively. In meta-regression, only the study setting (sepsis vs. heterogenous) was associated with heterogeneity. We observed substantial heterogeneity among eligible studies. Five studies had concerns in patient selection, and seven studies also had applicability concerns in patient selection for this review. Moderate certainty evidence showed that RAI >= 8 has good predicting ability in recognizing children at risk of severe AKI and receipt of KRT. The combination of urinary NGAL and RAI further improves the predicting ability (low-certainty evidence). Further studies are required on the context-driven assessment of novel biomarkers in the early prediction of AKI in RAI-positive children.