Abstract: FR-PO0057
Rapid (<20 mins) Multiplex Biomarker Panels Predict AKI-Sepsis Subphenotypes
Session Information
- AKI: Epidemiology, Risk Factors, and Prevention
November 07, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Acute Kidney Injury
- 101 AKI: Epidemiology, Risk Factors, and Prevention
Authors
- Mabrey, Frances Linzee, University of Washington, Seattle, Washington, United States
- Zelnick, Leila R., University of Washington, Seattle, Washington, United States
- Bailey, Zoie A., University of Washington, Seattle, Washington, United States
- Lo, Jordan J., University of Washington, Seattle, Washington, United States
- Shak, Joshua R, Roche Diagnostics Corporation, Pleasanton, California, United States
- Liles, Wayne Conrad, University of Washington, Seattle, Washington, United States
- Bhatraju, Pavan K., University of Washington, Seattle, Washington, United States
Background
Previous work has identified two molecular AKI sub-phenotypes (SP1, SP2) that show differences in treatment responses and these sub-phenotypes are present in sepsis patients at risk of AKI. However, clinical use is limited by retrospective classification, slow biomarker processing (6–8 hrs), and measurement variability. We conducted a pilot study using the Roche Cobas® e 411 platform to rapidly (<20 mins) measure interleukin-6 (IL-6), angiopoietin-2 (Ang-2) and soluble triggering receptor expressed on myeloid cell-1 (sTREM-1) and assess their utility in sub-phenotype classification.
Methods
We prospectively enrolled a sepsis cohort collecting plasma within 24 hours of ICU admission and measured IL-6, Ang-2, and sTREM-1 using the Cobas e 411 platform. These assays are labeled Research Use Only, and only the Elecsys® IL-6 assay is available commercially in the US under Emergency Use Authorization. Sub-phenotypes were derived using published validated models. We then applied LASSO regression to fit models for predicting sub-phenotypes using Roche biomarkers. C-statistics used 5-fold cross-validation; 95% CIs derived from 2000-iteration bootstraps.
Results
Of 332 ICU patients with sepsis, 223 (67%) were male, 90 (27%) had AKI at enrollment, and hospital mortality was 8%. All three biomarkers had low (<10%) coefficients of variation. 316 participants were classified into SP1 and 16 as SP2. IL-6, Ang-2, and sTREM-1 in combination were robustly predictive of sub-phenotypes (AUC: 0.93; 95% CI: 0.91, 0.95) (Figure). This 3-biomarker model’s performance was significantly better than an IL-6 reference model (AUC: 0.72; 95% CI: 0.65–0.79). Adding construct validity, Roche derived SP2 was associated with renal replacement therapy by day 3 (sHR): 8.8 (95% CI: 2.4-32.1), p-value: 0.001)) and hospital mortality (HR: 3.24 (95% CI: 1.1-9.5); p-value: 0.03).
Conclusion
Rapid measurement of IL-6, sTREM-1, and Ang-2 on the Cobas e 411 platform predicts AKI/sepsis sub-phenotypes and could serve as enrichment tools in trials to identify treatment responses within diverse sepsis populations.
Funding
- NIDDK Support – Roche Diagnostics