Abstract: TH-PO0746
Biomarkers for Sepsis-Associated AKI: A Systematic Review and Network Meta-Analysis of Diagnostic and Prognostic Test Accuracy Studies
Session Information
- AKI: Critical Care, CKRT, Cardiorenal Syndromes, and Clinical Outcomes
October 22, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Acute Kidney Injury
- 102 AKI: Clinical, Outcomes, and Trials
Authors
- Hu, Yuanzhi, Hong Kong Baptist University School of Chinese Medicine, Hong Kong SAR, China
- Cheng, Haojin, Hong Kong Baptist University School of Chinese Medicine, Hong Kong SAR, China
- Xiong, Zihui, The Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, China
- Ng, Edmond S. W., London School of Hygiene & Tropical Medicine Faculty of Epidemiology and Population Health, London, England, United Kingdom
- Chan, Kam Wa, Hong Kong Baptist University School of Chinese Medicine, Hong Kong SAR, China
Background
The comparative diagnostic and prognostic accuracy of biomarkers for diagnosing acute kidney injury (AKI) in sepsis patients and predicting mortality in sepsis-associated AKI (SA-AKI) patients remains unknown.
Methods
Embase, PubMed, CENTRAL, and seven other databases were searched until July 2025. Clinical trials and cohorts assessing biomarker performance in sepsis or SA-AKI patients were included. We constructed a main network based on biomarkers reported in more than one publication and a full network including all biomarkers. We used the Nyaga-ANOVA model for network meta-analysis. The primary outcome was diagnostic odds ratio (DOR) with 95% credible intervals (CrIs) for 1) AKI diagnosis in sepsis, and 2) mortality or renal replacement therapy (RRT) prediction in SA-AKI. This study was registered in PROSPERO (CRD42025639203).
Results
82 studies were included in the meta-analysis. For diagnostic accuracy assessment of AKI in sepsis patients (69 studies, n=17,008), the main network identified urine kidney injury molecule-1 (KIM-1) [DOR 14.98 (95% CrI 3.49, 57.55)], urine Actin [13.71 (1.36, 86.69)], and urine neutrophil gelatinase-associated lipocalin/creatinine (NGAL/Cr) [11.65 (1.13, 75.76)] as having the highest DORs for predicting AKI (Figure 1), confirmed by pairwise meta-analyses. In the full network, urine KIM-1 and urine Actin remained in the top three. For prognostic accuracy assessment of mortality in SA-AKI patients (15 studies, n=12,287), blood Signal Peptide-Complement C1r/C1s, Uegf, Bmp1-Epidermal Growth Factor-like Domain-Containing Protein 2 (SCUBE2) showed the highest DOR for in-hospital mortality [11.02 (0.58, 125.11)], blood citrullinated histone H3 (citH3) for 28/30-day mortality [12.17 (0.20, 954.99)], and the formula of urine tissue inhibitor of metalloproteinases-2×insulin-like growth factor-binding protein-7 (TIMP-2×IGFBP7)+C-C motif chemokine ligand 14 (CCL14) for RRT prediction [8.03 (0.29, 248.85)].
Conclusion
Urine KIM-1, Actin, and NGAL/Cr are promising biomarkers for predicting AKI in sepsis patients. Blood SCUBE2, citH3, and urine TIMP-2×IGFBP7+CCL14 potentially predicts in-hospital mortality, 28/30-day mortality, and need for RRT in SA-AKI patients, respectively.